Intro To Programming Using Java
Intro To Programming Using Java
David J. Eck
Hobart and William Smith Colleges
c
1996–2020, David J. Eck
Preface xi
i
CONTENTS ii
3 Control 71
3.1 Blocks, Loops, and Branches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
3.1.1 Blocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
3.1.2 The Basic While Loop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
3.1.3 The Basic If Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
3.1.4 Definite Assignment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
3.2 Algorithm Development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
3.2.1 Pseudocode and Stepwise Refinement . . . . . . . . . . . . . . . . . . . . 78
3.2.2 The 3N+1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
3.2.3 Coding, Testing, Debugging . . . . . . . . . . . . . . . . . . . . . . . . . . 84
3.3 while and do..while . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
3.3.1 The while Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
3.3.2 The do..while Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
3.3.3 break and continue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
3.4 The for Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
3.4.1 For Loops . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
3.4.2 Example: Counting Divisors . . . . . . . . . . . . . . . . . . . . . . . . . . 95
3.4.3 Nested for Loops . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
3.5 The if Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
3.5.1 The Dangling else Problem . . . . . . . . . . . . . . . . . . . . . . . . . . 101
3.5.2 Multiway Branching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
3.5.3 If Statement Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
3.5.4 The Empty Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
3.6 The switch Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
3.6.1 The Basic switch Statement . . . . . . . . . . . . . . . . . . . . . . . . . . 108
3.6.2 Menus and switch Statements . . . . . . . . . . . . . . . . . . . . . . . . . 110
3.6.3 Enums in switch Statements . . . . . . . . . . . . . . . . . . . . . . . . . 112
3.6.4 Definite Assignment and switch Statements . . . . . . . . . . . . . . . . . 112
3.6.5 A New switch Statement Syntax . . . . . . . . . . . . . . . . . . . . . . . 113
3.7 Exceptions and try..catch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
3.7.1 Exceptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
3.7.2 try..catch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
3.7.3 Exceptions in TextIO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
CONTENTS iii
4 Subroutines 141
4.1 Black Boxes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
4.2 Static Subroutines and Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
4.2.1 Subroutine Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
4.2.2 Calling Subroutines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
4.2.3 Subroutines in Programs . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
4.2.4 Member Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
4.3 Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
4.3.1 Using Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
4.3.2 Formal and Actual Parameters . . . . . . . . . . . . . . . . . . . . . . . . 153
4.3.3 Overloading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
4.3.4 Subroutine Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
4.3.5 Array Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
4.3.6 Command-line Arguments . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
4.3.7 Throwing Exceptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
4.3.8 Global and Local Variables . . . . . . . . . . . . . . . . . . . . . . . . . . 160
4.4 Return Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
4.4.1 The return statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
4.4.2 Function Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
4.4.3 3N+1 Revisited . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
4.5 Lambda Expressions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
4.5.1 First-class Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
4.5.2 Functional Interfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
4.5.3 Lambda Expressions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
4.5.4 Method References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
4.6 APIs, Packages, Modules, and Javadoc . . . . . . . . . . . . . . . . . . . . . . . . 171
4.6.1 Toolboxes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
4.6.2 Java’s Standard Packages . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
4.6.3 Using Classes from Packages . . . . . . . . . . . . . . . . . . . . . . . . . 174
4.6.4 About Modules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
4.6.5 Javadoc . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177
4.6.6 Static Import . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
4.7 More on Program Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180
4.7.1 Preconditions and Postconditions . . . . . . . . . . . . . . . . . . . . . . . 180
CONTENTS iv
Glossary 735
Preface
xi
Preface xii
Chapters 9 and 10 cover recursion and data structures, including generic programming and
the Java Collection Framework. Chapter 11 is about files and networking. Chapter 12 cov-
ers threads and parallel processing. Finally, Chapter 13 returns to the topic of graphical user
interface programming to cover some of JavaFX’s more advanced capabilities.
∗ ∗ ∗
The biggest change from the seventh edition of this textbook is a switch from Swing to
JavaFX for GUI programming. With just a few exceptions in code that the user is not meant
to read, all example code and exercises use JavaFX. (Unfortunately, even as I was working
on the new edition, the future of JavaFX has become unclear because, as of Java 11, JavaFX
is no longer included as part of the basic Java download. However, it is still be available to
be downloaded and installed separately. People who are interested in using Swing instead of
JavaFX might consider using Chapter 6 and Chapter 13 from Version 7 in place of the same
chapters from Version 8.) The eighth edition also incorporates two features that were introduced
in Java 8: lambda expressions and the stream API. Lambda expressions are covered fairly early,
as part of the material on subroutines in Chapter 4. They are used extensively with JavaFX in
Chapter 6 and Chapter 13, as well as with the stream API, which is introduced in Section 10.6.
Aside from these major changes, there are small improvements throughout, such as a short
discussion of loop invariants and class invariants in Section 8.2, increased coverage of Java’s
higher level concurrency support in Chapter 12, and the consistent use of factory methods like
Integer.valueOf(n) in preference to constructors.
Version 8.1 was a fairly small update to Version 8.0, and Version 8.1.1 makes no substantive
changes. The main motivation for the update to 8.1 was to make it clearer how to use the book
with Java 11 and later, but I have also taken the opportunity to briefly discuss a few features
that were introduced after Java 8. The major change to the language itself was the introduction
of modules in Java 9; these were already mentioned in Version 8.0, but that discussion has been
updated. (However, this book does not explain how to create modules.) Aside from that, there
have been few language changes that impact this textbook, but I do introduce the use of var
for declaring variables, several new methods related to the stream API, and the revisions made
to the switch statement in Java 14.
∗ ∗ ∗
The first version of the book was written in 1996, and there have been several editions since
then. All editions are archived (at least until my retirement) at the following Web addresses:
• First edition: http://math.hws.edu/eck/cs124/javanotes1/ (Covers Java 1.0.)
• Second edition: http://math.hws.edu/eck/cs124/javanotes2/ (Covers Java 1.1.)
• Third edition: http://math.hws.edu/eck/cs124/javanotes3/ (Covers Java 1.1.)
• Fourth edition: http://math.hws.edu/eck/cs124/javanotes4/ (Covers Java 1.4.)
• Fifth edition: http://math.hws.edu/eck/cs124/javanotes5/ (Covers Java 5.0.)
• Sixth edition: http://math.hws.edu/eck/cs124/javanotes6/ (Covers Java 5.0, with a bit
of 6.0.)
• Seventh edition: http://math.hws.edu/eck/cs124/javanotes7/ (Covers Java 7.)
• Eighth edition: http://math.hws.edu/eck/cs124/javanotes8/ (Covers Java 8.)
Introduction to Programming using Java is free, but it is not in the pub-
lic domain. Version 8 is published under the terms of the Creative Commons
Attribution-NonCommercial-ShareAlike 4.0 License. To view a copy of this license, visit
http://creativecommons.org/licenses/by-nc-sa/4.0/. For example, you can:
Preface xiii
• Post an unmodified copy of the on-line version on your own Web site (including the parts
that list the author and state the license under which it is distributed!).
• Give away unmodified copies of this book or sell them at cost of production, as long as
they meet the requirements of the license.
• Use the book as a textbook for a course that you are teaching (even though the students
are paying to take that course).
• Make modified copies of the complete book or parts of it and post them on the web or
otherwise distribute them non-commercially, provided that attribution to the author is
given, the modifications are clearly noted, and the modified copies are distributed under
the same license as the original. This includes translations to other languages.
For uses of the book in ways not covered by the license, permission of the author is required.
While it is not actually required by the license, I do appreciate hearing from people who
are using or distributing my work.
∗ ∗ ∗
A technical note on production: The on-line and PDF versions of this book are created
from a single source, which is written largely in XML. To produce the PDF version, the XML
is processed into a form that can be used by the TeX typesetting program. In addition to XML
files, the source includes DTDs, XSLT transformations, Java source code files, image files, a
TeX macro file, and a couple of scripts that are used in processing. The scripts work on Linux
and on Mac OS.
I have made the complete source files available for download at the following
address:
http://math.hws.edu/eck/cs124/downloads/javanotes8-full-source.zip
These files were not originally meant for publication, and therefore are not very cleanly
written. Furthermore, it requires a fair amount of expertise to use them. However, I have had
several requests for the sources and have made them available on an “as-is” basis. For more
information about the sources and how they are used see the README file from the source
download.
∗ ∗ ∗
Professor David J. Eck
Department of Mathematics and Computer Science
Hobart and William Smith Colleges
300 Pulteney Street
Geneva, New York 14456, USA
Email: [email protected]
WWW: http://math.hws.edu/eck/
Chapter 1
When you begin a journey, it’s a good idea to have a mental map of the terrain you’ll be
passing through. The same is true for an intellectual journey, such as learning to write computer
programs. In this case, you’ll need to know the basics of what computers are and how they
work. You’ll want to have some idea of what a computer program is and how one is created.
Since you will be writing programs in the Java programming language, you’ll want to know
something about that language in particular and about the modern computing environment for
which Java is designed.
As you read this chapter, don’t worry if you can’t understand everything in detail. (In fact,
it would be impossible for you to learn all the details from the brief expositions in this chapter.)
Concentrate on learning enough about the big ideas to orient yourself, in preparation for the
rest of the book. Most of what is covered in this chapter will be covered in much greater detail
later in the book.
1
CHAPTER 1. THE MENTAL LANDSCAPE 2
location. The CPU can also store information in memory by specifying the information to be
stored and the address of the location where it is to be stored.
On the level of machine language, the operation of the CPU is fairly straightforward (al-
though it is very complicated in detail). The CPU executes a program that is stored as a
sequence of machine language instructions in main memory. It does this by repeatedly reading,
or fetching , an instruction from memory and then carrying out, or executing , that instruc-
tion. This process—fetch an instruction, execute it, fetch another instruction, execute it, and
so on forever—is called the fetch-and-execute cycle. With one exception, which will be
covered in the next section, this is all that the CPU ever does. (This is all really somewhat
more complicated in modern computers. A typical processing chip these days contains several
CPU “cores,” which allows it to execute several instructions simultaneously. And access to
main memory is speeded up by memory “caches,” which can be more quickly accessed than
main memory and which are meant to hold data and instructions that the CPU is likely to
need soon. However, these complications don’t change the basic operation.)
A CPU contains an Arithmetic Logic Unit, or ALU, which is the part of the processor
that carries out operations such as addition and subtraction. It also holds a small number of
registers, which are small memory units capable of holding a single number. A typical CPU
might have 16 or 32 “general purpose” registers, which hold data values that are immediately
accessible for processing, and many machine language instructions refer to these registers. For
example, there might be an instruction that takes two numbers from two specified registers,
adds those numbers (using the ALU), and stores the result back into a register. And there
might be instructions for copying a data value from main memory into a register, or from a
register into main memory.
The CPU also includes special purpose registers. The most important of these is the pro-
gram counter , or PC. The CPU uses the PC to keep track of where it is in the program it
is executing. The PC simply stores the memory address of the next instruction that the CPU
should execute. At the beginning of each fetch-and-execute cycle, the CPU checks the PC to
see which instruction it should fetch. During the course of the fetch-and-execute cycle, the
number in the PC is updated to indicate the instruction that is to be executed in the next
cycle. Usually, but not always, this is just the instruction that sequentially follows the current
instruction in the program. Some machine language instructions modify the value that is stored
in the PC. This makes it possible for the computer to “jump” from one point in the program
to another point, which is essential for implementing the program features known as loops and
branches that are discussed in Section 1.4.
∗ ∗ ∗
A computer executes machine language programs mechanically—that is without under-
standing them or thinking about them—simply because of the way it is physically put together.
This is not an easy concept. A computer is a machine built of millions of tiny switches called
transistors, which have the property that they can be wired together in such a way that an
output from one switch can turn another switch on or off. As a computer computes, these
switches turn each other on or off in a pattern determined both by the way they are wired
together and by the program that the computer is executing.
Machine language instructions are expressed as binary numbers. A binary number is made
up of just two possible digits, zero and one. Each zero or one is called a bit. So, a machine
language instruction is just a sequence of zeros and ones. Each particular sequence encodes
some particular instruction. The data that the computer manipulates is also encoded as binary
numbers. In modern computers, each memory location holds a byte, which is a sequence of
CHAPTER 1. THE MENTAL LANDSCAPE 3
eight bits. A machine language instruction or a piece of data generally consists of several bytes,
stored in consecutive memory locations. For example, when a CPU reads an instruction from
memory, it might actually read four or eight bytes from four or eight memory locations; the
memory address of the instruction is the address of the first of those bytes.
A computer can work directly with binary numbers because switches can readily represent
such numbers: Turn the switch on to represent a one; turn it off to represent a zero. Machine
language instructions are stored in memory as patterns of switches turned on or off. When a
machine language instruction is loaded into the CPU, all that happens is that certain switches
are turned on or off in the pattern that encodes that instruction. The CPU is built to respond
to this pattern by executing the instruction it encodes; it does this simply because of the way
all the other switches in the CPU are wired together.
So, you should understand this much about how computers work: Main memory holds ma-
chine language programs and data. These are encoded as binary numbers. The CPU fetches
machine language instructions from memory one after another and executes them. Each in-
struction makes the CPU perform some very small task, such as adding two numbers or moving
data to or from memory. The CPU does all this mechanically, without thinking about or un-
derstanding what it does—and therefore the program it executes must be perfect, complete in
all details, and unambiguous because the CPU can do nothing but execute it exactly as written.
Here is a schematic view of this first-stage understanding of the computer:
Memory
CPU 10001010
00001100
(Location
(Location
0)
1)
10111000 (Location 2)
Registers Data to Memory 01000001 (Location 3)
00001011 (Location 4)
Data from Memory 11011101 (Location 5)
10110000 (Location 6)
01010010 (Location 7)
ALU 11111010 (Location 8)
01001100 (Location 9)
Address for 00100011 (Location 10)
PC
reading/writing
00011010 (Location 11)
data
. .
. .
. .
Input/ Data
Output Address
Controller Control
Now, devices such as keyboard, mouse, and network interface can produce input that needs
to be processed by the CPU. How does the CPU know that the data is there? One simple idea,
which turns out to be not very satisfactory, is for the CPU to keep checking for incoming data
over and over. Whenever it finds data, it processes it. This method is called polling , since
the CPU polls the input devices continually to see whether they have any input data to report.
Unfortunately, although polling is very simple, it is also very inefficient. The CPU can waste
an awful lot of time just waiting for input.
CHAPTER 1. THE MENTAL LANDSCAPE 5
To avoid this inefficiency, interrupts are generally used instead of polling. An interrupt
is a signal sent by another device to the CPU. The CPU responds to an interrupt signal by
putting aside whatever it is doing in order to respond to the interrupt. Once it has handled
the interrupt, it returns to what it was doing before the interrupt occurred. For example, when
you press a key on your computer keyboard, a keyboard interrupt is sent to the CPU. The
CPU responds to this signal by interrupting what it is doing, reading the key that you pressed,
processing it, and then returning to the task it was performing before you pressed the key.
Again, you should understand that this is a purely mechanical process: A device signals an
interrupt simply by turning on a wire. The CPU is built so that when that wire is turned on,
the CPU saves enough information about what it is currently doing so that it can return to
the same state later. This information consists of the contents of important internal registers
such as the program counter. Then the CPU jumps to some predetermined memory location
and begins executing the instructions stored there. Those instructions make up an interrupt
handler that does the processing necessary to respond to the interrupt. (This interrupt handler
is part of the device driver software for the device that signaled the interrupt.) At the end of
the interrupt handler is an instruction that tells the CPU to jump back to what it was doing;
it does that by restoring its previously saved state.
Interrupts allow the CPU to deal with asynchronous events. In the regular fetch-and-
execute cycle, things happen in a predetermined order; everything that happens is “synchro-
nized” with everything else. Interrupts make it possible for the CPU to deal efficiently with
events that happen “asynchronously,” that is, at unpredictable times.
As another example of how interrupts are used, consider what happens when the CPU needs
to access data that is stored on a hard disk. The CPU can access data directly only if it is
in main memory. Data on the disk has to be copied into memory before it can be accessed.
Unfortunately, on the scale of speed at which the CPU operates, the disk drive is extremely
slow. When the CPU needs data from the disk, it sends a signal to the disk drive telling it
to locate the data and get it ready. (This signal is sent synchronously, under the control of
a regular program.) Then, instead of just waiting the long and unpredictable amount of time
that the disk drive will take to do this, the CPU goes on with some other task. When the disk
drive has the data ready, it sends an interrupt signal to the CPU. The interrupt handler can
then read the requested data.
∗ ∗ ∗
Now, you might have noticed that all this only makes sense if the CPU actually has several
tasks to perform. If it has nothing better to do, it might as well spend its time polling for input
or waiting for disk drive operations to complete. All modern computers use multitasking to
perform several tasks at once. Some computers can be used by several people at once. Since the
CPU is so fast, it can quickly switch its attention from one user to another, devoting a fraction
of a second to each user in turn. This application of multitasking is called timesharing . But a
modern personal computer with just a single user also uses multitasking. For example, the user
might be typing a paper while a clock is continuously displaying the time and a file is being
downloaded over the network.
Each of the individual tasks that the CPU is working on is called a thread . (Or a process;
there are technical differences between threads and processes, but they are not important here,
since it is threads that are used in Java.) Many CPUs can literally execute more than one
thread simultaneously—such CPUs contain multiple “cores,” each of which can run a thread—
but there is always a limit on the number of threads that can be executed at the same time.
Since there are often more threads than can be executed simultaneously, the computer has to be
CHAPTER 1. THE MENTAL LANDSCAPE 6
able switch its attention from one thread to another, just as a timesharing computer switches
its attention from one user to another. In general, a thread that is being executed will continue
to run until one of several things happens:
• The thread might voluntarily yield control, to give other threads a chance to run.
• The thread might have to wait for some asynchronous event to occur. For example, the
thread might request some data from the disk drive, or it might wait for the user to press
a key. While it is waiting, the thread is said to be blocked , and other threads, if any, have
a chance to run. When the event occurs, an interrupt will “wake up” the thread so that
it can continue running.
• The thread might use up its allotted slice of time and be suspended to allow other threads
to run. Most computers can “forcibly” suspend a thread in this way; computers that
can do that are said to use preemptive multitasking . To do preemptive multitasking,
a computer needs a special timer device that generates an interrupt at regular intervals,
such as 100 times per second. When a timer interrupt occurs, the CPU has a chance to
switch from one thread to another, whether the thread that is currently running likes it
or not. All modern desktop and laptop computers, and even typical smartphones and
tablets, use preemptive multitasking.
Ordinary users, and indeed ordinary programmers, have no need to deal with interrupts and
interrupt handlers. They can concentrate on the different tasks that they want the computer to
perform; the details of how the computer manages to get all those tasks done are not important
to them. In fact, most users, and many programmers, can ignore threads and multitasking
altogether. However, threads have become increasingly important as computers have become
more powerful and as they have begun to make more use of multitasking and multiprocessing.
In fact, the ability to work with threads is fast becoming an essential job skill for programmers.
Fortunately, Java has good support for threads, which are built into the Java programming
language as a fundamental programming concept. Programming with threads will be covered
in Chapter 12.
Just as important in Java and in modern programming in general is the basic concept of
asynchronous events. While programmers don’t actually deal with interrupts directly, they do
often find themselves writing event handlers, which, like interrupt handlers, are called asyn-
chronously when specific events occur. Such “event-driven programming” has a very different
feel from the more traditional straight-through, synchronous programming. We will begin with
the more traditional type of programming, which is still used for programming individual tasks,
but we will return to threads and events later in the text, starting in Chapter 6
∗ ∗ ∗
By the way, the software that does all the interrupt handling, handles communication with
the user and with hardware devices, and controls which thread is allowed to run is called
the operating system. The operating system is the basic, essential software without which
a computer would not be able to function. Other programs, such as word processors and
Web browsers, are dependent upon the operating system. Common desktop operating systems
include Linux, various versions of Windows, and Mac OS. Operating systems for smartphones
and tablets include Android and iOS.
CHAPTER 1. THE MENTAL LANDSCAPE 7
essential features of Java: the same compiled program can be run on many different types of
computers.
Java Interpreter
for Mac OS
Java
Java Java Interpreter
Compiler Bytecode
Program for Windows
Program
Java Interpreter
for Linux
Why, you might wonder, use the intermediate Java bytecode at all? Why not just distribute
the original Java program and let each person compile it into the machine language of whatever
computer they want to run it on? There are several reasons. First of all, a compiler has to
understand Java, a complex high-level language. The compiler is itself a complex program.
A Java bytecode interpreter, on the other hand, is a relatively small, simple program. This
makes it easy to write a bytecode interpreter for a new type of computer; once that is done,
that computer can run any compiled Java program. It would be much harder to write a Java
compiler for the same computer.
Furthermore, some Java programs are meant to be downloaded over a network. This leads
to obvious security concerns: you don’t want to download and run a program that will damage
your computer or your files. The bytecode interpreter acts as a buffer between you and the
program you download. You are really running the interpreter, which runs the downloaded
program indirectly. The interpreter can protect you from potentially dangerous actions on the
part of that program.
When Java was still a new language, it was criticized for being slow: Since Java bytecode was
executed by an interpreter, it seemed that Java bytecode programs could never run as quickly
as programs compiled into native machine language (that is, the actual machine language of the
computer on which the program is running). However, this problem has been largely overcome
by the use of just-in-time compilers for executing Java bytecode. A just-in-time compiler
translates Java bytecode into native machine language. It does this while it is executing the
program. Just as for a normal interpreter, the input to a just-in-time compiler is a Java bytecode
program, and its task is to execute that program. But as it is executing the program, it also
translates parts of it into machine language. The translated parts of the program can then be
executed much more quickly than they could be interpreted. Since a given part of a program is
often executed many times as the program runs, a just-in-time compiler can significantly speed
up the overall execution time.
I should note that there is no necessary connection between Java and Java bytecode. A
program written in Java could certainly be compiled into the machine language of a real com-
puter. And programs written in other languages can be compiled into Java bytecode. However,
the combination of Java and Java bytecode is platform-independent, secure, and network-
compatible while allowing you to program in a modern high-level object-oriented language.
In the past few years, it has become fairly common to create new programming languages,
or versions of old languages, that compile into Java bytecode. The compiled bytecode programs
can then be executed by a standard JVM. New languages that have been developed specifically
CHAPTER 1. THE MENTAL LANDSCAPE 9
for programming the JVM include Scala, Groovy, Clojure, and Processing. Jython and JRuby
are versions of older languages, Python and Ruby, that target the JVM. These languages make
it possible to enjoy many of the advantages of the JVM while avoiding some of the technicalities
of the Java language. In fact, the use of other languages with the JVM has become important
enough that several new features have been added to the JVM specifically to add better support
for some of those languages. And this improvement to the JVM has in turn made possible some
new features in Java.
∗ ∗ ∗
I should also note that the really hard part of platform-independence is providing a “Graph-
ical User Interface”—with windows, buttons, etc.—that will work on all the platforms that
support Java. You’ll see more about this problem in Section 1.6.
other. However, this is obviously very limited: the computer would soon run out of instructions
to execute. Control structures are special instructions that can change the flow of control.
There are two basic types of control structure: loops, which allow a sequence of instructions
to be repeated over and over, and branches, which allow the computer to decide between two
or more different courses of action by testing conditions that occur as the program is running.
For example, it might be that if the value of the variable “principal” is greater than 10000,
then the “interest” should be computed by multiplying the principal by 0.05; if not, then the
interest should be computed by multiplying the principal by 0.04. A program needs some
way of expressing this type of decision. In Java, it could be expressed using the following “if
statement”:
if (principal > 10000)
interest = principal * 0.05;
else
interest = principal * 0.04;
(Don’t worry about the details for now. Just remember that the computer can test a condition
and decide what to do next on the basis of that test.)
Loops are used when the same task has to be performed more than once. For example,
if you want to print out a mailing label for each name on a mailing list, you might say, “Get
the first name and address and print the label; get the second name and address and print
the label; get the third name and address and print the label. . . ” But this quickly becomes
ridiculous—and might not work at all if you don’t know in advance how many names there are.
What you would like to say is something like “While there are more names to process, get the
next name and address, and print the label.” A loop can be used in a program to express such
repetition.
∗ ∗ ∗
Large programs are so complex that it would be almost impossible to write them if there
were not some way to break them up into manageable “chunks.” Subroutines provide one way to
do this. A subroutine consists of the instructions for performing some task, grouped together
as a unit and given a name. That name can then be used as a substitute for the whole set of
instructions. For example, suppose that one of the tasks that your program needs to perform
is to draw a house on the screen. You can take the necessary instructions, make them into
a subroutine, and give that subroutine some appropriate name—say, “drawHouse()”. Then
anyplace in your program where you need to draw a house, you can do so with the single
command:
drawHouse();
This will have the same effect as repeating all the house-drawing instructions in each place.
The advantage here is not just that you save typing. Organizing your program into sub-
routines also helps you organize your thinking and your program design effort. While writing
the house-drawing subroutine, you can concentrate on the problem of drawing a house without
worrying for the moment about the rest of the program. And once the subroutine is written,
you can forget about the details of drawing houses—that problem is solved, since you have a
subroutine to do it for you. A subroutine becomes just like a built-in part of the language which
you can use without thinking about the details of what goes on “inside” the subroutine.
∗ ∗ ∗
Variables, types, loops, branches, and subroutines are the basis of what might be called
“traditional programming.” However, as programs become larger, additional structure is needed
CHAPTER 1. THE MENTAL LANDSCAPE 11
to help deal with their complexity. One of the most effective tools that has been found is object-
oriented programming, which is discussed in the next section.
in known, well-defined ways. And it makes it easier for programs to use the module, since they
don’t have to worry about the details of how the data is represented. Information about the
representation of the data is hidden.
Modules that could support this kind of information-hiding became common in program-
ming languages in the early 1980s. Since then, a more advanced form of the same idea has
more or less taken over software engineering. This latest approach is called object-oriented
programming , often abbreviated as OOP.
The central concept of object-oriented programming is the object, which is a kind of module
containing data and subroutines. The point-of-view in OOP is that an object is a kind of self-
sufficient entity that has an internal state (the data it contains) and that can respond to
messages (calls to its subroutines). A mailing list object, for example, has a state consisting
of a list of names and addresses. If you send it a message telling it to add a name, it will
respond by modifying its state to reflect the change. If you send it a message telling it to print
itself, it will respond by printing out its list of names and addresses.
The OOP approach to software engineering is to start by identifying the objects involved in
a problem and the messages that those objects should respond to. The program that results is
a collection of objects, each with its own data and its own set of responsibilities. The objects
interact by sending messages to each other. There is not much “top-down” in the large-scale
design of such a program, and people used to more traditional programs can have a hard time
getting used to OOP. However, people who use OOP would claim that object-oriented programs
tend to be better models of the way the world itself works, and that they are therefore easier
to write, easier to understand, and more likely to be correct.
∗ ∗ ∗
You should think of objects as “knowing” how to respond to certain messages. Different
objects might respond to the same message in different ways. For example, a “print” message
would produce very different results, depending on the object it is sent to. This property of
objects—that different objects can respond to the same message in different ways—is called
polymorphism.
It is common for objects to bear a kind of “family resemblance” to one another. Objects
that contain the same type of data and that respond to the same messages in the same way
belong to the same class. (In actual programming, the class is primary; that is, a class is
created and then one or more objects are created using that class as a template.) But objects
can be similar without being in exactly the same class.
For example, consider a drawing program that lets the user draw lines, rectangles, ovals,
polygons, and curves on the screen. In the program, each visible object on the screen could be
represented by a software object in the program. There would be five classes of objects in the
program, one for each type of visible object that can be drawn. All the lines would belong to
one class, all the rectangles to another class, and so on. These classes are obviously related;
all of them represent “drawable objects.” They would, for example, all presumably be able to
respond to a “draw yourself” message. Another level of grouping, based on the data needed to
represent each type of object, is less obvious, but would be very useful in a program: We can
group polygons and curves together as “multipoint objects,” while lines, rectangles, and ovals
are “two-point objects.” (A line is determined by its two endpoints, a rectangle by two of its
corners, and an oval by two corners of the rectangle that contains it. The rectangles that I am
talking about here have sides that are vertical and horizontal, so that they can be specified by
just two points; this is the common meaning of “rectangle” in drawing programs.) We could
diagram these relationships as follows:
CHAPTER 1. THE MENTAL LANDSCAPE 13
DrawableObject
MultipointObject TwoPointObject
(If you would like to run this program, the source code, GUIDemo.java, is available on line.
For more information on using this and other examples from this textbook, see Section 2.6.)
Now, Java actually has three complete sets of GUI components. One of these, the AWT
or Abstract Windowing Toolkit, was available in the original version of Java. The second,
which is known as Swing , was introduced in Java version 1.2, and was the standard GUI toolkit
for many years. The third GUI toolkit, JavaFX, became a standard part of Java in Version 8
(but but has recently been removed, so that it requires separate installation in some versions
of Java). Although Swing, and even the AWT, can still be used, JavaFX is meant as a more
modern way to write GUI applications. This textbook covers JavaFX exclusively. (If you need
to learn Swing, you can take a look at the previous version of this book.)
When a user interacts with GUI components, “events” are generated. For example, clicking
a push button generates an event, and pressing a key on the keyboard generates an event.
Each time an event is generated, a message is sent to the program telling it that the event has
occurred, and the program responds according to its program. In fact, a typical GUI program
consists largely of “event handlers” that tell the program how to respond to various types of
events. In the above example, the program has been programmed to respond to each event by
displaying a message in the text area. In a more realistic example, the event handlers would
have more to do.
The use of the term “message” here is deliberate. Messages, as you saw in the previous sec-
tion, are sent to objects. In fact, Java GUI components are implemented as objects. Java
includes many predefined classes that represent various types of GUI components. Some of
CHAPTER 1. THE MENTAL LANDSCAPE 15
these classes are subclasses of others. Here is a diagram showing just a few of the JavaFX GUI
classes and their relationships:
Control
Don’t worry about the details for now, but try to get some feel about how object-oriented
programming and inheritance are used here. Note that all the GUI classes shown here are
subclasses, directly or indirectly, of a class called Control, which represents general properties
that are shared by many JavaFX components. In the diagram, two of the direct subclasses
of Control themselves have subclasses. The classes TextField and TextArea, which have certain
behaviors in common, are grouped together as subclasses of TextInputControl. Similarly But-
ton and CheckBox are subclasses of ButtonBase, which represents properties common to both
buttons and checkboxes. (ComboBox, by the way, is the class that represents pop-up menus.)
Just from this brief discussion, perhaps you can see how GUI programming can make effec-
tive use of object-oriented design. In fact, GUIs, with their “visible objects,” are probably a
major factor contributing to the popularity of OOP.
Programming with GUI components and events is one of the most interesting aspects of
Java. However, we will spend several chapters on the basics before returning to this topic in
Chapter 6.
use TCP/IP to send specific types of information such as web pages, electronic mail, and data
files.
All communication over the Internet is in the form of packets. A packet consists of some
data being sent from one computer to another, along with addressing information that indicates
where on the Internet that data is supposed to go. Think of a packet as an envelope with an
address on the outside and a message on the inside. (The message is the data.) The packet
also includes a “return address,” that is, the address of the sender. A packet can hold only
a limited amount of data; longer messages must be divided among several packets, which are
then sent individually over the Net and reassembled at their destination.
Every computer on the Internet has an IP address, a number that identifies it uniquely
among all the computers on the Net. (Actually, the claim about uniqueness is not quite true, but
the basic idea is valid, and the full truth is complicated.) The IP address is used for addressing
packets. A computer can only send data to another computer on the Internet if it knows that
computer’s IP address. Since people prefer to use names rather than numbers, most computers
are also identified by names, called domain names. For example, the main computer of
the Mathematics Department at Hobart and William Smith Colleges has the domain name
math.hws.edu. (Domain names are just for convenience; your computer still needs to know
IP addresses before it can communicate. There are computers on the Internet whose job it
is to translate domain names to IP addresses. When you use a domain name, your computer
sends a message to a domain name server to find out the corresponding IP address. Then, your
computer uses the IP address, rather than the domain name, to communicate with the other
computer.)
The Internet provides a number of services to the computers connected to it (and, of course,
to the users of those computers). These services use TCP/IP to send various types of data over
the Net. Among the most popular services are instant messaging, file sharing, electronic mail,
and the World-Wide Web. Each service has its own protocols, which are used to control
transmission of data over the network. Each service also has some sort of user interface, which
allows the user to view, send, and receive data through the service.
For example, the email service uses a protocol known as SMTP (Simple Mail Transfer
Protocol) to transfer email messages from one computer to another. Other protocols, such as
POP and IMAP, are used to fetch messages from an email account so that the recipient can
read them. A person who uses email, however, doesn’t need to understand or even know about
these protocols. Instead, they are used behind the scenes by computer programs to send and
receive email messages. These programs provide the user with an easy-to-use user interface to
the underlying network protocols.
The World-Wide Web is perhaps the most exciting of network services. The World-Wide
Web allows you to request pages of information that are stored on computers all over the
Internet. A Web page can contain links to other pages on the same computer from which
it was obtained or to other computers anywhere in the world. A computer that stores such
pages of information is called a web server . The user interface to the Web is the type of
program known as a web browser . Common web browsers include Microsoft Edge, Internet
Explorer, Firefox, Chrome, and Safari. You use a Web browser to request a page of information.
The browser sends a request for that page to the computer on which the page is stored, and
when a response is received from that computer, the web browser displays it to you in a neatly
formatted form. A web browser is just a user interface to the Web. Behind the scenes, the
web browser uses a protocol called HTTP (HyperText Transfer Protocol) to send each page
request and to receive the response from the web server.
CHAPTER 1. THE MENTAL LANDSCAPE 17
∗ ∗ ∗
Now just what, you might be thinking, does all this have to do with Java? In fact, Java
is intimately associated with the Internet and the World-Wide Web. When Java was first
introduced, one of its big attractions was the ability to write applets. An applet is a small
program that is transmitted over the Internet and that runs on a web page. Applets made it
possible for a web page to perform complex tasks and have complex interactions with the user.
Alas, applets have suffered from a variety of problems, and they have fallen out of use. There
are now other options for running programs on Web pages.
But applets were only one aspect of Java’s relationship with the Internet. Java can be
used to write complex, stand-alone applications that do not depend on a Web browser. Many
of these programs are network-related. For example many of the largest and most complex
web sites use web server software that is written in Java. Java includes excellent support for
network protocols, and its platform independence makes it possible to write network programs
that work on many different types of computer. You will learn about Java’s network support
in Chapter 11.
Its support for networking is not Java’s only advantage. But many good programming
languages have been invented only to be soon forgotten. Java has had the good luck to ride on
the coattails of the Internet’s immense and increasing popularity.
∗ ∗ ∗
As Java has matured, its applications have reached far beyond the Net. The standard
version of Java already comes with support for many technologies, such as cryptography, data
compression, sound processing, and three-dimensional graphics. And programmers have written
Java libraries to provide additional capabilities. Complex, high-performance systems can be
developed in Java. For example, Hadoop, a system for large scale data processing, is written in
Java. Hadoop is used by Yahoo, Facebook, and other Web sites to process the huge amounts
of data generated by their users.
Furthermore, Java is not restricted to use on traditional computers. Java can be used to
write programs for many smartphones (though not for the iPhone). It is the primary devel-
opment language for Android-based devices. (Android uses Google’s own version of Java and
does not use the same graphical user interface components as standard Java.) Java is also the
programming language for the Amazon Kindle eBook reader and for interactive features on
Blu-Ray video disks.
At this time, Java certainly ranks as one of the most widely used programming languages.
It is a good choice for almost any programming project that is meant to run on more than
one type of computing device, and is a reasonable choice even for many programs that will
run on only one device. It is probably still the most widely taught language at Colleges and
Universities. It is similar enough to other popular languages, such as C++, JavaScript, and
Python, that knowing it will give you a good start on learning those languages as well. Overall,
learning Java is a great starting point on the road to becoming an expert programmer. I hope
you enjoy the journey!
Quiz 18
Quiz on Chapter 1
(answers)
1. One of the components of a computer is its CPU. What is a CPU and what role does it
play in a computer?
5. If you have the source code for a Java program, and you want to run that program, you
will need both a compiler and an interpreter. What does the Java compiler do, and what
does the Java interpreter do?
6. What is a subroutine?
8. What is a variable? (There are four different ideas associated with variables in Java. Try
to mention all four aspects in your answer. Hint: One of the aspects is the variable’s
name.)
10. What is the “Internet”? Give some examples of how it is used. (What kind of services
does it provide?)
Chapter 2
On a basic level (the level of machine language), a computer can perform only very simple
operations. A computer performs complex tasks by stringing together large numbers of such
operations. Such tasks must be “scripted” in complete and perfect detail by programs. Creating
complex programs will never be really easy, but the difficulty can be handled to some extent by
giving the program a clear overall structure. The design of the overall structure of a program
is what I call “programming in the large.”
Programming in the small, which is sometimes called coding , would then refer to filling in
the details of that design. The details are the explicit, step-by-step instructions for performing
fairly small-scale tasks. When you do coding, you are working “close to the machine,” with some
of the same concepts that you might use in machine language: memory locations, arithmetic
operations, loops and branches. In a high-level language such as Java, you get to work with
these concepts on a level several steps above machine language. However, you still have to
worry about getting all the details exactly right.
This chapter and the next examine the facilities for programming in the small in the Java
programming language. Don’t be misled by the term “programming in the small” into thinking
that this material is easy or unimportant. This material is an essential foundation for all types
of programming. If you don’t understand it, you can’t write programs, no matter how good
you get at designing their large-scale structure.
The last section of this chapter discusses programming environments. That section
contains information about how to compile and run Java programs, and you should take a look
at it before trying to write and use your own programs or trying to use the sample programs
in this book.
19
CHAPTER 2. NAMES AND THINGS 20
using things like loops, branches, and subroutines. A syntactically correct program is one that
can be successfully compiled or interpreted; programs that have syntax errors will be rejected
(hopefully with a useful error message that will help you fix the problem).
So, to be a successful programmer, you have to develop a detailed knowledge of the syntax
of the programming language that you are using. However, syntax is only part of the story. It’s
not enough to write a program that will run—you want a program that will run and produce
the correct result! That is, the meaning of the program has to be right. The meaning of
a program is referred to as its semantics. More correctly, the semantics of a programming
language is the set of rules that determine the meaning of a program written in that language.
A semantically correct program is one that does what you want it to.
Furthermore, a program can be syntactically and semantically correct but still be a pretty
bad program. Using the language correctly is not the same as using it well. For example, a
good program has “style.” It is written in a way that will make it easy for people to read and
to understand. It follows conventions that will be familiar to other programmers. And it has
an overall design that will make sense to human readers. The computer is completely oblivious
to such things, but to a human reader, they are paramount. These aspects of programming are
sometimes referred to as pragmatics. (I will often use the more common term style.)
When I introduce a new language feature, I will explain the syntax, the semantics, and
some of the pragmatics of that feature. You should memorize the syntax; that’s the easy part.
Then you should get a feeling for the semantics by following the examples given, making sure
that you understand how they work, and, ideally, writing short programs of your own to test
your understanding. And you should try to appreciate and absorb the pragmatics—this means
learning how to use the language feature well, with style that will earn you the admiration of
other programmers.
Of course, even when you’ve become familiar with all the individual features of the language,
that doesn’t make you a programmer. You still have to learn how to construct complex programs
to solve particular problems. For that, you’ll need both experience and taste. You’ll find hints
about software development throughout this textbook.
∗ ∗ ∗
We begin our exploration of Java with the problem that has become traditional for such
beginnings: to write a program that displays the message “Hello World!”. This might seem like
a trivial problem, but getting a computer to do this is really a big first step in learning a new
programming language (especially if it’s your first programming language). It means that you
understand the basic process of:
1. getting the program text into the computer,
2. compiling the program, and
3. running the compiled program.
The first time through, each of these steps will probably take you a few tries to get right. I
won’t go into the details here of how you do each of these steps; it depends on the particular
computer and Java programming environment that you are using. See Section 2.6 for informa-
tion about creating and running Java programs in specific programming environments. But in
general, you will type the program using some sort of text editor and save the program in a file.
Then, you will use some command to try to compile the file. You’ll either get a message that the
program contains syntax errors, or you’ll get a compiled version of the program. In the case of
Java, the program is compiled into Java bytecode, not into machine language. Finally, you can
run the compiled program by giving some appropriate command. For Java, you will actually use
CHAPTER 2. NAMES AND THINGS 21
an interpreter to execute the Java bytecode. Your programming environment might automate
some of the steps for you—for example, the compilation step is often done automatically—but
you can be sure that the same three steps are being done in the background.
Here is a Java program to display the message “Hello World!”. Don’t expect to understand
what’s going on here just yet; some of it you won’t really understand until a few chapters from
now:
/** A program to display the message
* "Hello World!" on standard output.
*/
public class HelloWorld {
public static void main(String[] args) {
System.out.println("Hello World!");
}
} // end of class HelloWorld
The command that actually displays the message is:
System.out.println("Hello World!");
This command is an example of a subroutine call statement. It uses a “built-in subroutine”
named System.out.println to do the actual work. Recall that a subroutine consists of the
instructions for performing some task, chunked together and given a name. That name can be
used to “call” the subroutine whenever that task needs to be performed. A built-in subroutine
is one that is already defined as part of the language and therefore automatically available for
use in any program.
When you run this program, the message “Hello World!” (without the quotes) will be
displayed on standard output. Unfortunately, I can’t say exactly what that means! Java is
meant to run on many different platforms, and standard output will mean different things on
different platforms. However, you can expect the message to show up in some convenient or
inconvenient place. (If you use a command-line interface, like that in Oracle’s Java Development
Kit, you type in a command to tell the computer to run the program. The computer will type
the output from the program, Hello World!, on the next line. In an integrated development
environment such as Eclipse, the output might appear somewhere in one of the environment’s
windows.)
You must be curious about all the other stuff in the above program. Part of it consists of
comments. Comments in a program are entirely ignored by the computer; they are there for
human readers only. This doesn’t mean that they are unimportant. Programs are meant to be
read by people as well as by computers, and without comments, a program can be very difficult
to understand. Java has two types of comments. The first type begins with // and extends to
the end of a line. There is a comment of this form on the last line of the above program. The
computer ignores the // and everything that follows it on the same line. The second type of
comment starts with /* and ends with */, and it can extend over more than one line. The first
three lines of the program are an example of this second type of comment. (A comment that
actually begins with /**, like this one does, has special meaning; it is a “Javadoc” comment
that can be used to produce documentation for the program. See Subsection 4.6.5.)
Everything else in the program is required by the rules of Java syntax. All programming in
Java is done inside “classes.” The first line in the above program (not counting the comment)
says that this is a class named HelloWorld. “HelloWorld,” the name of the class, also serves as
CHAPTER 2. NAMES AND THINGS 22
the name of the program. Not every class is a program. In order to define a program, a class
must include a subroutine named main, with a definition that takes the form:
public static void main(String[] args) {
hstatements i
}
When you tell the Java interpreter to run the program, the interpreter calls this main()
subroutine, and the statements that it contains are executed. These statements make up the
script that tells the computer exactly what to do when the program is executed. The main()
routine can call other subroutines that are defined in the same class or even in other classes,
but it is the main() routine that determines how and in what order the other subroutines are
used.
The word “public” in the first line of main() means that this routine can be called from out-
side the program. This is essential because the main() routine is called by the Java interpreter,
which is something external to the program itself. The remainder of the first line of the routine
is harder to explain at the moment; for now, just think of it as part of the required syntax.
The definition of the subroutine—that is, the instructions that say what it does—consists of
the sequence of “statements” enclosed between braces, { and }. Here, I’ve used hstatementsi as
a placeholder for the actual statements that make up the program. Throughout this textbook,
I will always use a similar format: anything that you see in hthis style of texti (italic in angle
brackets) is a placeholder that describes something you need to type when you write an actual
program.
As noted above, a subroutine can’t exist by itself. It has to be part of a “class”. A program
is defined by a public class that takes the form:
hoptional-package-declaration i
hoptional-imports i
public class hprogram-name i {
hoptional-variable-declarations-and-subroutines i
public static void main(String[] args) {
hstatements i
}
hoptional-variable-declarations-and-subroutines i
}
The first two lines have to do with using packages. A package is a group of classes. You will
start learning about packages in Section 2.4, but our first few example programs will not use
them.
The hprogram-namei in the line that begins “public class” is the name of the program, as
well as the name of the class. (Remember, again, that hprogram-namei is a placeholder for the
actual name!) If the name of the class is HelloWorld, then the class must be saved in a file
called HelloWorld.java. When this file is compiled, another file named HelloWorld.class
will be produced. This class file, HelloWorld.class, contains the translation of the program
into Java bytecode, which can be executed by a Java interpreter. HelloWorld.java is called
the source code for the program. To execute the program, you only need the compiled class
file, not the source code.
The layout of the program on the page, such as the use of blank lines and indentation, is not
part of the syntax or semantics of the language. The computer doesn’t care about layout—you
CHAPTER 2. NAMES AND THINGS 23
could run the entire program together on one line as far as it is concerned. However, layout is
important to human readers, and there are certain style guidelines for layout that are followed
by most programmers.
Also note that according to the above syntax specification, a program can contain other
subroutines besides main(), as well as things called “variable declarations.” You’ll learn more
about these later, but not until Chapter 4.
2.2.1 Variables
Programs manipulate data that are stored in memory. In machine language, data can only be
referred to by giving the numerical address of the location in memory where the data is stored.
In a high-level language such as Java, names are used instead of numbers to refer to data. It
is the job of the computer to keep track of where in memory the data is actually stored; the
programmer only has to remember the name. A name used in this way—to refer to data stored
in memory—is called a variable.
Variables are actually rather subtle. Properly speaking, a variable is not a name for the
data itself but for a location in memory that can hold data. You should think of a variable as
a container or box where you can store data that you will need to use later. The variable refers
directly to the box and only indirectly to the data in the box. Since the data in the box can
change, a variable can refer to different data values at different times during the execution of
the program, but it always refers to the same box. Confusion can arise, especially for beginning
programmers, because when a variable is used in a program in certain ways, it refers to the
container, but when it is used in other ways, it refers to the data in the container. You’ll see
examples of both cases below.
In Java, the only way to get data into a variable—that is, into the box that the variable
names—is with an assignment statement. An assignment statement takes the form:
hvariable i = hexpression i;
where hexpressioni represents anything that refers to or computes a data value. When the
computer comes to an assignment statement in the course of executing a program, it evaluates
the expression and puts the resulting data value into the variable. For example, consider the
simple assignment statement
rate = 0.07;
The hvariablei in this assignment statement is rate, and the hexpressioni is the number 0.07.
The computer executes this assignment statement by putting the number 0.07 in the variable
rate, replacing whatever was there before. Now, consider the following more complicated
assignment statement, which might come later in the same program:
interest = rate * principal;
Here, the value of the expression “rate * principal” is being assigned to the variable
interest. In the expression, the * is a “multiplication operator” that tells the computer
to multiply rate times principal. The names rate and principal are themselves variables,
and it is really the values stored in those variables that are to be multiplied. We see that when
a variable is used in an expression, it is the value stored in the variable that matters; in this
case, the variable seems to refer to the data in the box, rather than to the box itself. When
the computer executes this assignment statement, it takes the value of rate, multiplies it by
the value of principal, and stores the answer in the box referred to by interest. When a
variable is used on the left-hand side of an assignment statement, it refers to the box that is
named by the variable.
(Note, by the way, that an assignment statement is a command that is executed by the
computer at a certain time. It is not a statement of fact. For example, suppose a program
includes the statement “rate = 0.07;”. If the statement “interest = rate * principal;”
is executed later in the program, can we say that the principal is multiplied by 0.07? No!
The value of rate might have been changed in the meantime by another statement. The
CHAPTER 2. NAMES AND THINGS 25
2.2.2 Types
A variable in Java is designed to hold only one particular type of data; it can legally hold that
type of data and no other. The compiler will consider it to be a syntax error if you try to
violate this rule by assigning a value of the wrong type to a variable. We say that Java is a
strongly typed language because it enforces this rule.
There are eight so-called primitive types built into Java. The primitive types are named
byte, short, int, long, float, double, char, and boolean. The first four types hold integers
(whole numbers such as 17, -38477, and 0). The four integer types are distinguished by the
ranges of integers they can hold. The float and double types hold real numbers (such as 3.6 and
-145.99). Again, the two real types are distinguished by their range and accuracy. A variable
of type char holds a single character from the Unicode character set. And a variable of type
boolean holds one of the two logical values true or false.
Any data value stored in the computer’s memory must be represented as a binary number,
that is as a string of zeros and ones. A single zero or one is called a bit. A string of eight
bits is called a byte. Memory is usually measured in terms of bytes. Not surprisingly, the byte
data type refers to a single byte of memory. A variable of type byte holds a string of eight
bits, which can represent any of the integers between -128 and 127, inclusive. (There are 256
integers in that range; eight bits can represent 256—two raised to the power eight—different
values.) As for the other integer types,
• short corresponds to two bytes (16 bits). Variables of type short have values in the range
-32768 to 32767.
• int corresponds to four bytes (32 bits). Variables of type int have values in the range
-2147483648 to 2147483647.
• long corresponds to eight bytes (64 bits). Variables of type long have values in the range
-9223372036854775808 to 9223372036854775807.
You don’t have to remember these numbers, but they do give you some idea of the size of
integers that you can work with. Usually, for representing integer data you should just stick to
the int data type, which is good enough for most purposes.
The float data type is represented in four bytes of memory, using a standard method for
encoding real numbers. The maximum value for a float is about 10 raised to the power 38.
A float can have about 7 significant digits. (So that 32.3989231134 and 32.3989234399 would
both have to be rounded off to about 32.398923 in order to be stored in a variable of type
float.) A double takes up 8 bytes, can range up to about 10 to the power 308, and has about
15 significant digits. Ordinarily, you should stick to the double type for real values.
A variable of type char occupies two bytes in memory. The value of a char variable is a
single character such as A, *, x, or a space character. The value can also be a special character
such a tab or a carriage return or one of the many Unicode characters that come from different
languages. Values of type char are closely related to integer values, since a character is actually
stored as a 16-bit integer code number. In fact, we will see that chars in Java can actually be
used like integers in certain situations.
It is important to remember that a primitive type value is represented using only a certain,
finite number of bits. So, an int can’t be an arbitrary integer; it can only be an integer
CHAPTER 2. NAMES AND THINGS 26
in a certain finite range of values. Similarly, float and double variables can only take on
certain values. They are not true real numbers in the mathematical sense. For example, the
mathematical constant π can only be approximated by a value of type float or double, since
it would require an infinite number of decimal places to represent it exactly. For that matter,
many simple numbers such as 1/3 can only be approximated by floats and doubles.
2.2.3 Literals
A data value is stored in the computer as a sequence of bits. In the computer’s memory, it
doesn’t look anything like a value written on this page. You need a way to include constant
values in the programs that you write. In a program, you represent constant values as literals.
A literal is something that you can type in a program to represent a value. It is a kind of name
for a constant value.
For example, to type a value of type char in a program, you must surround it with a pair
of single quote marks, such as ’A’, ’*’, or ’x’. The character and the quote marks make up a
literal of type char. Without the quotes, A would be an identifier and * would be a multiplication
operator. The quotes are not part of the value and are not stored in the variable; they are just
a convention for naming a particular character constant in a program. If you want to store the
character A in a variable ch of type char, you could do so with the assignment statement
ch = ’A’;
Certain special characters have special literals that use a backslash, \, as an “escape character.”
In particular, a tab is represented as ’\t’, a carriage return as ’\r’, a linefeed as ’\n’, the
single quote character as ’\’’, and the backslash itself as ’\\’. Note that even though you
type two characters between the quotes in ’\t’, the value represented by this literal is a single
tab character.
Numeric literals are a little more complicated than you might expect. Of course, there
are the obvious literals such as 317 and 17.42. But there are other possibilities for expressing
numbers in a Java program. First of all, real numbers can be represented in an exponential
form such as 1.3e12 or 12.3737e-108. The “e12” and “e-108” represent powers of 10, so that
1.3e12 means 1.3 times 1012 and 12.3737e-108 means 12.3737 times 10−108 . This format can be
used to express very large and very small numbers. Any numeric literal that contains a decimal
point or exponential is a literal of type double. To make a literal of type float, you have to
append an “F” or “f” to the end of the number. For example, “1.2F” stands for 1.2 considered
as a value of type float. (Occasionally, you need to know this because the rules of Java say that
you can’t assign a value of type double to a variable of type float, so you might be confronted
with a ridiculous-seeming error message if you try to do something like “x = 1.2;” if x is a
variable of type float. You have to say “x = 1.2F;". This is one reason why I advise sticking
to type double for real numbers.)
Even for integer literals, there are some complications. Ordinary integers such as 177777
and -32 are literals of type byte, short, or int, depending on their size. You can make a literal
of type long by adding “L” as a suffix. For example: 17L or 728476874368L. As another
complication, Java allows binary, octal (base-8), and hexadecimal (base-16) literals. I don’t
want to cover number bases in detail, but in case you run into them in other people’s programs,
it’s worth knowing a few things: Octal numbers use only the digits 0 through 7. In Java, a
numeric literal that begins with a 0 is interpreted as an octal number; for example, the octal
literal 045 represents the number 37, not the number 45. Octal numbers are rarely used, but
you need to be aware of what happens when you start a number with a zero. Hexadecimal
CHAPTER 2. NAMES AND THINGS 27
numbers use 16 digits, the usual digits 0 through 9 and the letters A, B, C, D, E, and F. Upper
case and lower case letters can be used interchangeably in this context. The letters represent
the numbers 10 through 15. In Java, a hexadecimal literal begins with 0x or 0X, as in 0x45
or 0xFF7A. Finally, binary literals start with 0b or 0B and contain only the digits 0 and 1; for
example: 0b10110.
As a final complication, numeric literals can include the underscore character (“ ”), which
can be used to separate groups of digits. For example, the integer constant for two billion could
be written 2 000 000 000, which is a good deal easier to decipher than 2000000000. There is
no rule about how many digits have to be in each group. Underscores can be especially useful
in long binary numbers; for example, 0b1010 1100 1011.
I will note that hexadecimal numbers can also be used in character literals to represent
arbitrary Unicode characters. A Unicode literal consists of \u followed by four hexadecimal
digits. For example, the character literal ’\u00E9’ represents the Unicode character that is an
“e” with an acute accent.
For the type boolean, there are precisely two literals: true and false. These literals are
typed just as I’ve written them here, without quotes, but they represent values, not variables.
Boolean values occur most often as the values of conditional expressions. For example,
rate > 0.05
is a boolean-valued expression that evaluates to true if the value of the variable rate is greater
than 0.05, and to false if the value of rate is less than or equal to 0.05. As you’ll see in
Chapter 3, boolean-valued expressions are used extensively in control structures. Of course,
boolean values can also be assigned to variables of type boolean. For example, if test is a
variable of type boolean, then both of the following assignment statements are legal:
test = true;
test = rate > 0.05;
What, after all, would the computer do with the value computed by the function in this case?
You have to tell the computer to do something with the value. You might tell the computer to
display it:
System.out.print( Math.sqrt(x) ); // Display the square root of x.
or you might use an assignment statement to tell the computer to store that value in a variable:
lengthOfSide = Math.sqrt(x);
The function call Math.sqrt(x) represents a value of type double, and it can be used anyplace
where a numeric literal of type double could be used. The x in this formula represents the
parameter to the subroutine; it could be a variable named “x”, or it could be replaced by any
expression that represents a numerical value. For example, Math.sqrt(2) computes the square
root of 2, and Math.sqrt(a*a+b*b) would be legal as long as a and b are numeric variables.
The Math class contains many static member functions. Here is a list of some of the more
important of them:
• Math.abs(x), which computes the absolute value of x.
• The usual trigonometric functions, Math.sin(x), Math.cos(x), and Math.tan(x). (For
all the trigonometric functions, angles are measured in radians, not degrees.)
• The inverse trigonometric functions arcsin, arccos, and arctan, which are written as:
Math.asin(x), Math.acos(x), and Math.atan(x). The return value is expressed in radi-
ans, not degrees.
• The exponential function Math.exp(x) for computing the number e raised to the power
x, and the natural logarithm function Math.log(x) for computing the logarithm of x in
the base e.
• Math.pow(x,y) for computing x raised to the power y.
• Math.floor(x), which rounds x down to the nearest integer value that is less than or
equal to x. Even though the return value is mathematically an integer, it is returned
as a value of type double, rather than of type int as you might expect. For example,
Math.floor(3.76) is 3.0, and Math.floor(-4.2) is -5. The function Math.round(x)
returns the integer that is closest to x, and Math.ceil(x) rounds x up to an integer.
(“Ceil” is short for “ceiling”, the opposite of “floor.”)
• Math.random(), which returns a randomly chosen double in the range 0.0 <=
Math.random() < 1.0. (The computer actually calculates so-called “pseudorandom”
numbers, which are not truly random but are effectively random enough for most pur-
poses.) We will find a lot of uses for Math.random in future examples.
For these functions, the type of the parameter—the x or y inside the parentheses—can be
any value of any numeric type. For most of the functions, the value returned by the function
is of type double no matter what the type of the parameter. However, for Math.abs(x), the
value returned will be the same type as x; if x is of type int, then so is Math.abs(x). So, for
example, while Math.sqrt(9) is the double value 3.0, Math.abs(9) is the int value 9.
Note that Math.random() does not have any parameter. You still need the parentheses,
even though there’s nothing between them. The parentheses let the computer know that this is
a subroutine rather than a variable. Another example of a subroutine that has no parameters
is the function System.currentTimeMillis(), from the System class. When this function is
executed, it retrieves the current time, expressed as the number of milliseconds that have passed
CHAPTER 2. NAMES AND THINGS 32
since a standardized base time (the start of the year 1970, if you care). One millisecond is one-
thousandth of a second. The return value of System.currentTimeMillis() is of type long (a
64-bit integer). This function can be used to measure the time that it takes the computer to
perform a task. Just record the time at which the task is begun and the time at which it is
finished and take the difference. For more accurate timing, you can use System.nanoTime()
instead. System.nanoTime() returns the number of nanoseconds since some arbitrary starting
time, where one nanosecond is one-billionth of a second. However, you should not expect the
time to be truly accurate to the nanosecond.
Here is a sample program that performs a few mathematical tasks and reports the time that
it takes for the program to run.
/**
* This program performs some mathematical computations and displays the
* results. It also displays the value of the constant Math.PI. It then
* reports the number of seconds that the computer spent on this task.
*/
public class TimedComputation {
public static void main(String[] args) {
long startTime; // Starting time of program, in nanoseconds.
long endTime; // Time when computations are done, in nanoseconds.
long compTime; // Run time in nanoseconds.
double seconds; // Time difference, in seconds.
startTime = System.nanoTime();
double width, height, hypotenuse; // sides of a triangle
width = 42.0;
height = 17.0;
hypotenuse = Math.sqrt( width*width + height*height );
System.out.print("A triangle with sides 42 and 17 has hypotenuse ");
System.out.println(hypotenuse);
System.out.println("\nMathematically, sin(x)*sin(x) + "
+ "cos(x)*cos(x) - 1 should be 0.");
System.out.println("Let’s check this for x = 100:");
System.out.print(" sin(100)*sin(100) + cos(100)*cos(100) - 1 is: ");
System.out.println( Math.sin(100)*Math.sin(100)
+ Math.cos(100)*Math.cos(100) - 1 );
System.out.println("(There can be round-off errors when"
+ " computing with real numbers!)");
System.out.print("\nHere is a random number: ");
System.out.println( Math.random() );
System.out.print("\nThe value of Math.PI is ");
System.out.println( Math.PI );
endTime = System.nanoTime();
compTime = endTime - startTime;
seconds = compTime / 1000000000.0;
System.out.print("\nRun time in nanoseconds was: ");
System.out.println(compTime);
System.out.println("(This is probably not perfectly accurate!");
System.out.print("\nRun time in seconds was: ");
CHAPTER 2. NAMES AND THINGS 33
System.out.println(seconds);
} // end main()
} // end class TimedComputation
prefer to use the more general term, “subroutine.” However, I should note that some people
prefer to use the term “method” from the beginning.
can also use s1.indexOf(ch) to search for a char, ch, in s1. To find the first occurrence
of x at or after position N, you can use s1.indexOf(x,N). To find the last occurrence of
x in s1, use s1.lastIndexOf(x).
• s1.compareTo(s2) is an integer-valued function that compares the two strings. If the
strings are equal, the value returned is zero. If s1 is less than s2, the value returned is
a number less than zero, and if s1 is greater than s2, the value returned is some number
greater than zero. There is also a function s1.compareToIgnoreCase(s2). (If both of
the strings consist entirely of lower case letters, or if they consist entirely of upper case
letters, then “less than” and “greater than” refer to alphabetical order. Otherwise, the
ordering is more complicated.)
• s1.toUpperCase() is a String -valued function that returns a new string that is equal to s1,
except that any lower case letters in s1 have been converted to upper case. For example,
"Cat".toUpperCase() is the string "CAT". There is also a function s1.toLowerCase().
• s1.trim() is a String -valued function that returns a new string that is equal to s1 except
that any non-printing characters such as spaces and tabs have been trimmed from the
beginning and from the end of the string. Thus, if s1 has the value "fred ", then
s1.trim() is the string "fred", with the spaces at the end removed.
For the functions s1.toUpperCase(), s1.toLowerCase(), and s1.trim(), note that the
value of s1 is not changed. Instead a new string is created and returned as the value of
the function. The returned value could be used, for example, in an assignment statement
such as “smallLetters = s1.toLowerCase();”. To change the value of s1, you could use an
assignment “s1 = s1.toLowerCase();”.
∗ ∗ ∗
Here is another extremely useful fact about strings: You can use the plus operator, +, to
concatenate two strings. The concatenation of two strings is a new string consisting of all the
characters of the first string followed by all the characters of the second string. For example,
"Hello" + "World" evaluates to "HelloWorld". (Gotta watch those spaces, of course—if you
want a space in the concatenated string, it has to be somewhere in the input data, as in
"Hello " + "World".)
Let’s suppose that name is a variable of type String and that it already refers to the name
of the person using the program. Then, the program could greet the user by executing the
statement:
System.out.println("Hello, " + name + ". Pleased to meet you!");
Even more surprising is that you can actually concatenate values of any type onto a String
using the + operator. The value is converted to a string, just as it would be if you printed it to
the standard output, and then that string is concatenated with the other string. For example,
the expression "Number" + 42 evaluates to the string "Number42". And the statements
System.out.print("After ");
System.out.print(years);
System.out.print(" years, the value is ");
System.out.print(principal);
can be replaced by the single statement:
System.out.print("After " + years +
" years, the value is " + principal);
CHAPTER 2. NAMES AND THINGS 36
Obviously, this is very convenient. It would have shortened some of the examples presented
earlier in this chapter.
You can print out an enum value with an output statement such as System.out.print(vacation).
The output value will be the name of the enum constant (without the “Season.”). In this case,
the output would be “SUMMER”.
Because an enum is technically a class, the enum values are technically objects. As ob-
jects, they can contain subroutines. One of the subroutines in every enum value is named
ordinal(). When used with an enum value, it returns the ordinal number of the value in
the list of values of the enum. The ordinal number simply tells the position of the value in
the list. That is, Season.SPRING.ordinal() is the int value 0, Season.SUMMER.ordinal() is
1, Season.FALL.ordinal() is 2, and Season.WINTER.ordinal() is 3. (You will see over and
over again that computer scientists like to start counting at zero!) You can, of course, use the
ordinal() method with a variable of type Season, such as vacation.ordinal().
Using enums can make a program more readable, since you can use meaningful names for
the values. And it can prevent certain types of errors, since a compiler can check that the values
assigned to an enum variable are in fact legal values for that variable. For now, you should just
appreciate them as the first example of an important concept: creating new types. Here is a
little example that shows enums being used in a complete program:
public class EnumDemo {
// Define two enum types -- remember that the definitions
// go OUTSIDE the main() routine!
enum Day { SUNDAY, MONDAY, TUESDAY, WEDNESDAY, THURSDAY, FRIDAY, SATURDAY }
enum Month { JAN, FEB, MAR, APR, MAY, JUN, JUL, AUG, SEP, OCT, NOV, DEC }
public static void main(String[] args) {
Day tgif; // Declare a variable of type Day.
Month libra; // Declare a variable of type Month.
tgif = Day.FRIDAY; // Assign a value of type Day to tgif.
libra = Month.OCT; // Assign a value of type Month to libra.
System.out.print("My sign is libra, since I was born in ");
System.out.println(libra); // Output value will be: OCT
System.out.print("That’s the ");
System.out.print( libra.ordinal() );
System.out.println("-th month of the year.");
System.out.println(" (Counting from 0, of course!)");
System.out.print("Isn’t it nice to get to ");
System.out.println(tgif); // Output value will be: FRIDAY
System.out.println( tgif + " is the " + tgif.ordinal()
+ "-th day of the week.");
}
}
(As I mentioned, an enum can actually be defined in a separate file. The sample program
SeparateEnumDemo.java is identical to EnumDemo.java, except that the enum types that it
uses are defined in files named Month.java and Day.java.)
CHAPTER 2. NAMES AND THINGS 38
minimum number of spaces that should be used for the output. If the integer that is being
output takes up fewer than 12 spaces, extra blank spaces are added in front of the integer to
bring the total up to 12. We say that the output is “right-justified in a field of length 12.” A
very large value is not forced into 12 spaces; if the value has more than 12 digits, all the digits
will be printed, with no extra spaces. The specifier %d means the same as %1d—that is, an
integer will be printed using just as many spaces as necessary. (The “d,” by the way, stands
for “decimal”—that is, base-10—numbers. You can replace the “d” with an “x” to output an
integer value in hexadecimal form.)
The letter “s” at the end of a format specifier can be used with any type of value. It
means that the value should be output in its default format, just as it would be in unformatted
output. A number, such as the “20” in %20s, can be added to specify the (minimum) number
of characters. The “s” stands for “string,” and it can be used for values of type String. It can
also be used for values of other types; in that case the value is converted into a String value in
the usual way.
The format specifiers for values of type double are more complicated. An “f”, as in %1.2f,
is used to output a number in “floating-point” form, that is with digits after a decimal point. In
%1.2f, the “2” specifies the number of digits to use after the decimal point. The “1” specifies
the (minimum) number of characters to output; a “1” in this position effectively means that
just as many characters as are necessary should be used. Similarly, %12.3f would specify a
floating-point format with 3 digits after the decimal point, right-justified in a field of length 12.
Very large and very small numbers should be written in exponential format, such as
6.00221415e23, representing “6.00221415 times 10 raised to the power 23.” A format speci-
fier such as %15.8e specifies an output in exponential form, with the “8” telling how many
digits to use after the decimal point. If you use “g” instead of “e”, the output will be in ex-
ponential form for very small values and very large values and in floating-point form for other
values. In %1.8g, the 8 gives the total number of digits in the answer, including both the digits
before the decimal point and the digits after the decimal point.
For numeric output, the format specifier can include a comma (“,”), which will cause the
digits of the number to be separated into groups, to make it easier to read big numbers. In
the United States, groups of three digits are separated by commas. For example, if x is one
billion, then System.out.printf("%,d",x) will output 1,000,000,000. In other countries, the
separator character and the number of digits per group might be different. The comma should
come at the beginning of the format specifier, before the field width; for example: %,12.3f.
If you want the output to be left-justified instead of right justified, add a minus sign to the
beginning of the format specifier: for example, %-20s.
In addition to format specifiers, the format string in a printf statement can include other
characters. These extra characters are just copied to the output. This can be a convenient way
to insert values into the middle of an output string. For example, if x and y are variables of
type int, you could say
System.out.printf("The product of %d and %d is %d", x, y, x*y);
When this statement is executed, the value of x is substituted for the first %d in the string, the
value of y for the second %d, and the value of the expression x*y for the third, so the output
would be something like “The product of 17 and 42 is 714” (quotation marks not included in
output!).
To output a percent sign, use the format specifier %% in the format string. You can use %n
to output a line feed. You can also use a backslash, \, as usual in strings to output special
characters such as tabs and double quote characters.
CHAPTER 2. NAMES AND THINGS 40
that the user typed will then be returned by the function, and it will be stored in the variable,
userInput. Here is a complete program that uses TextIO.getlnInt to read a number typed
by the user and then prints out the square of that number. Note the import directive on the
first line:
import textio.TextIO;
/**
* A program that reads an integer that is typed in by the
* user and computes and prints the square of that integer.
*/
public class PrintSquare {
public static void main(String[] args) {
int userInput; // The number input by the user.
int square; // The userInput, multiplied by itself.
System.out.print("Please type a number: ");
userInput = TextIO.getlnInt();
square = userInput * userInput;
System.out.println();
System.out.println("The number that you entered was " + userInput);
System.out.println("The square of that number is " + square);
System.out.println();
} // end of main()
} //end of class PrintSquare
When you run this program, it will display the message “Please type a number:” and will pause
until you type a response, including a carriage return after the number. Note that it is good
style to output a question or some other prompt to the user before reading input. Otherwise,
the user will have no way of knowing exactly what the computer is waiting for, or even that it
is waiting for the user to do something.
When you call one of these functions, you are guaranteed that it will return a legal value of
the correct type. If the user types in an illegal value as input—for example, if you ask for an
int and the user types in a non-numeric character or a number that is outside the legal range
of values that can be stored in a variable of type int—then the computer will ask the user to
re-enter the value, and your program never sees the first, illegal value that the user entered. For
TextIO.getlnBoolean(), the user is allowed to type in any of the following: true, false, t, f, yes,
no, y, n, 1, or 0. Furthermore, they can use either upper or lower case letters. In any case, the
user’s input is interpreted as a true/false value. It’s convenient to use TextIO.getlnBoolean()
to read the user’s response to a Yes/No question.
You’ll notice that there are two input functions that return Strings. The first, getlnWord(),
returns a string consisting of non-blank characters only. When it is called, it skips over any
spaces and carriage returns typed in by the user. Then it reads non-blank characters until
it gets to the next space or carriage return. It returns a String consisting of all the non-
blank characters that it has read. The second input function, getln(), simply returns a string
consisting of all the characters typed in by the user, including spaces, up to the next carriage
return. It gets an entire line of input text. The carriage return itself is not returned as part of
the input string, but it is read and discarded by the computer. Note that the String returned
by TextIO.getln() might be the empty string , "", which contains no characters at all. You
will get this return value if the user simply presses return, without typing anything else first.
TextIO.getln() does not skip blanks or end-of-lines before reading a value. But the
input functions getlnInt(), getlnDouble(), getlnBoolean(), and getlnChar() behave like
getlnWord() in that they will skip past any blanks and carriage returns in the input before
reading a value. When one of these functions skips over an end-of-line, it outputs a ’ ?’ to let
the user know that more input is expected.
Furthermore, if the user types extra characters on the line after the input value, all the
extra characters will be discarded, along with the carriage return at the end of the
line. If the program executes another input function, the user will have to type in another line
of input, even if they had typed more than one value on the previous line. It might not sound
like a good idea to discard any of the user’s input, but it turns out to be the safest thing to do
in most programs.
∗ ∗ ∗
Using TextIO for input and output, we can now improve the program from Section 2.2 for
computing the value of an investment. We can have the user type in the initial value of the
investment and the interest rate. The result is a much more useful program—for one thing, it
makes sense to run it more than once! Note that this program uses formatted output to print
out monetary values in their correct format.
import textio.TextIO;
/**
* This class implements a simple program that will compute
* the amount of interest that is earned on an investment over
* a period of one year. The initial amount of the investment
* and the interest rate are input by the user. The value of
* the investment at the end of the year is output. The
* rate must be input as a decimal, not a percentage (for
* example, 0.05 rather than 5).
*/
public class Interest2 {
CHAPTER 2. NAMES AND THINGS 43
When you call TextIO.writeFile, TextIO remembers the file and automatically sends any
output from TextIO.put or other output functions to that file. If you want to go back to
writing to standard output, you can call
TextIO.writeStandardOutput();
Here is a simple program that asks the user some questions and outputs the user’s responses to
a file named “profile.txt.” As an example, it uses TextIO for output to standard output as well
as to the file, but System.out could also have been used for the output to standard output.
import textio.TextIO;
public class CreateProfile {
public static void main(String[] args) {
String name; // The user’s name.
String email; // The user’s email address.
double salary; // the user’s yearly salary.
String favColor; // The user’s favorite color.
TextIO.putln("Good Afternoon! This program will create");
TextIO.putln("your profile file, if you will just answer");
TextIO.putln("a few simple questions.");
TextIO.putln();
/* Gather responses from the user. */
TextIO.put("What is your name? ");
name = TextIO.getln();
TextIO.put("What is your email address? ");
email = TextIO.getln();
TextIO.put("What is your yearly income? ");
salary = TextIO.getlnDouble();
TextIO.put("What is your favorite color? ");
favColor = TextIO.getln();
/* Write the user’s information to the file named profile.txt. */
TextIO.writeFile("profile.txt"); // subsequent output goes to file
TextIO.putln("Name: " + name);
TextIO.putln("Email: " + email);
TextIO.putln("Favorite Color: " + favColor);
TextIO.putf( "Yearly Income: %,1.2f%n", salary);
/* Print a final message to standard output. */
TextIO.writeStandardOutput();
TextIO.putln("Thank you. Your profile has been written to profile.txt.");
}
}
In many cases, you want to let the user select the file that will be used for output. You
could ask the user to type in the file name, but that is error-prone, and users are more familiar
with selecting a file from a file dialog box. The statement
TextIO.writeUserSelectedFile();
CHAPTER 2. NAMES AND THINGS 45
will open a typical graphical-user-interface file selection dialog where the user can specify the
output file. This also has the advantage of alerting the user if they are about to replace
an existing file. It is possible for the user to cancel the dialog box without selecting a file.
TextIO.writeUserSelectedFile is a function that returns a boolean value. The return value
is true if the user selected a file, and is false if the user canceled the dialog box. Your program
can check the return value if it needs to know whether it is actually going to write to a file or
not.
∗ ∗ ∗
TextIO can also read from files, as an alternative to reading from standard input. You can
specify an input source for TextIO’s various “get” functions. The default input source is standard
input. You can use the statement TextIO.readFile("data.txt") to read from a file named
“data.txt” instead, or you can let the user select the input file with a GUI-style dialog box by
saying TextIO.readUserSelectedFile(). After you have done this, any input will come from
the file instead of being typed by the user. You can go back to reading the user’s input with
TextIO.readStandardInput().
When your program is reading from standard input, the user gets a chance to correct any
errors in the input. This is not possible when the program is reading from a file. If illegal data
is found when a program tries to read from a file, an error occurs that will crash the program.
(Later, we will see that it is possible to “catch” such errors and recover from them.) Errors can
also occur, though more rarely, when writing to files.
A complete understanding of input/output in Java requires a knowledge of object oriented
programming. We will return to the topic later, in Chapter 11. The file I/O capabilities in
the TextIO class are rather primitive by comparison. Nevertheless, they are sufficient for many
applications, and they will allow you to get some experience with files sooner rather than later.
on the line has been read or discarded (by one of the “getln” functions). The user only gets to
type when the buffer is empty.
Note, by the way, that although the TextIO input functions will skip past blank spaces and
carriage returns while looking for input, they will not skip past other characters. For example,
if you try to read two ints and the user types “42,17”, the computer will read the first number
correctly, but when it tries to read the second number, it will see the comma. It will regard this
as an error and will force the user to retype the number. If you want to input several numbers
from one line, you should make sure that the user knows to separate them with spaces, not
commas. Alternatively, if you want to require a comma between the numbers, use getChar()
to read the comma before reading the second number.
There is another character input function, TextIO.getAnyChar(), which does not skip past
blanks or carriage returns. It simply reads and returns the next character typed by the user,
even if it’s a blank or carriage return. If the user typed a carriage return, then the char returned
by getAnyChar() is the special linefeed character ’\n’. There is also a function, TextIO.peek(),
that lets you look ahead at the next character in the input without actually reading it. After
you “peek” at the next character, it will still be there when you read the next item from input.
This allows you to look ahead and see what’s coming up in the input, so that you can take
different actions depending on what’s there.
The TextIO class provides a number of other functions. To learn more about them, you can
look at the comments in the source code file, TextIO.java.
Clearly, the semantics of input is much more complicated than the semantics of output!
Fortunately, for the majority of applications, it’s pretty straightforward in practice. You only
need to follow the details if you want to do something fancy. In particular, I strongly advise
you to use the “getln” versions of the input routines, rather than the “get” versions, unless you
really want to read several items from the same line of input, precisely because the semantics
of the “getln” versions is much simpler.
same as TextIO.getInt() except for two things: If the value entered by the user is not a legal
int, then stdin.nextInt() will crash rather than prompt the user to re-enter the value. And
the integer entered by the user must be followed by a blank space or by an end-of-line, whereas
TextIO.getInt() will stop reading at any character that is not a digit.
There are corresponding methods for reading other types of data, including
stdin.nextDouble(), stdin.nextLong(), and stdin.nextBoolean(). (stdin.nextBoolean()
will only accept “true” or “false” as input.) These subroutines can read more than one value
from a line, so they are more similar to the “get” versions of TextIO subroutines rather than
the “getln” versions. The method stdin.nextLine() is equivalent to TextIO.getln(), and
stdin.next(), like TextIO.getWord(), returns a string of non-blank characters.
As a simple example, here is a version of the sample program Interest2.java that uses Scanner
instead of TextIO for user input:
import java.util.Scanner;
public class Interest2WithScanner {
public static void main(String[] args) {
Scanner stdin = new Scanner( System.in ); // Create the Scanner.
double principal; // The value of the investment.
double rate; // The annual interest rate.
double interest; // The interest earned during the year.
System.out.print("Enter the initial investment: ");
principal = stdin.nextDouble();
System.out.print("Enter the annual interest rate (as a decimal): ");
rate = stdin.nextDouble();
interest = principal * rate; // Compute this year’s interest.
principal = principal + interest; // Add it to principal.
System.out.printf("The amount of interest is $%1.2f%n", interest);
System.out.printf("The value after one year is $%1.2f%n", principal);
} // end of main()
} // end of class Interest2WithScanner
Note the inclusion of the two lines given above to import Scanner and create stdin.
Also note the substitution of stdin.nextDouble() for TextIO.getlnDouble(). (In fact,
stdin.nextDouble() is really equivalent to TextIO.getDouble() rather than to the “getln”
version, but this will not affect the behavior of the program as long as the user types just one
number on each line of input.)
I will continue to use TextIO for input for the time being, but I will give a few more examples
of using Scanner in the on-line solutions to the end-of-chapter exercises. There will be more
detailed coverage of Scanner later in the book.
When two numerical values are combined (after doing type conversion on one of them, if
necessary), the answer will be of the same type. If you multiply two ints, you get an int; if you
multiply two doubles, you get a double. This is what you would expect, but you have to be
very careful when you use the division operator /. When you divide two integers, the answer
will always be an integer; if the quotient has a fractional part, it is discarded. For example, the
value of 7/2 is 3, not 3.5. If N is an integer variable, then N/100 is an integer, and 1/N is equal
to zero for any N greater than one! This fact is a common source of programming errors. You
can force the computer to compute a real number as the answer by making one of the operands
real: For example, when the computer evaluates 1.0/N, it first converts N to a real number in
order to match the type of 1.0, so you get a real number as the answer.
Java also has an operator for computing the remainder when one number is divided by
another. This operator is indicated by %. If A and B are integers, then A % B represents the
remainder when A is divided by B. (However, for negative operands, % is not quite the same as
the usual mathematical “modulus” operator, since if one of A or B is negative, then the value
of A % B will be negative.) For example, 7 % 2 is 1, while 34577 % 100 is 77, and 50 % 8 is
2. A common use of % is to test whether a given integer is even or odd: N is even if N % 2 is
zero, and it is odd if N % 2 is 1. More generally, you can check whether an integer N is evenly
divisible by an integer M by checking whether N % M is zero.
The % operator also works with real numbers. In general, A % B is what is left over after
you remove as many copies of B as possible from A. For example, 7.52 % 0.5 is 0.02.
Finally, you might need the unary minus operator, which takes the negative of a number.
For example, -X has the same value as (-1)*X. For completeness, Java also has a unary plus
operator, as in +X, even though it doesn’t really do anything.
By the way, recall that the + operator can also be used to concatenate a value of any type
onto a String. When you use + to combine a string with a value of some other type, it is another
example of type conversion, since any type can be automatically converted into type String.
Usually, the operators ++ or -- are used in statements like “x++;” or “x--;”. These state-
ments are commands to change the value of x. However, it is also legal to use x++, ++x, x--,
or --x as expressions, or as parts of larger expressions. That is, you can write things like:
y = x++;
y = ++x;
TextIO.putln(--x);
z = (++x) * (y--);
The statement “y = x++;” has the effects of adding 1 to the value of x and, in addition, assigning
some value to y. The value assigned to y is defined to be the old value of x, before the 1 is
added. Thus, if the value of x is 6, the statement “y = x++;” will change the value of x to 7,
but it will change the value of y to 6, because the value assigned to y is the old value of x. On
the other hand, the value of ++x is defined to be the new value of x, after the 1 is added. So if
x is 6, then the statement “y = ++x;” changes the values of both x and y to 7. The decrement
operator, --, works in a similar way.
Note in particular that the statement x = x++; does not change the value of x! This is
because the value that is being assigned to x is the old value of x, the one that it had before the
statement was executed. The net result is that x is incremented but then immediately changed
back to its previous value! You also need to remember that x++ is not the same as x + 1. The
expression x++ changes the value of x; the expression x + 1 does not.
This can be confusing, and I have seen many bugs in student programs resulting from the
confusion. My advice is: Don’t be confused. Use ++ and -- only as stand-alone statements,
not as expressions. I will follow this advice in almost all examples in these notes.
boolean sameSign;
sameSign = ((x > 0) == (y > 0));
One thing that you cannot do with the relational operators <, >, <=, and >= is to use them
to compare values of type String. You can legally use == and != to compare Strings, but
because of peculiarities in the way objects behave, they might not give the results you want.
(The == operator checks whether two objects are stored in the same memory location, rather
than whether they contain the same value. Occasionally, for some objects, you do want to
make such a check—but rarely for strings. I’ll get back to this in a later chapter.) Instead,
you should compare strings using subroutines such as equals() and compareTo(), which were
described in Subsection 2.3.3.
Another place where == and != don’t work as you would expect is with Double.NaN, the
constant that represents an undefined value of type double. The values of x == Double.NaN and
x != Double.NaN are both defined to be false in all cases, whether or not x is Double.NaN! To
test whether a real value x is the undefined value Double.NaN, use the boolean-valued function
Double.isNaN(x).
operator—that is, it has three operands—and it comes in two pieces, ? and :, that have to be
used together. It takes the form
hboolean-expression i ? hexpression1 i : hexpression2 i
The computer tests the value of hboolean-expressioni. If the value is true, it evaluates
hexpression1 i; otherwise, it evaluates hexpression2 i. For example:
next = (N % 2 == 0) ? (N/2) : (3*N+1);
will assign the value N/2 to next if N is even (that is, if N % 2 == 0 is true), and it will assign
the value (3*N+1) to next if N is odd. (The parentheses in this example are not required, but
they do make the expression easier to read.)
You can do type casts from any numeric type to any other numeric type. However, you should
note that you might change the numeric value of a number by type-casting it. For example,
(short)100000 is -31072. (The -31072 is obtained by taking the 4-byte int 100000 and throwing
away two of those bytes to obtain a short—you’ve lost the real information that was in those
two bytes.)
When you type-cast a real number to an integer, the fractional part is discarded. For
example, (int)7.9453 is 7. As another example of type casts, consider the problem of get-
ting a random integer between 1 and 6. The function Math.random() gives a real number
between 0.0 and 0.9999. . . , and so 6*Math.random() is between 0.0 and 5.999. . . . The type-
cast operator, (int), can be used to convert this to an integer: (int)(6*Math.random()).
Thus, (int)(6*Math.random()) is one of the integers 0, 1, 2, 3, 4, and 5. To get a number
between 1 and 6, we can add 1: “(int)(6*Math.random()) + 1”. (The parentheses around
6*Math.random() are necessary because of precedence rules; without the parentheses, the type
cast operator would apply only to the 6.)
The type char is almost an integer type. You can assign char values to int variables, and you
can assign integer constants in the range 0 to 65535 to char variables. You can also use explicit
type-casts between char and the numeric types. For example, (char)97 is ’a’, (int)’+’ is
43, and (char)(’A’ + 2) is ’C’.
∗ ∗ ∗
Type conversion between String and other types cannot be done with type-casts. One way to
convert a value of any type into a string is to concatenate it with an empty string. For example,
"" + 42 is the string "42". But a better way is to use the function String.valueOf(x), a static
member function in the String class. String.valueOf(x) returns the value of x, converted into
a string. For example, String.valueOf(42) is the string "42", and if ch is a char variable,
then String.valueOf(ch) is a string of length one containing the single character that is the
value of ch.
It is also possible to convert certain strings into values of other types. For example, the
string "10" should be convertible into the int value 10, and the string "17.42e-2" into the
double value 0.1742. In Java, these conversions are handled by built-in functions.
The standard class Integer contains a static member function for converting from String to
int. In particular, if str is any expression of type String, then Integer.parseInt(str) is a
function call that attempts to convert the value of str into a value of type int. For example, the
value of Integer.parseInt("10") is the int value 10. If the parameter to Integer.parseInt
does not represent a legal int value, then an error occurs.
Similarly, the standard class Double includes a function Double.parseDouble. If str is a
String, then the function call Double.parseDouble(str) tries to convert str into a value of
type double. An error occurs if str does not represent a legal double value.
∗ ∗ ∗
Getting back to assignment statements, Java has several variations on the assignment
operator, which exist to save typing. For example, “A += B” is defined to be the same as
“A = A + B”. Every operator in Java that applies to two operands, except for the relational
operators, gives rise to a similar assignment operator. For example:
x -= y; // same as: x = x - y;
x *= y; // same as: x = x * y;
x /= y; // same as: x = x / y;
x %= y; // same as: x = x % y;
q &&= p; // same as: q = q && p; (for booleans q and p)
CHAPTER 2. NAMES AND THINGS 54
The combined assignment operator += even works with strings. Recall that when the + operator
is used with a string as one of the operands, it represents concatenation. Since str += x is
equivalent to str = str + x, when += is used with a string on the left-hand side, it appends
the value on the right-hand side onto the string. For example, if str has the value “tire”, then
the statement str += ’d’; changes the value of str to “tired”.
This textbook covers Java 8, but as of the release of Version 8.1.1 of the book in May
2020, the current version of Java is Java 14, with version 15 due to be released in September
2020. Significant changes both in the Java language and in the Oracle corporation’s policies
concerning Java have been made since the release of Java 8. Unfortunately, the changes have
made it somewhat more difficult to install Java and get it up and running. In this section, I
will try to give you enough information to make it possible to install Java and use it with this
textbook.
One of the changes made to Java was to the release schedule. New versions of Java will
be released much more frequently than in the past, about twice a year, but only some of the
releases are “long-term support” (LTS) releases that will continue to receive bug fixes and
security updates over an extended period of time. Java 8 and Java 11 are long-term support
releases, but Java 9, 10, 12, 13, and 14 are not. In fact, only Java 8, 11, and 14 are currently
officially supported (although for the purposes of this textbook, using another version that is
no longer receiving support is not a real problem).
You should get the JavaFX “SDK”—not the “jmods”—for your operating system (Linux,
Mac OS, or Windows). You should get the version number that matches the version number
of the OpenJDK that you downloaded.
If you downloaded a JDK installer for Windows or Mac OS from Oracle or AdoptOpenJDK,
you can just double-click the installer file to start the installation, if it does not start automat-
ically. Otherwise, when you download OpenJDK or the JavaFX SDK, it will be in the form of
a compressed archive file. You will need to extract the contents of the archive. Usually, simply
double-clicking the icon of the archive file will either extract the contents or open a program
that you can use to extract the contents. You will get a directory with a name something like
jdk-11.0.7 or or javafx-sdk-11.0.2. You can put the directories anywhere on your computer, but
you will need to know where they are located.
Note that if you are using Linux, then a recent OpenJDK and JavaFX should be available in
the software repositories for your version of Linux, and all that you need to do is use your Linux
software installer tool to install them. (Exactly what that means depends on the version of
Linux that you are using. Search for packages with names containing “openjdk” and “openjfx”.)
You will still need to know where the JavaFX SDK directory is located; on my computer, it’s
/usr/share/openjfx.
And if you are using Mac OS and downloaded an OpenJDK as an archive file, then
when you extract the OpenJDK archive, you will get a directory that is named something
like jdk-11.0.7.jdk. It is advisable to move that JDK directory into the directory /Li-
brary/Java/JavaVirtualMachines/. (This will require administrative privileges on your Mac.)
Mac OS should automatically use the most recent JDK in /Library/Java/JavaVirtualMachines/
as the default version of Java.
As for Windows, the usual location for JDKs is in C:\Program Files\Java, so you might
consider moving the OpenJDK directory there.
command dir (on Windows) or ls (on Linux and Mac OS). When the window first opens,
the current directory is your home directory , where your personal files are stored. You can
change the current directory using the cd command with the name of the directory that you
want to use. For example, if the current directory is your home directory, then you can change
into your Desktop directory by typing the command cd Desktop (and then pressing return).
You might want to create a directory (that is, a folder) to hold your Java work. For example,
you might create a directory named javawork in your home directory. You can do this using
your computer’s GUI; another way is to use the command line: Open a command window. If
you want to put your work directory in a different folder from your home directory, cd into
the directory where you want to put it. Then enter the command mkdir javawork to make
the directory. When you want to work on programming, open a command window and use the
cd command to change into your Java work directory. Of course, you can have more than one
working directory for your Java work; you can organize your files any way you like.
∗ ∗ ∗
The most basic commands for using Java on the command line are javac and java. The
javac command is used to compile Java source code, and java is used to run Java programs.
These commands, and other commands for working with Java, can be found in a directory
named bin inside the JDK directory. If you set things up correctly on your computer, it should
recognize these commands when you type them on the command line. Try typing the commands
java -version and javac -version. The output from these commands should tell you which
version of Java is being used. If you get a message such as “Command not found,” then Java
is not correctly configured.
Java should already be configured correctly on Linux, if you have installed Java from
the Linux software repositories. The same is true on Mac OS, if you have used an in-
staller from AdoptOpenJDK or if you have manually moved the OpenJDK directory into
/Library/Java/JavaVirtualMachines/ — on Mac OS, javac and java will use the most re-
cent JDK in that directory. For Windows, an AdoptOpenJDK installer should set things up
correctly by default. If you use Oracle JDK or a JDK from a compressed archive file, you
need to add the JDK bin directory to the system path by hand. The installation instructions
for the Oracle JDK for Windows have information about how to do that; see the “Installation
Instructions” link on the following page, and look for “Setting the Path Environment Variable”
in the instructions for installing the JDK on Windows:
https://www.oracle.com/technetwork/java/javase/downloads/index.html
∗ ∗ ∗
To test the javac command, place a copy of HelloWorld.java into your working directory.
(If you downloaded the Web site of this book, you can find it in the directory named source;
you can use your computer’s GUI to copy-and-paste this file into your working directory. Al-
ternatively, you can navigate to HelloWorld.java on the book’s Web site and use the “Save As”
command in your Web browser to save a copy of the file into your working directory.) Type
the command:
javac HelloWorld.java
This will compile HelloWorld.java and will create a bytecode file named HelloWorld.class
in the same directory. Note that if the command succeeds, you will not get any response from
the computer; it will just redisplay the command prompt to tell you it’s ready for another
command. You will then be able to run the program using the java command:
java HelloWorld
CHAPTER 2. NAMES AND THINGS 58
The computer should respond by outputting the message “Hello World!”. Note that although
the program is stored in a file named HelloWorld.class, the java command uses the name of
the class, HelloWorld, not the name of the file.
Many of the sample programs for this book use TextIO to read input from the user (see
Subsection 2.4.3). Since TextIO is not a standard part of Java, you must make it available to any
program that uses it. This means that your working directory should contain a folder named
textio, and inside that folder should be the file TextIO.java. You can copy TextIO.java from
this book’s source directory, or your can download it from the web site, but you should be sure
to place it inside a folder named textio in the same directory as the program that uses TextIO.
Once you have TextIO.java you can run a sample program such as Interest2.java to test
user input. First, compile the program with the command
javac Interest2.java
If successful, this will create the compiled file named Interest2.class. But you will also notice
that it creates the file TextIO.class inside the textio folder, if that file does not already exist.
More generally, the javac command will compile not just the file that you specify but also any
additional Java files that are needed. Once you have Interest2.class, you can run it using
the command
java Interest2
You will be asked to enter some information, and you will respond by typing your answers into
the command window, pressing return at the end of each line. When the program ends, you
will see the command prompt, and you can enter another command. (Note, by the way, that
“java TextIO” would not make sense, since TextIO does not have a main() routine, and so it
is not possible to execute it as a program.)
You can follow a similar procedure to run all of the examples in this book that do not use
JavaFX. For running JavaFX programs, see the next subsection below.
∗ ∗ ∗
To create your own programs, you will need a text editor . A text editor is a computer
program that allows you to create and save documents that contain plain text. It is important
that the documents be saved as plain text, that is without any special encoding or formatting
information. Word processor documents are not appropriate, unless you can get your word
processor to save as plain text. A good text editor can make programming a lot more pleasant.
Linux comes with several text editors. On Windows, you can use notepad in a pinch, but you
will probably want something better. For Mac OS, you might download the BBEdit application,
which can be used for free. One possibility that will work on any platform is to use jedit, a
programmer’s text editor that is itself written in Java and that can be downloaded for free from
www.jedit.org. Another popular cross-platform programming editor is Atom, available from
atom.io.
To work on your programs, you can open a command line window and cd into the working
directory where you will store your source code files. Start up your text editor program, such
as by double-clicking its icon or selecting it from a Start menu. Type your code into the editor
window, or open an existing source code file that you want to modify. Save the file into your
working directory. Remember that the name of a Java source code file must end in “.java”, and
the rest of the file name must match the name of the class that is defined in the file. Once the
file is saved in your working directory, go to the command window and use the javac command
to compile it, as discussed above. If there are syntax errors in the code, they will be listed
in the command window. Each error message contains the line number in the file where the
CHAPTER 2. NAMES AND THINGS 59
computer found the error. Go back to the editor and try to fix one or more errors, save your
changes, and then try the javac command again. (It’s usually a good idea to just work on
the first few errors; sometimes fixing those will make other errors go away.) Remember that
when the javac command finally succeeds, you will get no message at all, or possibly just some
“warnings” that do not stop the program from running. Then you can use the java command
to run your program, as described above. Once you’ve compiled the program, you can run it
as many times as you like without recompiling it.
That’s really all there is to it: Keep both editor and command-line window open. Edit,
save, and compile until you have eliminated all the syntax errors. (Always remember to save
the file before compiling it—the compiler only sees the saved file, not the version in the editor
window.) When you run the program, you might find that it has semantic errors that cause it
to run incorrectly. In that case, you have to go back to the edit/save/compile loop to try to
find and fix the problem.
ranging from fairly simple wrappers around the JDK to highly complex applications with a
multitude of features. For a beginning programmer, there is a danger in using an IDE, since
the difficulty of learning to use the IDE, on top of the difficulty of learning to program, can be
daunting. However, for my own programming, I generally use the Eclipse IDE, and I introduce
my students to it after they have had some experience with the command line. I will discuss
Eclipse in some detail and a much simpler alternative, BlueJ, more briefly. IDEs have features
that are very useful even for a beginning programmer, although a beginner will want to ignore
many of their advanced features.
Unless you are using Oracle’s JDK for Java 8, 9, or 10, using Eclipse for JavaFX programs
will require some extra configuration. The next subsection discusses using JavaFX in Eclipse.
This subsection tells you how to use it for programs that use only standard Java classes.
You can download an Eclipse IDE from eclipse.org. When I install Eclipse, I get the “Eclipse
IDE for Java Developers” package from this web page:
https://www.eclipse.org/downloads/packages/
For Windows and Linux, the download is a compressed archive file. You can simply extract
the contents of the archive and place the resulting directory wherever you want it on your
computer. You will find the Eclipse application in that directory, and you can start Eclipse by
double-clicking the application icon. For Mac OS, the download is a .dmg file that contains the
Eclipse application. You can open the .dmg file and drag the application to any location that
you prefer (probably the Applications folder).
Eclipse is a free program. It is itself written in Java. It requires a JDK for Java 8 or later.
Before trying to run Eclipse, you should first install a JDK. (Note for Mac OS users: As of
June 2019, a bug in Eclipse stops it from running on JDK 12. Until that is fixed, you can stick
to JDK 11 as your default version of Java, if you want to use Eclipse.)
The first time you start Eclipse, you will be asked to specify a workspace, which is the
directory where your work will be stored. You can accept the default name, or provide one
of your own. You can use multiple workspaces and select the one that you want to use at
startup. When a new workspace is first opened, the Eclipse window will be filled by a large
“Welcome” screen that includes links to extensive documentation and tutorials. You should
close this screen, by clicking the “X” next to the word “Welcome”; you can get back to it later
by choosing “Welcome” from the “Help” menu.
The Eclipse GUI consists of one large window that is divided into several sections. Each
section contains one or more views. For example, a view can be a text editor, it can be a place
where a program can do I/O, or it can contain a list of your projects. If there are several views
in one section of the window, then there will be tabs at the top of the section to select the view
that is displayed in that section. Each view displays a different type of information. The whole
set of views is called a perspective. Eclipse uses different perspectives, that is, different sets of
views of different types of information, for different tasks. For compiling and running programs,
the only perspective that you will need is the “Java Perspective,” which is the default. As you
become more experienced, you might want to use the “Debug Perspective,” which has features
designed to help you find semantic errors in programs. There are small buttons in the Eclipse
toolbar that can be used to switch between perspectives.
The Java Perspective includes a large area in the center of the window that contains text
editor views. This is where you will create and edit your programs. To the left of this is the
Package Explorer view, which will contain a list of your Java projects and source code files. To
the right are some other views that I don’t find very useful; I suggest that you close them by
clicking the small “X” next to the name of each one. Several other views that will be useful
CHAPTER 2. NAMES AND THINGS 62
appear in a section of the window below the editing area. If you accidently close one of the
important views, such as the Package Explorer, you can get it back by selecting it from the
“Show View” submenu of the “Window” menu. You can also reset the whole window to its
default contents by selecting “Reset Perspective” from the “Window” menu.
∗ ∗ ∗
To do any work in Eclipse, you need a project. To start a Java project, go to the “New”
submenu in the “File” menu, and select the “Java Project” command. In the window that pops
up, it is only necessary to fill in a “Project Name” for the project and click the “Finish” button.
(In some circumstances, it might be necessary to select the JDK that you want to use for the
project.) The project name can be anything you like. Eclipse might ask you if you want to
create “module-info.java” — you should click “Don’t Create”; the programs in this book
do not use modules. (See Subsection 4.6.4 for information about modules.) The project should
appear in the “Package Explorer” view. Click on the small triangle or plus sign next to the
project name to see the contents of the project. Assuming that you use the default settings,
there should be a directory named “src,” which is where your Java source code files will go. It
also contains the “JRE System Library”; this is the collection of standard built-in classes that
come with Java.
To run any of the sample Java programs from this textbook, you need to copy the source
code file into your Eclipse Java project. Assuming that you have downloaded the source code
file onto your computer, you can copy-and-paste it into the Eclipse window. (Right-click the
file icon (or control-click on Mac OS); select “Copy” from the pop-up menu; then right-click
the project’s src folder in the Eclipse window, and select “Paste”. Be sure to paste it into the
src folder, not into the project itself; files outside the src folder are not treated as Java source
code files.) Alternatively, you can try dragging the file icon from a file browser window onto
the src folder in the Eclipse window.
To use the TextIO based examples from this textbook, you must add the source code file
TextIO.java to your project. This file has to be in a “package” named textio. If you have
downloaded TextIO.java and placed it into a folder named “textio,” as described above, then
you can simply copy-and-paste the textio folder into the “src” folder of your project. Alterna-
tively, you can create the textio package using the “New/Package” command from the “File”
menu. This will make a folder named “textio” in your project, and you can then copy-and-paste
TextIO.java into that folder. In any case, package textio should appear under “src” in your
project, with TextIO.java inside it. (You can drag files from one location to another in the
Package Explorer view, if you accidently put a file in the wrong location.)
Once a Java program is in the project, you can open it in an editor by double-clicking the
file name in the “Package Explorer” view. To run the program, right-click in the editor window,
or on the file name in the Package Explorer view (or control-click in Mac OS). In the menu that
pops up, go to the “Run As” submenu, and select “Java Application”. The program will be
executed. If the program writes to standard output, the output will appear in the “Console”
view, in the area of the Eclipse window below the editing area. If the program uses TextIO
or Scanner for input, you will have to type the required input into the “Console” view—click
the “Console” view before you start typing so that the characters that you type will be
sent to the correct part of the window. (For an easier way to run a program, find and click
the small “Run” button in Eclipse’s tool bar. This will run either the program in the editor
window, the program selected in the Package Explorer view, or the program that was run most
recently, depending on context.) Note that when you run a program in Eclipse, it is compiled
automatically. There is no separate compilation step.
CHAPTER 2. NAMES AND THINGS 63
You can have more than one program in the same Eclipse project, or you can create addi-
tional projects to organize your work better. Remember to place a copy of TextIO.java, inside
a folder named textio, in any project that requires it.
∗ ∗ ∗
To create a new Java program in Eclipse, you must create a new Java class. To do that,
right-click the Java project name in the “Project Explorer” view. Go to the “New” submenu
of the popup menu, and select “Class”. (Alternatively, there is a small icon in the toolbar at
the top of the Eclipse window that you can click to create a new Java class.) In the window
that opens, type in the name of the class that you want to create. The class name must be
a legal Java identifier. Note that you want the name of the class, not the name of the source
code file, so don’t add “.java” at the end of the name. The window also includes an input box
labeled “Package” where you can specify the name of a package to contain the class. Most
examples in this book use the “default package,” but you can create your own programs in any
package. To use the default package, the “Package” input box should be empty. Finally, click
the “Finish” button to create the class. The class should appear inside the “src” folder, in a
folder corresponding to its package. The new file should automatically open in the editing area
so that you can start typing in your program.
Eclipse has several features that aid you as you type your code. It will underline any syntax
error with a jagged red line, and in some cases will place an error marker in the left border
of the edit window. If you hover the mouse cursor over the error marker or over the error
itself, a description of the error will appear. Note that you do not have to get rid of every
error immediately as you type; some errors will go away as you type in more of the program.
If an error marker displays a small “light bulb,” Eclipse is offering to try to fix the error for
you. Click the light bulb—or simply hover your mouse over the actual error—to get a list of
possible fixes, then click the fix that you want to apply. For example, if you use an undeclared
variable in your program, Eclipse will offer to declare it for you. You can actually use this
error-correcting feature to get Eclipse to write certain types of code for you! Unfortunately,
you’ll find that you won’t understand a lot of the proposed fixes until you learn more about
the Java language, and it is not a good idea to apply a fix that you don’t understand—often
that will just make things worse in the end.
Eclipse will also look for spelling errors in comments and will underline them with jagged
red lines. Hover your mouse over the error to get a list of possible correct spellings.
Another essential Eclipse feature is content assist. Content assist can be invoked by typing
Control-Space. It will offer possible completions of whatever you are typing at the moment. For
example, if you type part of an identifier and hit Control-Space, you will get a list of identifiers
that start with the characters that you have typed; use the up and down arrow keys to select one
of the items in the list, and press Return or Enter. (You can also click an item with the mouse
to select it, or hit Escape to dismiss the list.) If there is only one possible completion when
you hit Control-Space, it will be inserted automatically. By default, Content Assist will also
pop up automatically, after a short delay, when you type a period or certain other characters.
For example, if you type “TextIO.” and pause for just a fraction of a second, you will get a
list of all the subroutines in the TextIO class. Personally, I find this auto-activation annoying.
You can disable it in the Eclipse Preferences. (Look under Java / Editor / Content Assist, and
turn off the “Enable auto activation” option.) You can still call up Code Assist manually with
Control-Space.
Once you have an error-free program, you can run it as described above. If you find a
problem when you run it, it’s very easy to go back to the editor, make changes, and run it
CHAPTER 2. NAMES AND THINGS 64
again.
If you are adding a new JRE, you need to specify the “JRE home.” This will be the directory
that contains the bin directory of the JDK that you want to use. Click the “Directory” button
to open a file browser, and navigate to that directory. This can be a JDK that you have simply
downloaded and have not installed officially on your computer, or it can be an installed JDK.
In the example, the JDK was officially installed on Linux. (On Windows or Linux, you want
CHAPTER 2. NAMES AND THINGS 65
the main JDK directory. For Mac OS, you need to navigate to the main JDK directory, then
go into the “Content” folder and finally into the “Home” folder; the JRE home will end with
/Content/Home.) If you enter a valid JDK directory, Eclipse will fill in a JRE name and system
library.
Remember that to use JavaFX, you need to make it available to your program both at
compile time and at run time. To make it available at compile time, you can add JavaFX
.jar files to the system libraries. Click the “Add External JARs” button, and navigate to the
lib directory in the JavaFX SDK that you downloaded or installed. You should see the seven
JavaFX .jar files. Select them all, and click “OK”. They should appear in the “JRE system
libraries” list. Optionally, make sure that all of the JavaFX jar files are selected in that list, and
click the “Source Attachment” button. In the “Source Attachment Configuration” dialog box,
select “External Location”, and click “External File”. Select the file src.zip from the JavaFX
lib directory, and click “OK”. This optional step will allow Eclipse to find the documentation
for JavaFX, which can be very useful when you are writing programs.
To make JavaFX available to your programs at run time, you need to fill in the “Default
VM arguments” box. You should type in the arguments by hand. This box contains options
that will be added to the java command every time you run a program in Eclipse using this
JDK. You should type the same options that you would use on the java command line, as
discussed above. The value of the -p (or --module-path) option is the location of the JavaFX
lib directory, and the value for --add-modules can be ALL-MODULE-PATH, as shown.
When everything is set up, click “Finish” and then click “Apply and Close” in the main
“Preferences” dialog box. When you create a new Java project in the workspace, make sure
that the JRE that you have added or edited is selected in the project creation dialog box. Then
you should be able to use JavaFX in that project. If not, check your JRE configuration.
It is, by the way, harmless to use a JRE that is configured for JavaFX even for standard
programs that do not use JavaFX. In particular, you can use the same configuration for all the
programs in this textbook.
2.6.6 BlueJ
Finally, I will mention BlueJ, a small IDE that is designed specifically for people who are
learning to program. It is much less complex than Eclipse, but it does have some features that
make it useful for education. BlueJ can be downloaded from bluej.org. The current version
requires at least Java 11 and JavaFX 11. You need to have both before you can run BlueJ. (An
earlier version that works with Java 8 is also available.) The first time you start BlueJ, it will
ask you to enter the locations of the JDK and of JavaFX. You will only have to do that once,
and once it is done, you will not have to take any special actions to compile and run JavaFX
programs in the IDE.
In BlueJ, you can begin a project with the “New Project” command in the “Project” menu.
A BlueJ project is simply a folder. When you create a project, you will have to select a folder
name that does not already exist. The folder will be created and a window will be opened to
show the contents of the folder. Files are shown as icons in the BlueJ window. You can drag
.java files from the file system onto that window to add files to the project; they will be copied
into the project folder as well as shown in the window. You can also copy files directly into
the project folder, but BlueJ won’t see them until the next time you open the project. When
you restart BlueJ, it should show the last project you were working on, but you can open any
project with a command from the “Project” menu.
CHAPTER 2. NAMES AND THINGS 66
There is a button in the project window for creating a new class. An icon for the class is
added to the window, and a .java source code file is created in the project folder. The file is not
automatically opened for editing. To edit a file, double-click its icon in the project window. An
editor will be opened in a separate window. (A newly created class will contain some default
code that you probably don’t want; you can erase it and add a main() routine instead.) The
BlueJ editor does not show errors as you type. Errors will be reported when you compile the
program. Also, it does not offer automatic fixes for errors. It has a less capable version of
Eclipse’s Content Assist, which seems only to work for getting a list of available subroutines in
a class or object; call up the list by hitting Control-Space after typing the period following the
name of a class or object.
An editor window contains a button for compiling the program in the window. There is
also a compile button in the project window, which compiles all the classes in the project.
To run a program, it must already be compiled. Right-click the icon of a compiled program.
In the menu that pops up, you will see “void main(String[] args)”. Select that option from
the menu to run the program. Just click “OK” in the dialog box that pops up. A separate
window will open for input/output.
One of the neatest features of BlueJ is that you can actually use it to run any subroutine,
not just main. If a class contains other subroutines, you will see them in the list that you get
by right-clicking its icon. A pop-up dialog allows you to enter any parameters required by the
routine, and if the routine is a function, you will get another dialog box after the routine has
been executed to tell you its return value. This allows easy testing of individual subroutines.
Furthermore, you can also use BlueJ to create new objects from a class. An icon for the object
will be added at the bottom of the project window, and you can right-click that icon to get
a list of subroutines in the object. This will, of course, not be useful to you until we get to
object-oriented programming in Chapter 5.
that are also in the default package. (In fact, in earlier versions of this textbook, TextIO was
in the default package. I have moved it to package textio for Version 8 of the book.)
When packages are used in a command-line environment, some complications arise. For
example, if a program is in a package named text.pkg, then the source code file must be in
a subdirectory named “pkg” inside a directory named “test” that is in turn inside your main
Java working directory. Nevertheless, when you compile or execute the program, you should
be working in the main directory, not in the subdirectory. When you compile the source
code file, you have to include the name of the directory in the command: For example, for a
program in package test.pkg use “javac test/pkg/ClassName.java” on Linux or Mac OS, or
“javac test\pkg\ClassName.java” on Windows. The command for executing the program is
then “java test.pkg.ClassName”, with a period separating the package name from the class
name.
Using jshell can be a great way to learn Java and to experiment with its features. I won’t
give any more detailed information about it in this book, but you can learn more at
https://docs.oracle.com/en/java/javase/11/jshell/introduction-jshell.html
Exercises 68
1. Write a program that will print your initials to standard output in letters that are nine (solution)
lines tall. Each big letter should be made up of a bunch of *’s. For example, if your initials
were “DJE”, then the output would look something like:
****** ************* **********
** ** ** **
** ** ** **
** ** ** **
** ** ** ********
** ** ** ** **
** ** ** ** **
** ** ** ** **
***** **** **********
2. Write a program that simulates rolling a pair of dice. You can simulate rolling one die by (solution)
choosing one of the integers 1, 2, 3, 4, 5, or 6 at random. The number you pick represents
the number on the die after it is rolled. As pointed out in Section 2.5, the expression
(int)(Math.random()*6) + 1
does the computation to select a random integer between 1 and 6. You can assign this
value to a variable to represent one of the dice that are being rolled. Do this twice and
add the results together to get the total roll. Your program should report the number
showing on each die as well as the total roll. For example:
The first die comes up 3
The second die comes up 5
Your total roll is 8
3. Write a program that asks the user’s name, and then greets the user by name. Before (solution)
outputting the user’s name, convert it to upper case letters. For example, if the user’s
name is Fred, then the program should respond “Hello, FRED, nice to meet you!”.
4. Write a program that helps the user count his change. The program should ask how many (solution)
quarters the user has, then how many dimes, then how many nickels, then how many
pennies. Then the program should tell the user how much money he has, expressed in
dollars.
5. If you have N eggs, then you have N/12 dozen eggs, with N%12 eggs left over. (This is (solution)
essentially the definition of the / and % operators for integers.) Write a program that asks
the user how many eggs she has and then tells the user how many dozen eggs she has and
how many extra eggs are left over.
A gross of eggs is equal to 144 eggs. Extend your program so that it will tell the user
how many gross, how many dozen, and how many left over eggs she has. For example, if
the user says that she has 1342 eggs, then your program would respond with
Your number of eggs is 9 gross, 3 dozen, and 10
Exercises 69
6. This exercise asks you to write a program that tests some of the built-in subroutines for (solution)
working with Strings. The program should ask the user to enter their first name and their
last name, separated by a space. Read the user’s response using TextIO.getln(). Break
the input string up into two strings, one containing the first name and one containing the
last name. You can do that by using the indexOf() subroutine to find the position of the
space, and then using substring() to extract each of the two names. Also output the
number of characters in each name, and output the user’s initials. (The initials are the
first letter of the first name together with the first letter of the last name.) A sample run
of the program should look something like this:
Please enter your first name and last name, separated by a space.
? Mary Smith
Your first name is Mary, which has 4 characters
Your last name is Smith, which has 5 characters
Your initials are MS
7. Suppose that a file named “testdata.txt” contains the following information: The first (solution)
line of the file is the name of a student. Each of the next three lines contains an integer.
The integers are the student’s scores on three exams. Write a program that will read
the information in the file and display (on standard output) a message that contains the
name of the student and the student’s average grade on the three exams. The average is
obtained by adding up the individual exam grades and then dividing by the number of
exams.
Quiz 70
Quiz on Chapter 2
(answers)
1. Briefly explain what is meant by the syntax and the semantics of a programming language.
Give an example to illustrate the difference between a syntax error and a semantics error.
2. What does the computer do when it executes a variable declaration statement. Give an
example.
4. One of the primitive types in Java is boolean. What is the boolean type? Where are
boolean values used? What are its possible values?
6. Explain what is meant by an assignment statement, and give an example. What are
assignment statements used for?
8. What is a literal ?
9. In Java, classes have two fundamentally different purposes. What are they?
10. What is the difference between the statement “x = TextIO.getDouble();” and the state-
ment “x = TextIO.getlnDouble();”
11. Explain why the value of the expression 2 + 3 + "test" is the string "5test" while the
value of the expression "test" + 2 + 3 is the string "test23". What is the value of
"test" + 2 * 3 ?
12. Integrated Development Environments such as Eclipse often use syntax coloring , which
assigns various colors to the characters in a program to reflect the syntax of the language.
A student notices that Eclipse colors the word String differently from int, double, and
boolean. The student asks why String should be a different color, since all these words
are names of types. What’s the answer to the student’s question?
13. What is the purpose of an import directive, such as import textio.TextIO or import
java.util.Scanner?
14. Write a complete program that asks the user to enter the number of “widgets” they want
to buy and the cost per widget. The program should then output the total cost for all
the widgets. Use System.out.printf to print the cost, with two digits after the decimal
point. You do not need to include any comments in the program.
Chapter 3
3.1.1 Blocks
The block is the simplest type of structured statement. Its purpose is simply to group a
sequence of statements into a single statement. The format of a block is:
71
CHAPTER 3. CONTROL 72
{
hstatements i
}
That is, it consists of a sequence of statements enclosed between a pair of braces, “{” and “}”.
In fact, it is possible for a block to contain no statements at all; such a block is called an empty
block , and can actually be useful at times. An empty block consists of nothing but an empty
pair of braces. Block statements usually occur inside other statements, where their purpose is
to group together several statements into a unit. However, a block can be legally used wherever
a statement can occur. There is one place where a block is required: As you might have already
noticed in the case of the main subroutine of a program, the definition of a subroutine is a
block, since it is a sequence of statements enclosed inside a pair of braces.
I should probably note again at this point that Java is what is called a free-format language.
There are no syntax rules about how the language has to be arranged on a page. So, for example,
you could write an entire block on one line if you want. But as a matter of good programming
style, you should lay out your program on the page in a way that will make its structure as
clear as possible. In general, this means putting one statement per line and using indentation
to indicate statements that are contained inside control structures. This is the format that I
will use in my examples.
Here are two examples of blocks:
{
System.out.print("The answer is ");
System.out.println(ans);
}
In this section, I’ll introduce the while loop and the if statement. I’ll give the full details of
these statements and of the other three control structures in later sections.
A while loop is used to repeat a given statement over and over. Of course, it’s not likely
that you would want to keep repeating it forever. That would be an infinite loop, which is
generally a bad thing. (There is an old story about computer pioneer Grace Murray Hopper,
who read instructions on a bottle of shampoo telling her to “lather, rinse, repeat.” As the
story goes, she claims that she tried to follow the directions, but she ran out of shampoo. (In
case you don’t get it, she was making a joke about the way that computers mindlessly follow
instructions.))
To be more specific, a while loop will repeat a statement over and over, but only so long
as a specified condition remains true. A while loop has the form:
while (hboolean-expression i)
hstatement i
Since the statement can be, and usually is, a block, most while loops have the form:
while (hboolean-expression i) {
hstatements i
}
Some programmers think that the braces should always be included as a matter of style, even
when there is only one statement between them, but I don’t always follow that advice myself.
The semantics of the while statement go like this: When the computer comes to a while
statement, it evaluates the hboolean-expressioni, which yields either true or false as its value.
If the value is false, the computer skips over the rest of the while loop and proceeds to the
next command in the program. If the value of the expression is true, the computer executes
the hstatementi or block of hstatementsi inside the loop. Then it returns to the beginning of
the while loop and repeats the process. That is, it re-evaluates the hboolean-expressioni, ends
the loop if the value is false, and continues it if the value is true. This will continue over and
over until the value of the expression is false when the computer evaluates it; if that never
happens, then there will be an infinite loop.
Here is an example of a while loop that simply prints out the numbers 1, 2, 3, 4, 5:
int number; // The number to be printed.
number = 1; // Start with 1.
while ( number < 6 ) { // Keep going as long as number is < 6.
System.out.println(number);
number = number + 1; // Go on to the next number.
}
System.out.println("Done!");
The variable number is initialized with the value 1. So when the computer evaluates the
expression “number < 6” for the first time, it is asking whether 1 is less than 6, which is
true. The computer therefore proceeds to execute the two statements inside the loop. The
first statement prints out “1”. The second statement adds 1 to number and stores the result
back into the variable number; the value of number has been changed to 2. The computer has
reached the end of the loop, so it returns to the beginning and asks again whether number is
less than 6. Once again this is true, so the computer executes the loop again, this time printing
out 2 as the value of number and then changing the value of number to 3. It continues in this
way until eventually number becomes equal to 6. At that point, the expression “number < 6”
evaluates to false. So, the computer jumps past the end of the loop to the next statement
CHAPTER 3. CONTROL 74
and prints out the message “Done!”. Note that when the loop ends, the value of number is 6,
but the last value that was printed was 5.
By the way, you should remember that you’ll never see a while loop standing by itself
in a real program. It will always be inside a subroutine which is itself defined inside some
class. As an example of a while loop used inside a complete program, here is a little program
that computes the interest on an investment over several years. This is an improvement over
examples from the previous chapter that just reported the results for one year:
import textio.TextIO;
/**
* This class implements a simple program that will compute the amount of
* interest that is earned on an investment over a period of 5 years. The
* initial amount of the investment and the interest rate are input by the
* user. The value of the investment at the end of each year is output.
*/
public class Interest3 {
You should study this program, and make sure that you understand what the computer does
step-by-step as it executes the while loop.
CHAPTER 3. CONTROL 75
Yes Yes
Do statement Do statement
CHAPTER 3. CONTROL 76
In these diagrams, the arrows represent the flow of time as the statement is executed. Control
enters the diagram at the top and leaves at the bottom. Similarly, a flow control diagram for an
if..else statement makes it clear that exactly one of the two nested statements is executed:
Yes No
Is condition true?
Do statement
∗ ∗ ∗
Of course, either or both of the hstatementsi in an if statement can be a block, and again
many programmers prefer to add the braces even when they contain just a single statement.
So an if statement often looks like:
if ( hboolean-expression i ) {
hstatements i
}
else {
hstatements i
}
or:
if ( hboolean-expression i ) {
hstatements i
}
As an example, here is an if statement that exchanges the value of two variables, x and y,
but only if x is greater than y to begin with. After this if statement has been executed, we
can be sure that the value of x is definitely less than or equal to the value of y:
if ( x > y ) {
int temp; // A temporary variable for use in this block.
temp = x; // Save a copy of the value of x in temp.
x = y; // Copy the value of y into x.
y = temp; // Copy the value of temp into y.
}
CHAPTER 3. CONTROL 77
Finally, here is an example of an if statement that includes an else part. See if you can
figure out what it does, and why it would be used:
if ( years > 1 ) { // handle case for 2 or more years
System.out.print("The value of the investment after ");
System.out.print(years);
System.out.print(" years is $");
}
else { // handle case for 1 year
System.out.print("The value of the investment after 1 year is $");
} // end of if statement
System.out.printf("%1.2f", principal); // this is done in any case
I’ll have more to say about control structures later in this chapter. But you already know
the essentials. If you never learned anything more about control structures, you would already
know enough to perform any possible computing task. Simple looping and branching are all
you really need!
statement and a pair of plain if statements. Here is another pair of code segments that might
seem to do the same thing, but don’t. What’s the value of x after each code segment is executed?
int x; int x;
x = -1; x = -1;
if (x < 0) if (x < 0)
x = 1; x = 1;
else if (x >= 0)
x = 2; x = 2;
After the code on the left is executed, x is 1; after the code on the right, x is 2. If you don’t
believe this, work though the code step-by-step, doing exactly what the computer does when it
executes each step.
Suppose you have a task in mind that you want the computer to perform. One way to
proceed is to write a description of the task, and take that description as an outline of the
algorithm you want to develop. Then you can refine and elaborate that description, gradually
adding steps and detail, until you have a complete algorithm that can be translated directly
into programming language. This method is called stepwise refinement, and it is a type of
top-down design. As you proceed through the stages of stepwise refinement, you can write out
descriptions of your algorithm in pseudocode—informal instructions that imitate the structure
of programming languages without the complete detail and perfect syntax of actual program
code.
As an example, let’s see how one might develop the program from the previous section, which
computes the value of an investment over five years. The task that you want the program to
perform is: “Compute and display the value of an investment for each of the next five years,
where the initial investment and interest rate are to be specified by the user.” You might then
write—or more likely just think—that this can be expanded as:
Get the user’s input
Compute the value of the investment after 1 year
Display the value
Compute the value after 2 years
Display the value
Compute the value after 3 years
Display the value
Compute the value after 4 years
Display the value
Compute the value after 5 years
Display the value
This is correct, but rather repetitive. And seeing that repetition, you might notice an
opportunity to use a loop. A loop would take less typing. More important, it would be more
general: Essentially the same loop will work no matter how many years you want to process.
So, you might rewrite the above sequence of steps as:
Get the user’s input
while there are more years to process:
Compute the value after the next year
Display the value
Following this algorithm would certainly solve the problem, but for a computer we’ll have
to be more explicit about how to “Get the user’s input,” how to “Compute the value after the
next year,” and what it means to say “there are more years to process.” We can expand the
step, “Get the user’s input” into
Ask the user for the initial investment
Read the user’s response
Ask the user for the interest rate
Read the user’s response
To fill in the details of the step “Compute the value after the next year,” you have to
know how to do the computation yourself. (Maybe you need to ask your boss or professor for
clarification?) Let’s say you know that the value is computed by adding some interest to the
previous value. Then we can refine the while loop to:
CHAPTER 3. CONTROL 80
algorithm used indentation to show which statements are inside the loop. In Java, indentation
is completely ignored by the computer, so you need a pair of braces to tell the computer which
statements are in the loop. If you leave out the braces, the only statement inside the loop would
be “years = years + 1;". The other statements would only be executed once, after the loop
ends. The nasty thing is that the computer won’t notice this error for you, like it would if you
left out the parentheses around “(years < 5)”. The parentheses are required by the syntax of
the while statement. The braces are only required semantically. The computer can recognize
syntax errors but not semantic errors.)
One thing you should have noticed here is that my original specification of the problem—
“Compute and display the value of an investment for each of the next five years”—was far from
being complete. Before you start writing a program, you should make sure you have a complete
specification of exactly what the program is supposed to do. In particular, you need to know
what information the program is going to input and output and what computation it is going
to perform. Here is what a reasonably complete specification of the problem might look like in
this example:
“Write a program that will compute and display the value of
an investment for each of the next five years. Each year, interest
is added to the value. The interest is computed by multiplying
the current value by a fixed interest rate. Assume that the initial
value and the rate of interest are to be input by the user when the
program is run.”
to know when to continue the loop rather than when to stop it: We want to continue as long
as the number is not 1. So, we can expand our pseudocode algorithm to:
Get a positive integer N from the user;
while N is not 1:
Compute N = next term;
Output N;
Count this term;
Output the number of terms;
In order to compute the next term, the computer must take different actions depending on
whether N is even or odd. We need an if statement to decide between the two cases:
Get a positive integer N from the user;
while N is not 1:
if N is even:
Compute N = N/2;
else
Compute N = 3 * N + 1;
Output N;
Count this term;
Output the number of terms;
We are almost there. The one problem that remains is counting. Counting means that you
start with zero, and every time you have something to count, you add one. We need a variable
to do the counting. The variable must be set to zero once, before the loop starts, and it must
be incremented within the loop. (Again, this is a common pattern that you should expect to
see over and over.) With the counter added, we get:
Get a positive integer N from the user;
Let counter = 0;
while N is not 1:
if N is even:
Compute N = N/2;
else
Compute N = 3 * N + 1;
Output N;
Add 1 to counter;
Output the counter;
We still have to worry about the very first step. How can we get a positive integer from the
user? If we just read in a number, it’s possible that the user might type in a negative number
or zero. If you follow what happens when the value of N is negative or zero, you’ll see that the
program will go on forever, since the value of N will never become equal to 1. This is bad. In
this case, the problem is probably no big deal, but in general you should try to write programs
that are foolproof. One way to fix this is to keep reading in numbers until the user types in a
positive number:
Ask user to input a positive number;
Let N be the user’s response;
while N is not positive:
Print an error message;
Read another value for N;
Let counter = 0;
while N is not 1:
CHAPTER 3. CONTROL 83
if N is even:
Compute N = N/2;
else
Compute N = 3 * N + 1;
Output N;
Add 1 to counter;
Output the counter;
The first while loop will end only when N is a positive number, as required. (A common
beginning programmer’s error is to use an if statement instead of a while statement here: “If
N is not positive, ask the user to input another value.” The problem arises if the second number
input by the user is also non-positive. The if statement is only executed once, so the second
input number is never tested, and the program proceeds into an infinite loop. With the while
loop, after the second number is input, the computer jumps back to the beginning of the loop
and tests whether the second number is positive. If not, it asks the user for a third number,
and it will continue asking for numbers until the user enters an acceptable input. After the
while loop ends, we can be absolutely sure that N is a positive number.)
Here is a Java program implementing this algorithm. It uses the operators <= to mean “is
less than or equal to” and != to mean “is not equal to.” To test whether N is even, it uses
“N % 2 == 0”. All the operators used here were discussed in Section 2.5.
import textio.TextIO;
/**
* This program prints out a 3N+1 sequence starting from a positive
* integer specified by the user. It also counts the number of
* terms in the sequence, and prints out that number.
*/
public class ThreeN1 {
public static void main(String[] args) {
int N; // for computing terms in the sequence
int counter; // for counting the terms
System.out.print("Starting point for sequence: ");
N = TextIO.getlnInt();
while (N <= 0) {
System.out.print(
"The starting point must be positive. Please try again: " );
N = TextIO.getlnInt();
}
// At this point, we know that N > 0
counter = 0;
while (N != 1) {
if (N % 2 == 0)
N = N / 2;
else
N = 3 * N + 1;
System.out.println(N);
counter = counter + 1;
}
System.out.println();
System.out.print("There were ");
CHAPTER 3. CONTROL 84
System.out.print(counter);
System.out.println(" terms in the sequence.");
} // end of main()
} // end of class ThreeN1
Two final notes on this program: First, you might have noticed that the first term of the
sequence—the value of N input by the user—is not printed or counted by this program. Is
this an error? It’s hard to say. Was the specification of the program careful enough to decide?
This is the type of thing that might send you back to the boss/professor for clarification. The
problem (if it is one!) can be fixed easily enough. Just replace the line “counter = 0” before
the while loop with the two lines:
System.out.println(N); // print out initial term
counter = 1; // and count it
Second, there is the question of why this problem might be interesting. Well, it’s interesting
to mathematicians and computer scientists because of a simple question about the problem that
they haven’t been able to answer: Will the process of computing the 3N+1 sequence finish after
a finite number of steps for all possible starting values of N? Although individual sequences are
easy to compute, no one has been able to answer the general question. To put this another
way, no one knows whether the process of computing 3N+1 sequences can properly be called
an algorithm, since an algorithm is required to terminate after a finite number of steps! (Note:
This discussion really applies to integers, not to values of type int! That is, it assumes that the
value of N can take on arbitrarily large integer values, which is not true for a variable of type
int in a Java program. When the value of N in the program becomes too large to be represented
as a 32-bit int, the values output by the program are no longer mathematically correct. So
the Java program does not compute the correct 3N+1 sequence if N becomes too large. See
Exercise 8.2.)
In general, when the compiler gives multiple error messages, don’t try to fix the second error
message from the compiler until you’ve fixed the first one. Once the compiler hits an error in
your program, it can get confused, and the rest of the error messages might just be guesses.
Maybe the best advice is: Take the time to understand the error before you try to fix it.
Programming is not an experimental science.
When your program compiles without error, you are still not done. You have to test the
program to make sure it works correctly. Remember that the goal is not to get the right output
for the two sample inputs that the professor gave in class. The goal is a program that will
work correctly for all reasonable inputs. Ideally, when faced with an unreasonable input, it
should respond by gently chiding the user rather than by crashing. Test your program on a
wide variety of inputs. Try to find a set of inputs that will test the full range of functionality
that you’ve coded into your program. As you begin writing larger programs, write them in
stages and test each stage along the way. You might even have to write some extra code to
do the testing—for example to call a subroutine that you’ve just written. You don’t want to
be faced, if you can avoid it, with 500 newly written lines of code that have an error in there
somewhere.
The point of testing is to find bugs—semantic errors that show up as incorrect behavior
rather than as compilation errors. And the sad fact is that you will probably find them. Again,
you can minimize bugs by careful design and careful coding, but no one has found a way to
avoid them altogether. Once you’ve detected a bug, it’s time for debugging . You have to
track down the cause of the bug in the program’s source code and eliminate it. Debugging is a
skill that, like other aspects of programming, requires practice to master. So don’t be afraid of
bugs. Learn from them. One essential debugging skill is the ability to read source code—the
ability to put aside preconceptions about what you think it does and to follow it the way the
computer does—mechanically, step-by-step—to see what it really does. This is hard. I can still
remember the time I spent hours looking for a bug only to find that a line of code that I had
looked at ten times had a “1” where it should have had an “i”, or the time when I wrote a
subroutine named WindowClosing which would have done exactly what I wanted except that
the computer was looking for windowClosing (with a lower case “w”). Sometimes it can help
to have someone who doesn’t share your preconceptions look at your code.
Often, it’s a problem just to find the part of the program that contains the error. Most
programming environments come with a debugger , which is a program that can help you find
bugs. Typically, your program can be run under the control of the debugger. The debugger
allows you to set “breakpoints” in your program. A breakpoint is a point in the program where
the debugger will pause the program so you can look at the values of the program’s variables.
The idea is to track down exactly when things start to go wrong during the program’s execution.
The debugger will also let you execute your program one line at a time, so that you can watch
what happens in detail once you know the general area in the program where the bug is lurking.
I will confess that I only occasionally use debuggers myself. A more traditional approach to
debugging is to insert debugging statements into your program. These are output statements
that print out information about the state of the program. Typically, a debugging statement
would say something like
System.out.println("At start of while loop, N = " + N);
You need to be able to tell from the output where in your program the output is coming from,
and you want to know the value of important variables. Sometimes, you will find that the
computer isn’t even getting to a part of the program that you think it should be executing.
Remember that the goal is to find the first point in the program where the state is not what
CHAPTER 3. CONTROL 86
But how can we test whether there are more integers to process? A typical solution is to
tell the user to type in zero after all the data have been entered. This will work because we
are assuming that all the data are positive numbers, so zero is not a legal data value. The zero
is not itself part of the data to be averaged. It’s just there to mark the end of the real data.
A data value used in this way is sometimes called a sentinel value. So now the test in the
while loop becomes “while the input integer is not zero”. But there is another problem! The
first time the test is evaluated, before the body of the loop has ever been executed, no integer
has yet been read. There is no “input integer” yet, so testing whether the input integer is zero
doesn’t make sense. So, we have to do something before the while loop to make sure that the
test makes sense. Setting things up so that the test in a while loop makes sense the first time
it is executed is called priming the loop. In this case, we can simply read the first integer
before the beginning of the loop. Here is a revised algorithm:
Let sum = 0
Let count = 0
Read an integer
while the integer is not zero:
Add the integer to the sum
Count it
Read an integer
Divide sum by count to get the average
Print out the average
Notice that I’ve rearranged the body of the loop. Since an integer is read before the loop, the
loop has to begin by processing that integer. At the end of the loop, the computer reads a new
integer. The computer then jumps back to the beginning of the loop and tests the integer that
it has just read. Note that when the computer finally reads the sentinel value, the loop ends
before the sentinel value is processed. It is not added to the sum, and it is not counted. This
is the way it’s supposed to work. The sentinel is not part of the data. The original algorithm,
even if it could have been made to work without priming, was incorrect since it would have
summed and counted all the integers, including the sentinel. (Since the sentinel is zero, the sum
would still be correct, but the count would be off by one. Such so-called off-by-one errors
are very common. Counting turns out to be harder than it looks!)
We can easily turn the algorithm into a complete program. Note that the program cannot
use the statement “average = sum/count;” to compute the average. Since sum and count
are both variables of type int, the value of sum/count is an integer. The average should be
a real number. We’ve seen this problem before: we have to convert one of the int values to
a double to force the computer to compute the quotient as a real number. This can be done
by type-casting one of the variables to type double. The type cast “(double)sum” converts
the value of sum to a real number, so in the program the average is computed as “average =
((double)sum) / count;”. Another solution in this case would have been to declare sum to
be a variable of type double in the first place.
One other issue is addressed by the program: If the user enters zero as the first input value,
there are no data to process. We can test for this case by checking whether count is still equal
to zero after the while loop. This might seem like a minor point, but a careful programmer
should cover all the bases.
Here is the full source code for the program (with comments added, of course!):
import textio.TextIO;
/**
CHAPTER 3. CONTROL 88
has been moved to the end. The word “do” is added to mark the beginning of the loop. A
do..while statement has the form
do
hstatement i
while ( hboolean-expression i );
or, since, as usual, the hstatementi can be a block,
do {
hstatements i
} while ( hboolean-expression i );
Note the semicolon, ’;’, at the very end. This semicolon is part of the statement, just as
the semicolon at the end of an assignment statement or declaration is part of the statement.
Omitting it is a syntax error. (More generally, every statement in Java ends either with a
semicolon or a right brace, ’}’.)
To execute a do loop, the computer first executes the body of the loop—that is, the statement
or statements inside the loop—and then it evaluates the boolean expression. If the value of
the expression is true, the computer returns to the beginning of the do loop and repeats the
process; if the value is false, it ends the loop and continues with the next part of the program.
Since the condition is not tested until the end of the loop, the body of a do loop is always
executed at least once.
For example, consider the following pseudocode for a game-playing program. The do loop
makes sense here instead of a while loop because with the do loop, you know there will be at
least one game. Also, the test that is used at the end of the loop wouldn’t even make sense at
the beginning:
do {
Play a Game
Ask user if he wants to play another game
Read the user’s response
} while ( the user’s response is yes );
Let’s convert this into proper Java code. Since I don’t want to talk about game playing at the
moment, let’s say that we have a class named Checkers, and that the Checkers class contains
a static member subroutine named playGame() that plays one game of checkers against the
user. Then, the pseudocode “Play a game” can be expressed as the subroutine call statement
“Checkers.playGame();”. We need a variable to store the user’s response. The TextIO class
makes it convenient to use a boolean variable to store the answer to a yes/no question. The
input function TextIO.getlnBoolean() allows the user to enter the value as “yes” or “no”
(among other acceptable responses). “Yes” is considered to be true, and “no” is considered to
be false. So, the algorithm can be coded as
boolean wantsToContinue; // True if user wants to play again.
do {
Checkers.playGame();
System.out.print("Do you want to play again? ");
wantsToContinue = TextIO.getlnBoolean();
} while (wantsToContinue == true);
When the value of the boolean variable is set to false, it is a signal that the loop should end.
When a boolean variable is used in this way—as a signal that is set in one part of the program
CHAPTER 3. CONTROL 90
and tested in another part—it is sometimes called a flag or flag variable (in the sense of a
signal flag).
By the way, a more-than-usually-pedantic programmer would sneer at the test
“while (wantsToContinue == true)”. This test is exactly equivalent to “while
(wantsToContinue)”. Testing whether “wantsToContinue == true” is true amounts to the
same thing as testing whether “wantsToContinue” is true. A little less offensive is an expression
of the form “flag == false”, where flag is a boolean variable. The value of “flag == false”
is exactly the same as the value of “!flag”, where ! is the boolean negation operator. So
you can write “while (!flag)” instead of “while (flag == false)”, and you can write
“if (!flag)” instead of “if (flag == false)”.
Although a do..while statement is sometimes more convenient than a while statement,
having two kinds of loops does not make the language more powerful. Any problem that can be
solved using do..while loops can also be solved using only while statements, and vice versa.
In fact, if hdoSomethingi represents any block of program code, then
do {
hdoSomething i
} while ( hboolean-expression i );
has exactly the same effect as
hdoSomething i
while ( hboolean-expression i ) {
hdoSomething i
}
Similarly,
while ( hboolean-expression i ) {
hdoSomething i
}
can be replaced by
if ( hboolean-expression i ) {
do {
hdoSomething i
} while ( hboolean-expression i );
}
without changing the meaning of the program in any way.
∗ ∗ ∗
The continue statement is related to break, but less commonly used. A continue state-
ment tells the computer to skip the rest of the current iteration of the loop. However, instead
of jumping out of the loop altogether, it jumps back to the beginning of the loop and continues
with the next iteration (including evaluating the loop’s continuation condition to see whether
CHAPTER 3. CONTROL 92
any further iterations are required). As with break, when a continue is in a nested loop, it
will continue the loop that directly contains it; a “labeled continue” can be used to continue
the containing loop instead.
break and continue can be used in while loops and do..while loops. They can also be
used in for loops, which are covered in the next section. In Section 3.6, we’ll see that break can
also be used to break out of a switch statement. A break can occur inside an if statement,
but only if the if statement is nested inside a loop or inside a switch statement. In that case,
it does not mean to break out of the if. Instead, it breaks out of the loop or switch statement
that contains the if statement. The same consideration applies to continue statements inside
ifs.
The continuation condition is executed before each execution of the loop (including the first
execution), and the loop ends when this condition is false. The update part is executed at
the end of each execution of the loop, just before jumping back to check the condition.
The formal syntax of the for statement is as follows:
for ( hinitialization i; hcontinuation-condition i; hupdate i )
hstatement i
or, using a block statement:
for ( hinitialization i; hcontinuation-condition i; hupdate i ) {
hstatements i
}
The hcontinuation-conditioni must be a boolean-valued expression. The hinitializationi is usu-
ally a declaration or an assignment statement, but it can be any expression that would be
allowed as a statement in a program. The hupdatei can be any simple statement, but is usually
an increment, a decrement, or an assignment statement. Any of the three parts can be empty.
If the continuation condition is empty, it is treated as if it were “true,” so the loop will be
repeated forever or until it ends for some other reason, such as a break statement. (Some
people like to begin an infinite loop with “for (;;)” instead of “while (true)”.) Here’s a
flow control diagram for a for statement:
For Loop Flow of Control
Is condition true?
No
Yes
Do statement
Usually, the initialization part of a for statement assigns a value to some variable, and the
update changes the value of that variable with an assignment statement or with an increment
or decrement operation. The value of the variable is tested in the continuation condition, and
the loop ends when this condition evaluates to false. A variable used in this way is called a
loop control variable. In the example given above, the loop control variable was years.
Certainly, the most common type of for loop is the counting loop, where a loop control
variable takes on all integer values between some minimum and some maximum value. A
counting loop has the form
CHAPTER 3. CONTROL 94
// the loop.
for (N = 2; N <= 20; N = N + 2) {
System.out.println( N );
}
Perhaps it is worth stressing one more time that a for statement, like any statement except
for a variable declaration, never occurs on its own in a real program. A statement must be
inside the main routine of a program or inside some other subroutine. And that subroutine
must be defined inside a class. I should also remind you that every variable must be declared
before it can be used, and that includes the loop control variable in a for statement. In all
the examples that you have seen so far in this section, the loop control variables should be
declared to be of type int. It is not required that a loop control variable be an integer. Here,
for example, is a for loop in which the variable, ch, is of type char, using the fact that the ++
operator can be applied to characters as well as to numbers:
// Print out the alphabet on one line of output.
char ch; // The loop control variable;
// one of the letters to be printed.
for ( ch = ’A’; ch <= ’Z’; ch++ )
System.out.print(ch);
System.out.println();
the ones that actually do divide N evenly. (This is a correct solution, but is certainly not the
most efficient way to perform this task.) In pseudocode, the algorithm takes the form
Get a positive integer, N, from the user
Let divisorCount = 0
for each number, testDivisor, in the range from 1 to N:
if testDivisor is a divisor of N:
Count it by adding 1 to divisorCount
Output the count
This algorithm displays a common programming pattern that is used when some, but not all,
of a sequence of items are to be processed. The general pattern is
for each item in the sequence:
if the item passes the test:
process it
The for loop in our divisor-counting algorithm can be translated into Java code as
for (testDivisor = 1; testDivisor <= N; testDivisor++) {
if ( N % testDivisor == 0 )
divisorCount++;
}
On a modern computer, this loop can be executed very quickly. It is not impossible to run
it even for the largest legal int value, 2147483647. (If you wanted to run it for even larger
values, you could use variables of type long rather than int.) However, it does take a significant
amount of time for very large numbers. So when I implemented this algorithm, I decided to
output a dot every time the computer has tested ten million possible divisors. In the improved
version of the program, there are two types of counting going on. We have to count the number
of divisors and we also have to count the number of possible divisors that have been tested.
So the program needs two counters. When the second counter reaches 10000000, the program
outputs a ’.’ and resets the counter to zero so that we can start counting the next group of ten
million. Reverting to pseudocode, the algorithm now looks like
Get a positive integer, N, from the user
Let divisorCount = 0 // Number of divisors found.
Let numberTested = 0 // Number of possible divisors tested
// since the last period was output.
for each number, testDivisor, in the range from 1 to N:
if testDivisor is a divisor of N:
Count it by adding 1 to divisorCount
Add 1 to numberTested
if numberTested is 10000000:
print out a ’.’
Reset numberTested to 0
Output the count
Finally, we can translate the algorithm into a complete Java program:
import textio.TextIO;
/**
* This program reads a positive integer from the user.
* It counts how many divisors that number has, and
* then it prints the result.
*/
CHAPTER 3. CONTROL 97
combination of one control structure inside another is possible. We say that one structure is
nested inside another. You can even have multiple levels of nesting, such as a while loop
inside an if statement inside another while loop. The syntax of Java does not set a limit on
the number of levels of nesting. As a practical matter, though, it’s difficult to understand a
program that has more than a few levels of nesting.
Nested for loops arise naturally in many algorithms, and it is important to understand how
they work. Let’s look at a couple of examples. First, consider the problem of printing out a
multiplication table like this one:
1 2 3 4 5 6 7 8 9 10 11 12
2 4 6 8 10 12 14 16 18 20 22 24
3 6 9 12 15 18 21 24 27 30 33 36
4 8 12 16 20 24 28 32 36 40 44 48
5 10 15 20 25 30 35 40 45 50 55 60
6 12 18 24 30 36 42 48 54 60 66 72
7 14 21 28 35 42 49 56 63 70 77 84
8 16 24 32 40 48 56 64 72 80 88 96
9 18 27 36 45 54 63 72 81 90 99 108
10 20 30 40 50 60 70 80 90 100 110 120
11 22 33 44 55 66 77 88 99 110 121 132
12 24 36 48 60 72 84 96 108 120 132 144
The data in the table are arranged into 12 rows and 12 columns. The process of printing them
out can be expressed in a pseudocode algorithm as
for each rowNumber = 1, 2, 3, ..., 12:
Print the first twelve multiples of rowNumber on one line
Output a carriage return
The first step in the for loop can itself be expressed as a for loop. We can expand “Print the
first twelve multiples of rowNumber on one line” as:
for N = 1, 2, 3, ..., 12:
Print N * rowNumber
so a refined algorithm for printing the table has one for loop nested inside another:
for each rowNumber = 1, 2, 3, ..., 12:
for N = 1, 2, 3, ..., 12:
Print N * rowNumber
Output a carriage return
We want to print the output in neat columns, with each output number taking up four spaces.
This can be done using formatted output with format specifier %4d. Assuming that rowNumber
and N have been declared to be variables of type int, the algorithm can be expressed in Java as
for ( rowNumber = 1; rowNumber <= 12; rowNumber++ ) {
for ( N = 1; N <= 12; N++ ) {
// print in 4-character columns
System.out.printf( "%4d", N * rowNumber ); // No carriage return !
}
System.out.println(); // Add a carriage return at end of the line.
}
CHAPTER 3. CONTROL 99
This section has been weighed down with lots of examples of numerical processing. For our
next example, let’s do some text processing. Consider the problem of finding which of the 26
letters of the alphabet occur in a given string. For example, the letters that occur in “Hello
World” are D, E, H, L, O, R, and W. More specifically, we will write a program that will list all
the letters contained in a string and will also count the number of different letters. The string
will be input by the user. Let’s start with a pseudocode algorithm for the program.
Ask the user to input a string
Read the response into a variable, str
Let count = 0 (for counting the number of different letters)
for each letter of the alphabet:
if the letter occurs in str:
Print the letter
Add 1 to count
Output the count
Since we want to process the entire line of text that is entered by the user, we’ll use
TextIO.getln() to read it. The line of the algorithm that reads “for each letter of the al-
phabet” can be expressed as “for (letter=’A’; letter<=’Z’; letter++)”. But the if
statement inside the for loop needs still more thought before we can write the program. How
do we check whether the given letter, letter, occurs in str? One idea is to look at each
character in the string in turn, and check whether that character is equal to letter. We can
get the i-th character of str with the function call str.charAt(i), where i ranges from 0 to
str.length() - 1.
One more difficulty: A letter such as ’A’ can occur in str in either upper or lower case, ’A’
or ’a’. We have to check for both of these. But we can avoid this difficulty by converting str
to upper case before processing it. Then, we only have to check for the upper case letter. We
can now flesh out the algorithm fully:
Ask the user to input a string
Read the response into a variable, str
Convert str to upper case
Let count = 0
for letter = ’A’, ’B’, ..., ’Z’:
for i = 0, 1, ..., str.length()-1:
if letter == str.charAt(i):
Print letter
Add 1 to count
break // jump out of the loop, to avoid counting letter twice
Output the count
Note the use of break in the nested for loop. It is required to avoid printing or counting a given
letter more than once (in the case where it occurs more than once in the string). The break
statement breaks out of the inner for loop, but not the outer for loop. Upon executing the
break, the computer continues the outer loop with the next value of letter. You should try
to figure out exactly what count would be at the end of this program, if the break statement
were omitted. Here is the complete program:
import textio.TextIO;
/**
* This program reads a line of text entered by the user.
* It prints a list of the letters that occur in the text,
* and it reports how many different letters were found.
CHAPTER 3. CONTROL 100
*/
public class ListLetters {
public static void main(String[] args) {
String str; // Line of text entered by the user.
int count; // Number of different letters found in str.
char letter; // A letter of the alphabet.
System.out.println("Please type in a line of text.");
str = TextIO.getln();
str = str.toUpperCase();
count = 0;
System.out.println("Your input contains the following letters:");
System.out.println();
System.out.print(" ");
for ( letter = ’A’; letter <= ’Z’; letter++ ) {
int i; // Position of a character in str.
for ( i = 0; i < str.length(); i++ ) {
if ( letter == str.charAt(i) ) {
System.out.print(letter);
System.out.print(’ ’);
count++;
break;
}
}
}
System.out.println();
System.out.println();
System.out.println("There were " + count + " different letters.");
} // end main()
} // end class ListLetters
In fact, there is actually an easier way to determine whether a given letter occurs in a string,
str. The built-in function str.indexOf(letter) will return -1 if letter does not occur in
the string. It returns a number greater than or equal to zero if it does occur. So, we could
check whether letter occurs in str simply by checking “if (str.indexOf(letter) >= 0)”.
If we used this technique in the above program, we wouldn’t need a nested for loop. This gives
you a preview of how subroutines can be used to deal with complexity.
As usual, the statements inside an if statement can be blocks. The if statement represents
a two-way branch. The else part of an if statement—consisting of the word “else” and the
statement that follows it—can be omitted.
However, since the computer doesn’t care how a program is laid out on the page, this is almost
always written in the format:
if (hboolean-expression-1 i)
hstatement-1 i
else if (hboolean-expression-2 i)
hstatement-2 i
else
hstatement-3 i
You should think of this as a single statement representing a three-way branch. When the
computer executes this, one and only one of the three statements—hstatement-1 i, hstatement-
2 i, or hstatement-3 i—will be executed. The computer starts by evaluating hboolean-expression-
1 i. If it is true, the computer executes hstatement-1 i and then jumps all the way to the end of
the outer if statement, skipping the other two hstatementsi. If hboolean-expression-1 i is false,
the computer skips hstatement-1 i and executes the second, nested if statement. To do this,
it tests the value of hboolean-expression-2 i and uses it to decide between hstatement-2 i and
hstatement-3 i.
Here is an example that will print out one of three different messages, depending on the
value of a variable named temperature:
if (temperature < 50)
System.out.println("It’s cold.");
else if (temperature < 80)
System.out.println("It’s nice.");
else
System.out.println("It’s hot.");
If temperature is, say, 42, the first test is true. The computer prints out the message “It’s
cold”, and skips the rest—without even evaluating the second condition. For a temperature of
75, the first test is false, so the computer goes on to the second test. This test is true, so
the computer prints “It’s nice” and skips the rest. If the temperature is 173, both of the tests
evaluate to false, so the computer says “It’s hot” (unless its circuits have been fried by the
heat, that is).
You can go on stringing together “else-if’s” to make multiway branches with any number
of cases:
if (htest-1 i)
hstatement-1 i
else if (htest-2 i)
hstatement-2 i
else if (htest-3 i)
hstatement-3 i
.
. // (more cases)
.
else if (htest-N i)
hstatement-N i
else
hstatement-(N+1) i
The computer evaluates the tests, which are boolean expressions, one after the other until it
comes to one that is true. It executes the associated statement and skips the rest. If none
of the boolean expressions evaluate to true, then the statement in the else part is executed.
CHAPTER 3. CONTROL 103
This statement is called a multiway branch because one and only one of the statements will be
executed. The final else part can be omitted. In that case, if all the boolean expressions are
false, none of the statements are executed. Of course, each of the statements can be a block,
consisting of a number of statements enclosed between { and }. Admittedly, there is lot of
syntax here; as you study and practice, you’ll become comfortable with it. It might be useful
to look at a flow control diagram for the general “if..else if” statement shown above:
Yes No
Yes No
Do statemen
Yes
Do statement-2
Yes No
Do statement-N Do statement-(N+1)
Determining the relative order of y and z requires another if statement, so this becomes
if (x < y && x < z) { // x comes first
if (y < z)
System.out.println( x + " " + y + " " + z );
else
System.out.println( x + " " + z + " " + y );
}
else if (x > y && x > z) { // x comes last
if (y < z)
System.out.println( y + " " + z + " " + x );
else
System.out.println( z + " " + y + " " + x );
}
else { // x in the middle
if (y < z)
System.out.println( y + " " + x + " " + z);
else
System.out.println( z + " " + x + " " + y);
}
You might check that this code will work correctly even if some of the values are the same. If
the values of two variables are the same, it doesn’t matter which order you print them in.
Note, by the way, that even though you can say in English “if x is less than y and z,” you
can’t say in Java “if (x < y && z)”. The && operator can only be used between boolean
values, so you have to make separate tests, x<y and x<z, and then combine the two tests with
&&.
There is an alternative approach to this problem that begins by asking, “which order should
x and y be printed in?” Once that’s known, you only have to decide where to stick in z. This
line of thought leads to different Java code:
if ( x < y ) { // x comes before y
if ( z < x ) // z comes first
System.out.println( z + " " + x + " " + y);
else if ( z > y ) // z comes last
System.out.println( x + " " + y + " " + z);
else // z is in the middle
System.out.println( x + " " + z + " " + y);
}
else { // y comes before x
if ( z < y ) // z comes first
System.out.println( z + " " + y + " " + x);
else if ( z > x ) // z comes last
System.out.println( y + " " + x + " " + z);
else // z is in the middle
System.out.println( y + " " + z + " " + x);
}
Once again, we see how the same problem can be solved in many different ways. The two
approaches to this problem have not exhausted all the possibilities. For example, you might
start by testing whether x is greater than y. If so, you could swap their values. Once you’ve
done that, you know that x should be printed before y.
∗ ∗ ∗
CHAPTER 3. CONTROL 105
Finally, let’s write a complete program that uses an if statement in an interesting way. I
want a program that will convert measurements of length from one unit of measurement to
another, such as miles to yards or inches to feet. So far, the problem is extremely under-
specified. Let’s say that the program will only deal with measurements in inches, feet, yards,
and miles. It would be easy to extend it later to deal with other units. The user will type in
a measurement in one of these units, such as “17 feet” or “2.73 miles”. The output will show
the length in terms of each of the four units of measure. (This is easier than asking the user
which units to use in the output.) An outline of the process is
Read the user’s input measurement and units of measure
Express the measurement in inches, feet, yards, and miles
Display the four results
The program can read both parts of the user’s input from the same line by using
TextIO.getDouble() to read the numerical measurement and TextIO.getlnWord() to read
the unit of measure. The conversion into different units of measure can be simplified by first
converting the user’s input into inches. From there, the number of inches can easily be con-
verted into feet, yards, and miles. Before converting into inches, we have to test the input to
determine which unit of measure the user has specified:
Let measurement = TextIO.getDouble()
Let units = TextIO.getlnWord()
if the units are inches
Let inches = measurement
else if the units are feet
Let inches = measurement * 12 // 12 inches per foot
else if the units are yards
Let inches = measurement * 36 // 36 inches per yard
else if the units are miles
Let inches = measurement * 12 * 5280 // 5280 feet per mile
else
The units are illegal!
Print an error message and stop processing
Let feet = inches / 12.0
Let yards = inches / 36.0
Let miles = inches / (12.0 * 5280.0)
Display the results
Since units is a String, we can use units.equals("inches") to check whether the spec-
ified unit of measure is “inches”. However, it would be nice to allow the units to be spec-
ified as “inch” or abbreviated to “in”. To allow these three possibilities, we can check if
(units.equals("inches") || units.equals("inch") || units.equals("in")). It would
also be nice to allow upper case letters, as in “Inches” or “IN”. We can do this by converting
units to lower case before testing it or by substituting the function units.equalsIgnoreCase
for units.equals.
In my final program, I decided to make things more interesting by allowing the user to
repeat the process of entering a measurement and seeing the results of the conversion for each
measurement. The program will end only when the user inputs 0. To program that, I just had
to wrap the above algorithm inside a while loop, and make sure that the loop ends when the
user inputs a 0. Here’s the complete program:
import textio.TextIO;
CHAPTER 3. CONTROL 106
/**
* This program will convert measurements expressed in inches,
* feet, yards, or miles into each of the possible units of
* measure. The measurement is input by the user, followed by
* the unit of measure. For example: "17 feet", "1 inch", or
* "2.73 mi". Abbreviations in, ft, yd, and mi are accepted.
* The program will continue to read and convert measurements
* until the user enters an input of 0.
*/
public class LengthConverter {
public static void main(String[] args) {
double measurement; // Numerical measurement, input by user.
String units; // The unit of measure for the input, also
// specified by the user.
double inches, feet, yards, miles; // Measurement expressed in
// each possible unit of
// measure.
System.out.println("Enter measurements in inches, feet, yards, or miles.");
System.out.println("For example: 1 inch 17 feet 2.73 miles");
System.out.println("You can use abbreviations: in ft yd mi");
System.out.println("I will convert your input into the other units");
System.out.println("of measure.");
System.out.println();
while (true) {
/* Get the user’s input, and convert units to lower case. */
System.out.print("Enter your measurement, or 0 to end: ");
measurement = TextIO.getDouble();
if (measurement == 0)
break; // Terminate the while loop.
units = TextIO.getlnWord();
units = units.toLowerCase(); // convert units to lower case
/* Convert the input measurement to inches. */
if (units.equals("inch") || units.equals("inches")
|| units.equals("in")) {
inches = measurement;
}
else if (units.equals("foot") || units.equals("feet")
|| units.equals("ft")) {
inches = measurement * 12;
}
else if (units.equals("yard") || units.equals("yards")
|| units.equals("yd")) {
inches = measurement * 36;
}
else if (units.equals("mile") || units.equals("miles")
|| units.equals("mi")) {
inches = measurement * 12 * 5280;
}
CHAPTER 3. CONTROL 107
else {
System.out.println("Sorry, but I don’t understand \""
+ units + "\".");
continue; // back to start of while loop
}
/* Convert measurement in inches to feet, yards, and miles. */
feet = inches / 12;
yards = inches / 36;
miles = inches / (12*5280);
/* Output measurement in terms of each unit of measure. */
System.out.println();
System.out.println("That’s equivalent to:");
System.out.printf("%14.5g inches%n", inches);
System.out.printf("%14.5g feet%n", feet);
System.out.printf("%14.5g yards%n", yards);
System.out.printf("%14.5g miles%n", miles);
System.out.println();
} // end while
System.out.println();
System.out.println("OK! Bye for now.");
} // end main()
} // end class LengthConverter
(Note that this program uses formatted output with the “g” format specifier. In this pro-
gram, we have no control over how large or how small the numbers might be. It could easily
make sense for the user to enter very large or very small measurements. The “g” format will
print a real number in exponential form if it is very large or very small, and in the usual decimal
form otherwise. Remember that in the format specification %14.5g, the 5 is the total number
of significant digits that are to be printed, so we will always get the same number of significant
digits in the output, no matter what the size of the number. If we had used an “f” format
specifier such as %14.5f, the output would be in decimal form with 5 digits after the decimal
point. This would print the number 0.000000000745482 as 0.00000, with no significant digits
at all! With the “g” format specifier, the output would be 7.4549e-10.)
if ( done )
; // Empty statement
else
System.out.println( "Not done yet.");
does nothing when the boolean variable done is true, and prints out “Not done yet” when
it is false. You can’t just leave out the semicolon in this example, since Java syntax requires
an actual statement between the if and the else. I prefer, though, to use an empty block,
consisting of { and } with nothing between, for such cases.
Occasionally, stray empty statements can cause annoying, hard-to-find errors in a program.
For example, the following program segment prints out “Hello” just once, not ten times:
for (i = 0; i < 10; i++);
System.out.println("Hello");
Why? Because the “;” at the end of the first line is a statement, and it is this empty statement
that is executed ten times. The System.out.println statement is not really inside the for
statement at all, so it is executed just once, after the for loop has completed. The for loop
just does nothing, ten times!
.
case hconstant-N i:
hstatements-N i
break;
default: // optional default case
hstatements-(N+1) i
} // end of switch statement
This has exactly the same effect as the following multiway if statement, but the switch
statement can be more efficient because the computer can evaluate one expression and jump
directly to the correct case, whereas in the if statement, the computer must evaluate up to N
expressions before it knows which set of statements to execute:
if (hexpression i == hconstant-1 i) { // but use .equals for String!!
hstatements-1 i
}
else if (hexpression i == hconstant-2 i) {
hstatements-2 i
}
.
.
.
else if (hexpression i == hconstant-N i) {
hstatements-N i
}
else {
hstatements-(N+1) i
}
The break statements in the switch are not actually required by the syntax of the switch
statement. The effect of a break is to make the computer jump past the end of the switch state-
ment, skipping over all the remaining cases. If you leave out the break statement, the computer
will just forge ahead after completing one case and will execute the statements associated with
the next case label. This is rarely what you want, but it is legal. (I will note here—although
you won’t understand it until you get to the next chapter—that inside a subroutine, the break
statement is sometimes replaced by a return statement, which terminates the subroutine as
well as the switch.)
Note that you can leave out one of the groups of statements entirely (including the break).
You then have two case labels in a row, containing two different constants. This just means
that the computer will jump to the same place and perform the same action for each of the two
constants.
Here is an example of a switch statement. This is not a useful example, but it should be
easy for you to follow. Note, by the way, that the constants in the case labels don’t have to be
in any particular order, but they must all be different:
switch ( N ) { // (Assume N is an integer variable.)
case 1:
System.out.println("The number is 1.");
break;
case 2:
case 4:
case 8:
System.out.println("The number is 2, 4, or 8.");
System.out.println("(That’s a power of 2!)");
CHAPTER 3. CONTROL 110
break;
case 3:
case 6:
case 9:
System.out.println("The number is 3, 6, or 9.");
System.out.println("(That’s a multiple of 3!)");
break;
case 5:
System.out.println("The number is 5.");
break;
default:
System.out.println("The number is 7 or is outside the range 1 to 9.");
}
The switch statement is pretty primitive as control structures go, and it’s easy to make mis-
takes when you use it. Java takes all its control structures directly from the older programming
languages C and C++. The switch statement is certainly one place where the designers of Java
should have introduced some improvements.
case 2:
System.out.println("Enter the number of feet: ");
measurement = TextIO.getlnDouble();
inches = measurement * 12;
break;
case 3:
System.out.println("Enter the number of yards: ");
measurement = TextIO.getlnDouble();
inches = measurement * 36;
break;
case 4:
System.out.println("Enter the number of miles: ");
measurement = TextIO.getlnDouble();
inches = measurement * 12 * 5280;
break;
default:
System.out.println("Error! Illegal option number! I quit!");
System.exit(1);
} // end switch
/* Now go on to convert inches to feet, yards, and miles... */
This example could instead be written using a String in the switch statement:
String units; // Unit of measurement, entered by user.
double measurement; // A numerical measurement, input by the user.
double inches; // The same measurement, converted into inches.
/* Read the user’s unit of measurement. */
System.out.println("What unit of measurement does your input use?");
System.out.print("Legal responses: inches, feet, yards, or miles : ");
units = TextIO.getln().toLowerCase();
/* Read user’s measurement and convert to inches. */
System.out.print("Enter the number of " + units + ": ");
measurement = TextIO.getlnDouble();
switch ( units ) {
case "inches":
inches = measurement;
break;
case "feet":
inches = measurement * 12;
break;
case "yards":
inches = measurement * 36;
break;
case "miles":
inches = measurement * 12 * 5280;
break;
default:
System.out.println("Wait a minute! Illegal unit of measure! I quit!");
System.exit(1);
} // end switch
CHAPTER 3. CONTROL 112
String computerMove;
switch ( (int)(3*Math.random()) ) {
case 0:
computerMove = "Rock";
break;
case 1:
computerMove = "Paper";
break;
case 2:
computerMove = "Scissors";
break;
}
System.out.println("The computer’s move is " + computerMove); // ERROR!
Now there is a subtle error on the last line! The problem here is due to definite assignment,
the idea that the Java compiler must be able to determine that a variable has definitely been
assigned a value before its value is used. Definite assignment was introduced in Subsection 3.1.4.
In this example, it’s true that the three cases in the switch cover all the possibilities, but the
compiler is not smart enough to figure that out; it just sees that there is an integer-valued
expression in the switch but not all possible integer values are covered by the given cases.
A simple solution is to replace the final case in the switch statement with default. With
a default case, all possible values of the expression in the switch are certainly covered, and
the compiler knows that computerMove is definitely assigned a value:
String computerMove;
switch ( (int)(3*Math.random()) ) {
case 0:
computerMove = "Rock";
break;
case 1:
computerMove = "Paper";
break;
default:
computerMove = "Scissors";
break;
}
System.out.println("The computer’s move is " + computerMove); // OK!
3.7.1 Exceptions
The term exception is used to refer to the type of event that one might want to handle with
a try..catch. An exception is an exception to the normal flow of control in the program.
The term is used in preference to “error” because in some cases, an exception might not be
considered to be an error at all. You can sometimes think of an exception as just another way
to organize a program.
Exceptions in Java are represented as objects of type Exception. Actual exceptions are
usually defined by subclasses of Exception. Different subclasses represent different types of
CHAPTER 3. CONTROL 115
exceptions. We will look at only two types of exception in this section: NumberFormatException
and IllegalArgumentException.
A NumberFormatException can occur when an attempt is made to convert a string
into a number. Such conversions are done by the functions Integer.parseInt
and Double.parseDouble. (See Subsection 2.5.7.) Consider the function call
Integer.parseInt(str) where str is a variable of type String. If the value of str is the
string "42", then the function call will correctly convert the string into the int 42. However,
if the value of str is, say, "fred", the function call will fail because "fred" is not a legal
string representation of an int value. In this case, an exception of type NumberFormatException
occurs. If nothing is done to handle the exception, the program will crash.
An IllegalArgumentException can occur when an illegal value is passed as a parameter to a
subroutine. For example, if a subroutine requires that a parameter be greater than or equal to
zero, an IllegalArgumentException might occur when a negative value is passed to the subroutine.
How to respond to the illegal value is up to the person who wrote the subroutine, so we
can’t simply say that every illegal parameter value will result in an IllegalArgumentException.
However, it is a common response.
3.7.2 try..catch
When an exception occurs, we say that the exception is “thrown.” For example, we say that
Integer.parseInt(str) throws an exception of type NumberFormatException when the value
of str is illegal. When an exception is thrown, it is possible to “catch” the exception and
prevent it from crashing the program. This is done with a try..catch statement. In simplified
form, the syntax for a try..catch statement can be:
try {
hstatements-1 i
}
catch ( hexception-class-name i hvariable-name i ) {
hstatements-2 i
}
The hexception-class-namei could be NumberFormatException, IllegalArgumentException, or
some other exception class. When the computer executes this try..catch statement, it ex-
ecutes hstatements-1 i, the statements inside the try part. If no exception occurs during the
execution of hstatements-1 i, then the computer just skips over the catch part and proceeds
with the rest of the program. However, if an exception of type hexception-class-namei occurs
during the execution of hstatements-1 i, the computer immediately jumps from the point where
the exception occurs to the catch part and executes hstatements-2 i, skipping any remaining
statements in hstatements-1 i. Note that only one type of exception is caught; if some other
type of exception occurs during the execution of hstatements-1 i, it will crash the program as
usual.
During the execution of hstatements-2 i, the hvariable-namei represents the exception object,
so that you can, for example, print it out. The exception object contains information about
the cause of the exception. This includes an error message, which will be displayed if you print
out the exception object.
After the end of the catch part, the computer proceeds with the rest of the program; the
exception has been caught and handled and does not crash the program.
By the way, note that the braces, { and }, are part of the syntax of the try..catch
statement. They are required even if there is only one statement between the braces. This is
CHAPTER 3. CONTROL 116
different from the other statements we have seen, where the braces around a single statement
are optional.
As an example, suppose that str is a variable of type String whose value might or might
not represent a legal real number. Then we could say:
double x;
try {
x = Double.parseDouble(str);
System.out.println( "The number is " + x );
}
catch ( NumberFormatException e ) {
System.out.println( "Not a legal number." );
x = Double.NaN;
}
If an error is thrown by the call to Double.parseDouble(str), then the output statement in
the try part is skipped, and the statement in the catch part is executed. (In this example,
I set x to be the value Double.NaN when an exception occurs. Double.NaN is the special
“not-a-number” value for type double.)
It’s not always a good idea to catch exceptions and continue with the program. Often that
can just lead to an even bigger mess later on, and it might be better just to let the exception
crash the program at the point where it occurs. However, sometimes it’s possible to recover
from an error.
Suppose, for example, we want a program that will find the average of a sequence of real
numbers entered by the user, and we want the user to signal the end of the sequence by entering
a blank line. (This is similar to the sample program ComputeAverage.java from Section 3.3, but
in that program the user entered a zero to signal end-of-input.) If we use TextIO.getlnInt()
to read the user’s input, we will have no way of detecting the blank line, since that function
simply skips over blank lines. A solution is to use TextIO.getln() to read the user’s input.
This allows us to detect a blank input line, and we can convert non-blank inputs to numbers
using Double.parseDouble. And we can use try..catch to avoid crashing the program when
the user’s input is not a legal number. Here’s the program:
import textio.TextIO;
public class ComputeAverage2 {
public static void main(String[] args) {
String str; // The user’s input.
double number; // The input converted into a number.
double total; // The total of all numbers entered.
double avg; // The average of the numbers.
int count; // The number of numbers entered.
total = 0;
count = 0;
System.out.println("Enter your numbers, press return to end.");
while (true) {
System.out.print("? ");
str = TextIO.getln();
if (str.equals("")) {
break; // Exit the loop, since the input line was blank.
}
try {
CHAPTER 3. CONTROL 117
number = Double.parseDouble(str);
// If an error occurs, the next 2 lines are skipped!
total = total + number;
count = count + 1;
}
catch (NumberFormatException e) {
System.out.println("Not a legal number! Try again.");
}
}
avg = total/count;
System.out.printf("The average of %d numbers is %1.6g%n", count, avg);
}
}
* successfully read from the file. The name of the file will be
* input by the user.
*/
public class AverageNumbersFromFile {
public static void main(String[] args) {
while (true) {
String fileName; // The name of the file, to be input by the user.
System.out.print("Enter the name of the file: ");
fileName = TextIO.getln();
try {
TextIO.readFile( fileName ); // Try to open the file for input.
break; // If that succeeds, break out of the loop.
}
catch ( IllegalArgumentException e ) {
System.out.println("Can’t read from the file \"" + fileName + "\".");
System.out.println("Please try again.\n");
}
}
/* At this point, TextIO is reading from the file. */
double number; // A number read from the data file.
double sum; // The sum of all the numbers read so far.
int count; // The number of numbers that were read.
sum = 0;
count = 0;
try {
while (true) { // Loop ends when an exception occurs.
number = TextIO.getDouble();
count++; // This is skipped when the exception occurs
sum += number;
}
}
catch ( IllegalArgumentException e ) {
// We expect this to occur when the end-of-file is encountered.
// We don’t consider this to be an error, so there is nothing to do
// in this catch clause. Just proceed with the rest of the program.
}
// At this point, we’ve read the entire file.
System.out.println();
System.out.println("Number of data values read: " + count);
System.out.println("The sum of the data values: " + sum);
if ( count == 0 )
System.out.println("Can’t compute an average of 0 values.");
else
System.out.println("The average of the values: " + (sum/count));
}
}
CHAPTER 3. CONTROL 119
form can be used like any other variable: You can assign a value to it, print it out, use it in an
expression, and so on.
An array also contains a kind of variable representing its length. For example, you can refer
to the length of the array namelist as namelist.length. However, you cannot assign a value
to namelist.length, since the length of an array cannot be changed.
Before you can use a variable to refer to an array, that variable must be declared, and it
must have a type. For an array of Strings, for example, the type for the array variable would
be String[ ], and for an array of ints, it would be int[ ]. In general, an array type consists of the
base type of the array followed by a pair of empty square brackets. Array types can be used to
declare variables; for example,
String[] namelist;
int[] A;
double[] prices;
and variables declared in this way can refer to arrays. However, declaring a variable does not
make the actual array. Like all variables, an array variable has to be assigned a value before
it can be used. In this case, the value is an array. Arrays have to be created using a special
syntax. (The syntax is related to the fact that arrays in Java are actually objects, but that
doesn’t need to concern us here.) Arrays are created with an operator named new . Here are
some examples:
namelist = new String[1000];
A = new int[5];
prices = new double[100];
The general syntax is
harray-variable i = new hbase-type i[harray-length i];
The length of the array can be given as either an integer or an integer-valued expression. For
example, after the assignment statement “A = new int[5];”, A is an array containing the five
integer elements A[0], A[1], A[2], A[3], and A[4]. Also, A.length would have the value 5.
It’s useful to have a picture in mind:
When you create an array of int, each element of the array is automatically initialized to
zero. Any array of numbers is filled with zeros when it is created. An array of boolean is filled
with the value false. And an array of char is filled with the character that has Unicode code
number zero. (For an array of String, the initial value is null, a special value used for objects
that we won’t encounter officially until Section 5.1.)
CHAPTER 3. CONTROL 121
want to average the non-zero elements. In this case, the number of items that we add up can
be less than the length of the array, so we will need to keep a count of the number of items
added to the sum:
double total; // The sum of the non-zero numbers in the array.
int count; // The number of non-zero numbers.
double average; // The average of the non-zero numbers.
int i;
total = 0;
count = 0;
for ( i = 0; i < A.length; i++ ) {
if ( A[i] != 0 ) {
total = total + A[i]; // Add element to the total
count = count + 1; // and count it.
}
}
if (count == 0) {
System.out.println("There were no non-zero elements.");
}
else {
average = total / count; // Divide by number of items
System.out.printf("Average of %d elements is %1.5g%n",
count, average);
}
we select someone whose birthday is day number i, we first check whether used[i] is true.
If it is true, then this is the second person with that birthday. We are done. On the other
hand, if used[i] is false, we set used[i] to be true to record the fact that we’ve encountered
someone with that birthday, and we go on to the next person. Here is a program that carries
out the simulated experiment (of course, in the program, there are no simulated people, only
simulated birthdays):
/**
* Simulate choosing people at random and checking the day of the year they
* were born on. If the birthday is the same as one that was seen previously,
* stop, and output the number of people who were checked.
*/
public class BirthdayProblem {
public static void main(String[] args) {
boolean[] used; // For recording the possible birthdays
// that have been seen so far. A value
// of true in used[i] means that a person
// whose birthday is the i-th day of the
// year has been found.
int count; // The number of people who have been checked.
used = new boolean[365]; // Initially, all entries are false.
count = 0;
while (true) {
// Select a birthday at random, from 0 to 364.
// If the birthday has already been used, quit.
// Otherwise, record the birthday as used.
int birthday; // The selected birthday.
birthday = (int)(Math.random()*365);
count++;
System.out.printf("Person %d has birthday number %d%n", count, birthday);
if ( used[birthday] ) {
// This day was found before; it’s a duplicate. We are done.
break;
}
used[birthday] = true;
} // end while
System.out.println();
System.out.println("A duplicate birthday was found after "
+ count + " tries.");
}
} // end class BirthdayProblem
You should study the program to understand how it works and how it uses the array. Also, try
it out! You will probably find that a duplicate birthday tends to occur sooner than you expect.
CHAPTER 3. CONTROL 124
System.out.println( numbers[i] );
}
} // end main();
} // end class ReverseInputNumbers
It is especially important to note how the variable count plays a dual role. It is the number
of items that have been entered into the array. But it is also the index of the next available
spot in the array.
When the time comes to print out the numbers in the array, the last occupied spot in the
array is location count - 1, so the for loop prints out values starting from location count - 1
and going down to 0. This is also a nice example of processing the elements of an array in reverse
order.
∗ ∗ ∗
You might wonder what would happen in this program if the user tries to input more than
100 numbers. The result would be an error that would crash the program. When the user enters
the 101-st number, the program tries to store that number in an array element number[100].
However, there is no such array element. There are only 100 items in the array, and the
index of the last item is 99. The attempt to use number[100] generates an exception of type
ArrayIndexOutOfBoundsException. Exceptions of this type are a common source of run-time
errors in programs that use arrays.
0 1 2 3 4 5 6
0 13 7 33 54 -5 -1 92
1 -3 0 8 42 18 0 67
2 44 78 90 79 -5 72 22
3 43 -6 17 100 1 -12 12
4 2 0 58 58 36 21 87
This 5-by-7 grid contains a total of 35 elements. The rows in a 2D array are numbered 0, 1, 2,
. . . , up to the number of rows minus one. Similarly, the columns are numbered from zero up
to the number of columns minus one. Each individual element in the array can be picked out
by specifying its row number and its column number. (The illustration shown here is not what
the array actually looks like in the computer’s memory, but it does show the logical structure
of the array.)
In Java, the syntax for two-dimensional arrays is similar to the syntax for one-dimensional
arrays, except that an extra index is involved, since picking out an element requires both a row
number and a column number. For example, if A is a 2D array of int, then A[3][2] would be
the element in row 3, column 2. That would pick out the number 17 in the array shown above.
CHAPTER 3. CONTROL 126
The type for A would be given as int[ ][ ], with two pairs of empty brackets. To declare the array
variable and create the array, you could say,
int[][] A;
A = new int[5][7];
The second line creates a 2D array with 5 rows and 7 columns. Two-dimensional arrays are
often processed using nested for loops. For example, the following code segment will print out
the elements of A in neat columns:
int row, col; // loop-control-variables for accessing rows and columns in A
for ( row = 0; row < 5; row++ ) {
for ( col = 0; col < 7; col++ ) {
System.out.printf( "%7d", A[row][col] );
}
System.out.println();
}
The base type of a 2D array can be anything, so you can have arrays of type double[ ][ ],
String[ ][ ], and so on.
There are some natural uses for 2D arrays. For example, a 2D array can be used to store the
contents of the board in a game such as chess or checkers. And an example in Subsection 4.7.3
uses a 2D array to hold the colors of a grid of colored squares. But sometimes two-dimensional
arrays are used in problems in which the grid is not so visually obvious. Consider a company
that owns 25 stores. Suppose that the company has data about the profit earned at each store
for each month in the year 2018. If the stores are numbered from 0 to 24, and if the twelve
months from January 2018 through December 2018 are numbered from 0 to 11, then the profit
data could be stored in an array, profit, created as follows:
double[][] profit;
profit = new double[25][12];
profit[3][2] would be the amount of profit earned at store number 3 in March, and more
generally, profit[storeNum][monthNum] would be the amount of profit earned in store number
storeNum in month number monthNum (where the numbering, remember, starts from zero).
Let’s assume that the profit array has already been filled with data. This data can be
processed in a lot of interesting ways. For example, the total profit for the company—for the
whole year from all its stores—can be calculated by adding up all the entries in the array:
double totalProfit; // Company’s total profit in 2018.
int store, month; // variables for looping through the stores and the months
totalProfit = 0;
for ( store = 0; store < 25; store++ ) {
for ( month = 0; month < 12; month++ )
totalProfit += profit[store][month];
}
Sometimes it is necessary to process a single row or a single column of an array, not the
entire array. For example, to compute the total profit earned by the company in December,
that is, in month number 11, you could use the loop:
double decemberProfit;
int storeNum;
decemberProfit = 0.0;
for ( storeNum = 0; storeNum < 25; storeNum++ ) {
CHAPTER 3. CONTROL 127
decemberProfit += profit[storeNum][11];
}
Two-dimensional arrays are sometimes useful, but they are much less common than one-
dimensional arrays. Java actually allows arrays of even higher dimension, but they are only
rarely encountered in practice.
0 width
x
0
Hello World
height
Assuming that the drawing area is 800-by-500 pixels, the rectangle in the upper left of the
picture would have, approximately, width 200, height 150, and upper left corner at coordinates
(50,50).
∗ ∗ ∗
Drawing in Java is done using a graphics context. A graphics context is an object. As an
object, it can include subroutines and data. Among the subroutines in a graphics context are
routines for drawing basic shapes such as lines, rectangles, ovals, and text. (When text appears
on the screen, the characters have to be drawn there by the computer, just like the computer
draws any other shapes.) Among the data in a graphics context are the color and font that
are currently selected for drawing. (A font determines the style and size of characters.) One
other piece of data in a graphics context is the “drawing surface” on which the drawing is done.
Different graphics context objects can draw to different drawing surfaces. For us, the drawing
surface will be the content area of a window, not including its border or title bar.
There are two ways to draw a shape in JavaFX: You can fill the shape, meaning you can
set the color of each of the pixels inside the shape. Or you can stroke the shape, meaning that
you set the color of the pixels that lie along the border of the shape. Some shapes, such as a
line, can only be stroked. A graphics context actually keeps track of two separate colors, one
used for filling shapes and one used for stroking shapes. Stroking a shape is like dragging a
pen along the border of the shape. The properties of that pen (such as its size and whether it
produces a solid line or a dashed line) are properties of the graphics context.
A graphics context is represented by a variable. The type for the variable is GraphicsContext
(just like the type for a string variable is String ). The variable is often named g, but the name
of the variable is of course up to the programmer. Here are a few of the subroutines that are
available in a graphics context g. Note that all numerical parameter values can be of type
double.
• g.setFill(c) is called to set the color to be used for filling shapes. The parameter,
c is an object belonging to a class named Color. There are many constants represent-
ing standard colors that can be used as the parameter in this subroutine. The stan-
dard colors range from common colors such as Color.BLACK, Color.WHITE, Color.RED,
Color.GREEN, Color.BLUE, and Color.YELLOW, to more exotic color names such as
Color.CORNFLOWERBLUE. (Later, we will see that it is also possible to create new colors.)
For example, if you want to fill shapes with red, you would say “g.setFill(Color.RED);”.
The specified color is used for all subsequent fill operations up until the next time
g.setFill() is called. Note that previously drawn shapes are not affected!
• g.setStroke(c) is called to set the color to be used for stroking shapes. It works similarly
to g.setFill.
CHAPTER 3. CONTROL 129
• g.setLineWidth(w) sets the size of the pen that will be used for subsequent stroke oper-
ations, where w is measured in pixels.
• g.strokeLine(x1,y1,x2,y2) draws a line from the point with coordinates (x1,y1) to
the point with coordinates (x2,y2). The width of the line is 1, unless a different line
width has been set by calling g.setLineWidth(), and the color is black unless a different
color has been set by calling g.setStroke().
• g.strokeRect(x,y,w,h) draws the outline of a rectangle with vertical and horizontal
sides. This subroutine draws the outline of the rectangle whose top-left corner is x pixels
from the left edge of the drawing area and y pixels down from the top. The horizontal
width of the rectangle is w pixels, and the vertical height is h pixels. Color and line width
are set by calling g.setStroke() and g.setLineWidth().
• g.fillRect(x,y,w,h) is similar to g.strokeRect() except that it fills in the inside of
the rectangle instead of drawing an outline, and it uses the color set by g.setFill().
• g.strokeOval(x,y,w,h) draws the outline of an oval. The oval just fits inside the rect-
angle that would be drawn by g.strokeRect(x,y,w,h). To get a circle, use the same
values for w and for h.
• g.fillOval(x,y,w,h) is similar to g.strokeOval() except that it fills in the inside of
the oval instead of drawing an outline.
This is enough information to draw some pictures using Java graphics. To start with
something simple, let’s say that we want to draw a set of ten parallel lines, something like this:
Let’s say that the lines are 200 pixels long and that the distance from each line to the next
is 10 pixels, and let’s put the start of the first line at the pixel with coordinates (100,50). To
draw one line, we just have to call g.strokeLine(x1,y1,x2,y2) with appropriate values for
the parameters. Now, all the lines start at x -coordinate 100, so we can use the constant 100 as
the value for x1. Since the lines are 200 pixels long, we can use the constant 300 as the value
for x2. The y-coordinates of the lines are different, but we can see that both endpoints of a
line have the same y-coordinates, so we can use a single variable as the value for y1 and for
y2. Using y as the name of that variable, the command for drawing one of the lines becomes
g.strokeLine(100,y,300,y). The value of y is 50 for the top line and increases by 10 each
time we move down from one line to the next. We just need to make sure that y takes on the
correct sequence of values. We can use a for loop that counts from 1 to 10:
int y; // y-coordinate for the line
int i; // loop control variable
y = 50; // y starts at 50 for the first line
for ( i = 1; i <= 10; i++ ) {
g.strokeLine( 100, y, 300, y );
y = y + 10; // increase y by 10 before drawing the next line.
}
CHAPTER 3. CONTROL 130
Alternatively, we could use y itself as the loop control variable, noting that the value of y for
the last line is 140:
int y;
for ( y = 50; y <= 140; y = y + 10 )
g.strokeLine( 100, y, 300, y );
If we wanted the lines to be blue, we could do that by calling g.setStroke(Color.BLUE)
before drawing them. If we just draw the lines without setting the color, they will be black.
If we wanted the lines to be 3 pixels wide, we could call g.setLineWidth(3) before drawing
the lines.
For something a little more complicated, let’s draw a large number of randomly colored,
randomly positioned, filled circles. Since we only know a few colors, I will randomly select the
color to be red, green, blue, or yellow. That can be done with a simple switch statement, similar
to the ones in Subsection 3.6.4:
switch ( (int)(4*Math.random()) ) {
case 0:
g.setFill( Color.RED );
break;
case 1:
g.setFill( Color.GREEN );
break;
case 2:
g.setFill( Color.BLUE );
break;
case 3:
g.setFill( Color.YELLOW );
break;
}
I will choose the center points of the circles at random. Let’s say that the width of the
drawing area is given by a variable, width. Then we want a random value in the range 0 to
width-1 for the horizontal position of the center. Similarly, the vertical position of the center
will be a random value in the range 0 to height-1. That leaves the size of the circle to be
determined; I will make the radius of each circle equal to 50 pixels. We can draw the circle with
a statement of the form g.fillOval(x,y,w,h). However, in this command, x and y are not
the coordinates of the center of the circle; they are the upper left corner of a rectangle drawn
around the circle. To get values for x and y, we have to move back from the center of the circle
by 50 pixels, an amount equal to the radius of the circle. The parameters w and h give the
width and height of the rectangle, which have to be twice the radius, or 100 pixels in this case.
Taking all this into account, here is a code segment for drawing a random circle:
centerX = (int)(width*Math.random());
centerY = (int)(height*Math.random());
g.fillOval( centerX - 50, centerY - 50, 100, 100 );
This code comes after the color-setting code given above. In the end, I found that the picture
looks better if I also draw a black outline around each filled circle, so I added this code at the
end:
g.setStroke( Color.BLACK );
g.strokeOval( centerX - 50, centerY - 50, 100, 100 );
CHAPTER 3. CONTROL 131
Finally, to get a large number of circles, I put all of the above code into a for loop that runs
for 500 executions. Here’s a typical drawing from the program, shown at reduced size:
g.setStroke(Color.BLACK);
g.strokeOval( centerX - 50, centerY - 50, 100, 100 );
}
} // end drawPicture()
This is the first subroutine definition that you have seen, other than main(), but you will
learn all about defining subroutines in the next chapter. The first line of the definition makes
available certain values that are used in the subroutine: the graphics context g and the width
and height of the drawing area. These values come from outside the subroutine, but the
subroutine can use them. The point here is that to draw something, you just have to fill in the
inside of the subroutine, just as you write a program by filling in the inside of main().
The subroutine definition still has to go inside a class that defines the program. In this case,
the class is named SimpleGraphicsStarter, and the complete program is available in the sample
source code file SimpleGraphicsStarter.java. You can run that program to see the drawing. You
can use this sample program as a starting point for drawing your own pictures.
There’s a lot in the program that you won’t understand. To make your own drawing, all you
have to do is erase the inside of the drawPicture() routine in the source code and substitute
your own drawing code. You don’t need to understand the rest.
(By the way, you might notice that the main() subroutine uses the word static in its
definition, but drawPicture() does not. This has to do with the fact that drawPicture is a
subroutine in an object rather than in a class. The difference between static and non-static
subroutines is important but not something that we need to worry about for the time being. It
will become important for us in Chapter 5.)
3.9.3 Animation
We can extend the idea of drawing pictures to making animations. A computer animation is
simply a sequence of individual pictures, displayed quickly one after the other. If the change
from each picture to the next is small, the user will perceive the sequence of images as a
continuous animation. Each picture in the animation is called a frame. The sample program
SimpleAnimationStarter.java can be used as a starting point for writing animations. It contains
a subroutine named drawFrame() that draws one frame in an animation. You can create an
animation by filling in the definition of this subroutine. In addition to the graphics context
and the width and height of the drawing area, you can use the value of two other variables in
your code: frameNumber and elapsedSeconds. The drawFrame subroutine will automatically
be called about 60 times per second. The variable frameNumber takes on the values 0, 1, 2,
3, . . . in successive calls to the subroutine, and the value of elapsedSeconds is the number of
seconds that the animation has been running. By using either of these variables in your code,
you can draw a different picture each time drawFrame() is called, and the user will see the
series of pictures as an animation.
As an example of animation, we look at drawing a set of nested rectangles. The rectangles
will shrink towards the center of the drawing, giving an illusion of infinite motion. One frame
from the animation looks like this:
CHAPTER 3. CONTROL 133
Consider how to draw a picture like this one. The rectangles can be drawn with a while
loop, which draws the rectangles starting from the one on the outside and moving in. Think
about what variables will be needed and how they change from one iteration of the while loop to
the next. Each time through the loop, the rectangle that is drawn is smaller than the previous
one and is moved down and over a bit. The difference between two rectangles is in their sizes
and in the coordinates of their upper left corners. We need variables to represent the width
and height of the rectangle, which I call rectWidth and rectHeight. The x and y-coordinates
of the upper left corner are the same, and they can be represented by the same variable. I call
that variable inset, since it is the amount by which the edges of the rectangle are inset from
the edges of the drawing area. The width and height decrease from one rectangle to the next,
while the inset increases. The while loop ends when either the width or the height becomes
less than or equal to zero. In general outline, the algorithm for drawing one frame is
Fill the drawing area with white
Set the amount of inset for the first rectangle
Set the width and height for the first rectangle
Set the stroke color to black
while the width and height are both greater than zero:
draw a rectangle (using the g.strokeRect subroutine)
increase the inset (to move the next rectangle over and down)
decrease the width and height (to make the next rectangle smaller)
In my program, each rectangle is 15 pixels away from the rectangle that surrounds it, so the
inset is increased by 15 each time through the while loop. The rectangle shrinks by 15 pixels
on the left and by 15 pixels on the right, so the width of the rectangle shrinks by 30 before
drawing the next rectangle. The height also shrinks by 30 pixels each time through the loop.
The pseudocode is then easy to translate into Java, except that we need to know what
initial values to use for the inset, width, and height of the first rectangle. To figure that out,
we have to think about the fact that the picture is animated, so that what we draw will depend
in some way on the frame number. From one frame to the next frame of the animation, the
top-left corner of the outer rectangle moves over and down; that is, the inset for the outer
rectangle increases from one frame to the next. We can make this happen by setting the inset
for frame number 0 to 0, the inset for frame number 1 to 1, and so on. But that can’t go on
forever, or eventually all the rectangles would disappear. In fact, when the animation gets to
frame 15, a new rectangle should appear at the outside of the drawing area—but it’s not really
a “new rectangle,” it’s just that the inset for the outer rectangle goes back to zero. So, as the
CHAPTER 3. CONTROL 134
animation proceeds, the inset should go through the sequence of values 0, 1, 2, . . . , 14 over and
over. We can accomplish that very easily by setting
inset = frameNumber % 15;
Finally, note that the first rectangle fills the drawing area except for a border of size inset
around the outside of the rectangle. This means that the width of the rectangle is the width
of the drawing area minus two times the inset, and similarly for the height. Here, then is the
drawFrame() subroutine for the moving rectangle example:
public void drawFrame(GraphicsContext g, int frameNumber,
double elapsedSeconds, int width, int height) {
g.setFill(Color.WHITE);
g.fillRect(0,0,width,height); // Fill drawing area with white.
double inset; // Gap between edges of drawing area and outer rectangle.
double rectWidth, rectHeight; // The size of one of the rectangles.
g.setStroke(Color.BLACK); // Draw the rectangle outlines in black.
inset = frameNumber % 15 + 0.5; // (The 0.5 is a technicality that gives
// a sharper picture.)
rectWidth = width - 2*inset;
rectHeight = height - 2*inset;
while (rectWidth >= 0 && rectHeight >= 0) {
g.strokeRect(inset, inset, rectWidth, rectHeight);
inset += 15; // rectangles are 15 pixels apart
rectWidth -= 30;
rectHeight -= 30;
}
}
You can find the full source code for the program is in the sample program MovingRects.java.
Take a look! It’s a neat effect. For another example of animation, see the sample program
RandomCircles.java. That program adds one random colored disk to the picture in each frame;
it illustrates the fact that the image from one frame is not automatically erased before the next
frame is drawn.
Exercises 135
1. How many times do you have to roll a pair of dice before they come up snake eyes? You (solution)
could do the experiment by rolling the dice by hand. Write a computer program that
simulates the experiment. The program should report the number of rolls that it makes
before the dice come up snake eyes. (Note: “Snake eyes” means that both dice show a
value of 1.) Exercise 2.2 explained how to simulate rolling a pair of dice.
2. Which integer between 1 and 10000 has the largest number of divisors, and how many (solution)
divisors does it have? Write a program to find the answers and print out the results. It is
possible that several integers in this range have the same, maximum number of divisors.
Your program only has to print out one of them. An example in Subsection 3.4.2 discussed
divisors. The source code for that example is CountDivisors.java.
You might need some hints about how to find a maximum value. The basic idea is
to go through all the integers, keeping track of the largest number of divisors that you’ve
seen so far. Also, keep track of the integer that had that number of divisors.
3. Write a program that will evaluate simple expressions such as 17 + 3 and 3.14159 * 4.7. (solution)
The expressions are to be typed in by the user. The input always consists of a number,
followed by an operator, followed by another number. The operators that are allowed are
+, -, *, and /. You can read the numbers with TextIO.getDouble() and the operator
with TextIO.getChar(). Your program should read an expression, print its value, read
another expression, print its value, and so on. The program should end when the user
enters 0 as the first number on the line.
4. Write a program that reads one line of input text and breaks it up into words. The (solution)
words should be output one per line. A word is defined to be a sequence of letters. Any
characters in the input that are not letters should be discarded. For example, if the user
inputs the line
He said, "That’s not a good idea."
then the output of the program should be
He
said
That
s
not
a
good
idea
An improved version of the program would list “that’s” as a single word. An apostrophe
can be considered to be part of a word if there is a letter on each side of the apostrophe.
To test whether a character is a letter, you might use (ch >= ’a’ && ch <= ’z’) ||
(ch >= ’A’ && ch <= ’Z’). However, this only works in English and similar languages.
A better choice is to call the standard function Character.isLetter(ch), which returns
a boolean value of true if ch is a letter and false if it is not. This works for any Unicode
character.
Exercises 136
5. Suppose that a file contains information about sales figures for a company in various cities. (solution)
Each line of the file contains a city name, followed by a colon (:) followed by the data for
that city. The data is a number of type double. However, for some cities, no data was
available. In these lines, the data is replaced by a comment explaining why the data is
missing. For example, several lines from the file might look like:
San Francisco: 19887.32
Chicago: no report received
New York: 298734.12
Write a program that will compute and print the total sales from all the cities together.
The program should also report the number of cities for which data was not available.
The name of the file is “sales.dat”.
To complete this program, you’ll need one fact about file input with TextIO that was
not covered in Subsection 2.4.4. Since you don’t know in advance how many lines there
are in the file, you need a way to tell when you have gotten to the end of the file. When
TextIO is reading from a file, the function TextIO.eof() can be used to test for end of
file. This boolean-valued function returns true if the file has been entirely read and
returns false if there is more data to read in the file. This means that you can read the
lines of the file in a loop while (TextIO.eof() == false).... The loop will end when
all the lines of the file have been read.
Suggestion: For each line, read and ignore characters up to the colon. Then read the
rest of the line into a variable of type String. Try to convert the string into a number, and
use try..catch to test whether the conversion succeeds.
6. Exercise 3.2 asked you to find the number in the range 1 to 10000 that has the largest (solution)
number of divisors. You only had to print out one such number. Revise the program so
that it will print out all numbers that have the maximum number of divisors. Use an array
as follows: As you count the divisors for each number, store each count in an array. Then
at the end of the program, you can go through the array and print out all the numbers
that have the maximum count. The output from the program should look something like
this:
Among integers between 1 and 10000,
The maximum number of divisors was 64
Numbers with that many divisors include:
7560
9240
7. An example in Subsection 3.8.3 tried to answer the question, How many random people do (solution)
you have to select before you find a duplicate birthday? The source code for that program
can be found in the file BirthdayProblem.java. Here are some related questions:
• How many random people do you have to select before you find three people who
share the same birthday? (That is, all three people were born on the same day in
the same month, but not necessarily in the same year.)
• Suppose you choose 365 people at random. How many different birthdays will they
have? (The number could theoretically be anywhere from 1 to 365).
• How many different people do you have to check before you’ve found at least one
person with a birthday on each of the 365 days of the year?
Exercises 137
Write three programs to answer these questions. Each of your programs should sim-
ulate choosing people at random and checking their birthdays. (In each case, ignore the
possibility of leap years.)
8. Write a GUI program that draws a checkerboard. Base your solution on the sample pro- (solution)
gram SimpleGraphicsStarter.java You will draw the checkerboard in the drawPicture()
subroutine, after erasing the code that it already contains.
The checkerboard should be 400-by-400 pixels. You can change the size of the drawing
area in SimpleGraphicsStarter.java by modifying the first two lines of the start()
subroutine to set width and height to 400 instead of 800 and 600. A checkerboard
contains 8 rows and 8 columns of squares. If the size of the drawing area is 400, that
means that each square should be 50-by-50 pixels. The squares are red and black (or
whatever other colors you choose). Here is a tricky way to determine whether a given
square should be red or black: The rows and columns can be thought of as numbered from
0 to 7. If the row number of the square and the column number of the square are either
both even or both odd, then the square is red. Otherwise, it is black. Note that a square
is just a rectangle in which the height is equal to the width, so you can use the subroutine
g.fillRect() to draw the squares. Here is a reduced-size image of the checkerboard that
you want to draw:
9. Often, some element of an animation repeats over and over, every so many frames. Some- (solution)
times, the repetition is “cyclic,” meaning that at the end it jumps back to the start.
Sometimes the repetition is “oscillating,” like a back-and-forth motion where the second
half is the same as the first half played in reverse.
Write an animation that demonstrates both cyclic and oscillating motions at various
speeds. For cyclic motion, you can use a square that moves across the drawing area, then
jumps back to the start, and then repeats the same motion over and over. For oscillating
motion, you can do something similar, but the square should move back and forth between
the two edges of the drawing area; that is, it moves left-to-right during the first half of
the animation and then backwards from right-to-left during the second half. To write the
program, you can start with a copy of the sample program SimpleAnimationStarter.java.
A cyclic motion has to repeat every N frames for some value of N. What you draw in
some frame of the animation depends on the frameNumber. The frameNumber just keeps
increasing forever. To implement cyclic motion, what you really want is a “cyclic frame
Exercises 138
Quiz on Chapter 3
(answers)
1. What is an algorithm?
2. Explain briefly what is meant by “pseudocode” and how is it useful in the development
of algorithms.
3. What is a block statement? How are block statements used in Java programs?
4. What is the main difference between a while loop and a do..while loop?
7. Write a for loop that will print out all the multiples of 3 from 3 to 36, that is: 3 6 9 12
15 18 21 24 27 30 33 36.
8. Fill in the following main() routine so that it will ask the user to enter an integer, read
the user’s response, and tell the user whether the number entered is even or odd. (You can
use TextIO.getInt() to read the integer. Recall that an integer n is even if n % 2 == 0.)
public static void main(String[] args) {
// Fill in the body of this subroutine!
}
9. Write a code segment that will print out two different random integers selected from the
range 1 to 10. All possible outputs should have the same probability. Hint: You can easily
select two random numbers, but you have to account for the fact that the two numbers
that you pick might be the same.
10. Suppose that s1 and s2 are variables of type String, whose values are expected to be
string representations of values of type int. Write a code segment that will compute and
print the integer sum of those values, or will print an error message if the values cannot
successfully be converted into integers. (Use a try..catch statement.)
11. Show the exact output that would be produced by the following main() routine:
public static void main(String[] args) {
int N;
N = 1;
while (N <= 32) {
N = 2 * N;
System.out.println(N);
}
}
12. Show the exact output produced by the following main() routine:
Quiz 140
13. What output is produced by the following program segment? Why? (Recall that
name.charAt(i) is the i-th character in the string, name.)
String name;
int i;
boolean startWord;
name = "Richard M. Nixon";
startWord = true;
for (i = 0; i < name.length(); i++) {
if (startWord)
System.out.println(name.charAt(i));
if (name.charAt(i) == ’ ’)
startWord = true;
else
startWord = false;
}
14. Suppose that numbers is an array of type int[ ]. Write a code segment that will count and
output the number of times that the number 42 occurs in the array.
15. Define the range of an array of numbers to be the maximum value in the array minus
the minimum value. Suppose that raceTimes is an array of type double[ ]. Write a code
segment that will find and print the range of raceTimes.
Chapter 4
One way to break up a complex program into manageable pieces is to use subroutines.
A subroutine consists of the instructions for carrying out a certain task, grouped together and
given a name. Elsewhere in the program, that name can be used as a stand-in for the whole set
of instructions. As a computer executes a program, whenever it encounters a subroutine name,
it executes all the instructions necessary to carry out the task associated with that subroutine.
Subroutines can be used over and over, at different places in the program. A subroutine
can even be used inside another subroutine. This allows you to write simple subroutines and
then use them to help write more complex subroutines, which can then be used in turn in other
subroutines. In this way, very complex programs can be built up step-by-step, where each step
in the construction is reasonably simple.
Subroutines in Java can be either static or non-static. This chapter covers static subroutines.
Non-static subroutines, which are used in true object-oriented programming, will be covered in
the next chapter.
141
CHAPTER 4. SUBROUTINES 142
The contract of a subroutine says, essentially, “Here is what you have to do to use me,
and here is what I will do for you, guaranteed.” When you write a subroutine, the comments
that you write for the subroutine should make the contract very clear. (I should admit that
in practice, subroutines’ contracts are often inadequately specified, much to the regret and
annoyance of the programmers who have to use them.)
For the rest of this chapter, I turn from general ideas about black boxes and subroutines
in general to the specifics of writing and using subroutines in Java. But keep the general ideas
and principles in mind. They are the reasons that subroutines exist in the first place, and they
are your guidelines for using them. This should be especially clear in Section 4.7, where I will
discuss subroutines as a tool in program development.
∗ ∗ ∗
You should keep in mind that subroutines are not the only example of black boxes in
programming. For example, a class is also a black box. We’ll see that a class can have a
“public” part, representing its interface, and a “private” part that is entirely inside its hidden
implementation. All the principles of black boxes apply to classes as well as to subroutines.
system when the system runs your program.) For example, the playGame() method given as
an example above could be called using the following subroutine call statement:
playGame();
This statement could occur anywhere in the same class that includes the definition of
playGame(), whether in a main() method or in some other subroutine. Since playGame()
is a public method, it can also be called from other classes, but in that case, you have to tell
the computer which class it comes from. Since playGame() is a static method, its full name
includes the name of the class in which it is defined. Let’s say, for example, that playGame() is
defined in a class named Poker. Then to call playGame() from outside the Poker class, you
would have to say
Poker.playGame();
The use of the class name here tells the computer which class to look in to find the method. It
also lets you distinguish between Poker.playGame() and other potential playGame() methods
defined in other classes, such as Roulette.playGame() or Blackjack.playGame().
More generally, a subroutine call statement for a static subroutine takes the form
hsubroutine-name i(hparameters i);
if the subroutine that is being called is in the same class, or
hclass-name i.hsubroutine-name i(hparameters i);
if the subroutine is defined elsewhere, in a different class. (Non-static methods belong to objects
rather than classes, and they are called using objects instead of class names. More on that later.)
Note that the parameter list can be empty, as in the playGame() example, but the parentheses
must be there even if there is nothing between them. The number of parameters that you
provide when you call a subroutine must match the number specified in the parameter list in
the subroutine definition, and the types of the parameters in the call statement must match
the types in the subroutine definition.
The test for whether the game is over is complicated, since the game ends if either the user
makes a correct guess or the number of guesses is six. As in many cases, the easiest thing to
do is to use a “while (true)” loop and use break to end the loop whenever we find a reason
to do so. Also, if we are going to end the game after six guesses, we’ll have to keep track of the
number of guesses that the user has made. Filling out the algorithm gives:
Let computersNumber be a random number between 1 and 100
Let guessCount = 0
while (true):
Get the user’s guess
Count the guess by adding 1 to guess count
if the user’s guess equals computersNumber:
Tell the user he won
break out of the loop
if the number of guesses is 6:
Tell the user he lost
break out of the loop
if the user’s guess is less than computersNumber:
Tell the user the guess was low
else if the user’s guess is higher than computersNumber:
Tell the user the guess was high
With variable declarations added and translated into Java, this becomes the definition of the
playGame() routine. A random integer between 1 and 100 can be computed as (int)(100 *
Math.random()) + 1. I’ve cleaned up the interaction with the user to make it flow better.
static void playGame() {
int computersNumber; // A random number picked by the computer.
int usersGuess; // A number entered by user as a guess.
int guessCount; // Number of guesses the user has made.
computersNumber = (int)(100 * Math.random()) + 1;
// The value assigned to computersNumber is a randomly
// chosen integer between 1 and 100, inclusive.
guessCount = 0;
System.out.println();
System.out.print("What is your first guess? ");
while (true) {
usersGuess = TextIO.getInt(); // Get the user’s guess.
guessCount++;
if (usersGuess == computersNumber) {
System.out.println("You got it in " + guessCount
+ " guesses! My number was " + computersNumber);
break; // The game is over; the user has won.
}
if (guessCount == 6) {
System.out.println("You didn’t get the number in 6 guesses.");
System.out.println("You lose. My number was " + computersNumber);
break; // The game is over; the user has lost.
}
// If we get to this point, the game continues.
// Tell the user if the guess was too high or too low.
if (usersGuess < computersNumber)
System.out.print("That’s too low. Try again: ");
else if (usersGuess > computersNumber)
CHAPTER 4. SUBROUTINES 148
// Tell the user if the guess was too high or too low.
if (usersGuess < computersNumber)
System.out.print("That’s too low. Try again: ");
else if (usersGuess > computersNumber)
System.out.print("That’s too high. Try again: ");
}
System.out.println();
} // end of playGame()
} // end of class GuessingGame
Take some time to read the program carefully and figure out how it works. And try to
convince yourself that even in this relatively simple case, breaking up the program into two
methods makes the program easier to understand and probably made it easier to write each
piece.
named Poker, then code in the Poker class would refer to it simply as numberOfPlayers, while
code in another class would refer to it as Poker.numberOfPlayers.
As an example, let’s add a couple of static member variables to the GuessingGame class that
we wrote earlier in this section. We add a variable named gamesPlayed to keep track of how
many games the user has played and another variable named gamesWon to keep track of the
number of games that the user has won. The variables are declared as static member variables:
static int gamesPlayed;
static int gamesWon;
In the playGame() routine, we always add 1 to gamesPlayed, and we add 1 to gamesWon if the
user wins the game. At the end of the main() routine, we print out the values of both variables.
It would be impossible to do the same thing with local variables, since both subroutines need
to access the variables, and local variables exist in only one subroutine. Furthermore, global
variables keep their values between one subroutine call and the next. Local variables do not; a
local variable gets a new value each time that the subroutine that contains it is called.
When you declare a local variable in a subroutine, you have to assign a value to that variable
before you can do anything with it. Member variables, on the other hand are automatically
initialized with a default value. The default values are the same as those that are used when
initializing the elements of an array: For numeric variables, the default value is zero; for boolean
variables, the default is false; for char variables, it’s the character that has Unicode code
number zero; and for objects, such as Strings, the default initial value is the special value null.
Since they are of type int, the static member variables gamesPlayed and gamesWon au-
tomatically get zero as their initial value. This happens to be the correct initial value for a
variable that is being used as a counter. You can, of course, assign a value to a variable at the
beginning of the main() routine if you are not satisfied with the default initial value, or if you
want to make the initial value more explicit.
Here’s the revised version of GuessingGame.java. The changes from the above version are
shown in italic:
import textio.TextIO;
public class GuessingGame2 {
static int gamesPlayed; // The number of games played.
static int gamesWon; // The number of games won.
public static void main(String[] args) {
gamesPlayed = 0;
gamesWon = 0; // This is actually redundant, since 0 is
// the default initial value.
System.out.println("Let’s play a game. I’ll pick a number between");
System.out.println("1 and 100, and you try to guess it.");
boolean playAgain;
do {
playGame(); // call subroutine to play one game
System.out.print("Would you like to play again? ");
playAgain = TextIO.getlnBoolean();
} while (playAgain);
System.out.println();
System.out.println("You played " + gamesPlayed + " games,");
System.out.println("and you won " + gamesWon + " of those games.");
System.out.println("Thanks for playing. Goodbye.");
CHAPTER 4. SUBROUTINES 151
} // end of main()
static void playGame() {
int computersNumber; // A random number picked by the computer.
int usersGuess; // A number entered by user as a guess.
int guessCount; // Number of guesses the user has made.
gamesPlayed++; // Count this game.
computersNumber = (int)(100 * Math.random()) + 1;
// The value assigned to computersNumber is a randomly
// chosen integer between 1 and 100, inclusive.
guessCount = 0;
System.out.println();
System.out.print("What is your first guess? ");
while (true) {
usersGuess = TextIO.getInt(); // Get the user’s guess.
guessCount++;
if (usersGuess == computersNumber) {
System.out.println("You got it in " + guessCount
+ " guesses! My number was " + computersNumber);
gamesWon++; // Count this win.
break; // The game is over; the user has won.
}
if (guessCount == 6) {
System.out.println("You didn’t get the number in 6 guesses.");
System.out.println("You lose. My number was " + computersNumber);
break; // The game is over; the user has lost.
}
// If we get to this point, the game continues.
// Tell the user if the guess was too high or too low.
if (usersGuess < computersNumber)
System.out.print("That’s too low. Try again: ");
else if (usersGuess > computersNumber)
System.out.print("That’s too high. Try again: ");
}
System.out.println();
} // end of playGame()
} // end of class GuessingGame2
∗ ∗ ∗
(By the way, notice that in my example programs, I didn’t mark the static subroutines or
variables as being public or private. You might wonder what it means to leave out both
modifiers. Recall that global variables and subroutines with no access modifier can be used
anywhere in the same package as the class where they are defined, but not in other packages.
Classes that don’t declare a package are in the default package. So, any class in the default
package would have access to gamesPlayed, gamesWon, and playGame()—and that includes
most of the classes in this book. In fact, it is considered to be good practice to make member
variables and subroutines private, unless there is a reason for doing otherwise. (But then
again, it’s also considered good practice to avoid using the default package.))
CHAPTER 4. SUBROUTINES 152
4.3 Parameters
If a subroutine is a black box, then a parameter is something that provides a mechanism
for passing information from the outside world into the box. Parameters are part of the interface
of a subroutine. They allow you to customize the behavior of a subroutine to adapt it to a
particular situation.
As an analogy, consider a thermostat—a black box whose task it is to keep your house
at a certain temperature. The thermostat has a parameter, namely the dial that is used to
set the desired temperature. The thermostat always performs the same task: maintaining a
constant temperature. However, the exact task that it performs—that is, which temperature
it maintains—is customized by the setting on its dial.
that the expression computes a value of the correct type. The type of the actual parameter must
be one that could legally be assigned to the formal parameter with an assignment statement.
For example, if the formal parameter is of type double, then it would be legal to pass an int as
the actual parameter since ints can legally be assigned to doubles. When you call a subroutine,
you must provide one actual parameter for each formal parameter in the subroutine’s definition.
Consider, for example, a subroutine
static void doTask(int N, double x, boolean test) {
// statements to perform the task go here
}
This subroutine might be called with the statement
doTask(17, Math.sqrt(z+1), z >= 10);
When the computer executes this statement, it has essentially the same effect as the block of
statements:
{
int N; // Allocate memory locations for the formal parameters.
double x;
boolean test;
N = 17; // Assign 17 to the first formal parameter, N.
x = Math.sqrt(z+1); // Compute Math.sqrt(z+1), and assign it to
// the second formal parameter, x.
test = (z >= 10); // Evaluate "z >= 10" and assign the resulting
// true/false value to the third formal
// parameter, test.
// statements to perform the task go here
}
(There are a few technical differences between this and “doTask(17,Math.sqrt(z+1),z>=10);”
—besides the amount of typing—because of questions about scope of variables and what hap-
pens when several variables or parameters have the same name.)
Beginning programming students often find parameters to be surprisingly confusing. Call-
ing a subroutine that already exists is not a problem—the idea of providing information to the
subroutine in a parameter is clear enough. Writing the subroutine definition is another matter.
A common beginner’s mistake is to assign values to the formal parameters at the beginning
of the subroutine, or to ask the user to input their values. This represents a fundamen-
tal misunderstanding. By the time the computer starts executing the statements in the
subroutine, the formal parameters have already been assigned initial values! The computer
automatically assigns values to the formal parameters before it starts executing the code inside
the subroutine. The values come from the actual parameters in the subroutine call statement.
Remember that a subroutine is not independent. It is called by some other routine, and it is
the subroutine call statement’s responsibility to provide appropriate values for the parameters.
4.3.3 Overloading
In order to call a subroutine legally, you need to know its name, you need to know how many
formal parameters it has, and you need to know the type of each parameter. This information is
called the subroutine’s signature. The signature of the subroutine doTask, used as an example
above, can be expressed as: doTask(int,double,boolean). Note that the signature does not
include the names of the parameters; in fact, if you just want to use the subroutine, you don’t
CHAPTER 4. SUBROUTINES 155
even need to know what the formal parameter names are, so the names are not part of the
interface.
Java is somewhat unusual in that it allows two different subroutines in the same class to
have the same name, provided that their signatures are different. When this happens, we say
that the name of the subroutine is overloaded because it has several different meanings. The
computer doesn’t get the subroutines mixed up. It can tell which one you want to call by the
number and types of the actual parameters that you provide in the subroutine call statement.
You have already seen overloading used with System.out. This object includes many different
methods named println, for example. These methods all have different signatures, such as:
println(int) println(double)
println(char) println(boolean)
println()
The computer knows which of these subroutines you want to use based on the type of the
actual parameter that you provide. System.out.println(17) calls the subroutine with sig-
nature println(int), while System.out.println(’A’) calls the subroutine with signature
println(char). Of course all these different subroutines are semantically related, which is
why it is acceptable programming style to use the same name for them all. But as far as the
computer is concerned, printing out an int is very different from printing out a char, which is
different from printing out a boolean, and so forth—so that each of these operations requires
a different subroutine.
Note, by the way, that the signature does not include the subroutine’s return type. It is
illegal to have two subroutines in the same class that have the same signature but that have
different return types. For example, it would be a syntax error for a class to contain two
subroutines defined as:
int getln() { ... }
double getln() { ... }
This is why in the TextIO class, the subroutines for reading different types are not all named
getln(). In a given class, there can only be one routine that has the name getln with no
parameters. So, the input routines in TextIO are distinguished by having different names, such
as getlnInt() and getlnDouble().
in the body of the subroutine should do. Since we are only working with static subroutines
for now, we’ll need to use static as a modifier. We could add an access modifier (public or
private), but in the absence of any instructions, I’ll leave it out. Since we are not told to
return a value, the return type is void. Since no names are specified, we’ll have to make up
names for the formal parameter and for the subroutine itself. I’ll use N for the parameter and
printDivisors for the subroutine name. The subroutine will look like
static void printDivisors( int N ) {
hstatements i
}
and all we have left to do is to write the statements that make up the body of the routine. This
is not difficult. Just remember that you have to write the body assuming that N already has
a value! The algorithm is: “For each possible divisor D in the range from 1 to N, if D evenly
divides N, then print D.” Written in Java, this becomes:
/**
* Print all the divisors of N.
* We assume that N is a positive integer.
*/
static void printDivisors( int N ) {
int D; // One of the possible divisors of N.
System.out.println("The divisors of " + N + " are:");
for ( D = 1; D <= N; D++ ) {
if ( N % D == 0 ) // Does D evenly divide N?
System.out.println(D);
}
}
I’ve added a comment before the subroutine definition indicating the contract of the
subroutine—that is, what it does and what assumptions it makes. The contract includes the
assumption that N is a positive integer. It is up to the caller of the subroutine to make sure
that this assumption is satisfied.
As a second short example, consider the problem: Write a private subroutine named
printRow. It should have a parameter ch of type char and a parameter N of type int. The
subroutine should print out a line of text containing N copies of the character ch.
Here, we are told the name of the subroutine and the names of the two parameters, and we
are told that the subroutine is private, so we don’t have much choice about the first line of
the subroutine definition. The task in this case is pretty simple, so the body of the subroutine
is easy to write. The complete subroutine is given by
/**
* Write one line of output containing N copies of the
* character ch. If N <= 0, an empty line is output.
*/
private static void printRow( char ch, int N ) {
int i; // Loop-control variable for counting off the copies.
for ( i = 1; i <= N; i++ ) {
System.out.print( ch );
}
System.out.println();
}
CHAPTER 4. SUBROUTINES 157
Note that in this case, the contract makes no assumption about N, but it makes it clear what
will happen in all cases, including the unexpected case that N <= 0.
Finally, let’s do an example that shows how one subroutine can build on another. Let’s write
a subroutine that takes a String as a parameter. For each character in the string, it should
print a line of output containing 25 copies of that character. It should use the printRow()
subroutine to produce the output.
Again, we get to choose a name for the subroutine and a name for the parameter. I’ll call
the subroutine printRowsFromString and the parameter str. The algorithm is pretty clear:
For each position i in the string str, call printRow(str.charAt(i),25) to print one line of
the output. So, we get:
/**
* For each character in str, write a line of output
* containing 25 copies of that character.
*/
private static void printRowsFromString( String str ) {
int i; // Loop-control variable for counting off the chars.
for ( i = 0; i < str.length(); i++ ) {
printRow( str.charAt(i), 25 );
}
}
We could then use printRowsFromString in a main() routine such as
public static void main(String[] args) {
String inputLine; // Line of text input by user.
System.out.print("Enter a line of text: ");
inputLine = TextIO.getln();
System.out.println();
printRowsFromString( inputLine );
}
Of course, the three routines, main(), printRowsFromString(), and printRow(), would
have to be collected together inside the same class. The program is rather useless, but it does
demonstrate the use of subroutines. You’ll find the program in the file RowsOfChars.java, if
you want to take a look.
}
To use this subroutine, you need an actual array. Here is a legal, though not very realistic, code
segment that creates an array just to pass it as an argument to the subroutine:
int[] numbers;
numbers = new int[3];
numbers[0] = 42;
numbers[1] = 17;
numbers[2] = 256;
printValuesInList( numbers );
The output produced by the last statement would be [42,17,256].
by copying one line at a time from the original file to the copy, using TextIO. The function
TextIO.eof() is a boolean-valued function that is true if the end of the file has been reached.
input textio.TextIO;
/**
* Requires two command line arguments, which must be file names. The
* first must be the name of an existing file. The second is the name
* of a file to be created by the program. The contents of the first file
* are copied into the second. WARNING: If the second file already
* exists when the program is run, its previous contents will be lost!
* This program only works for plain text files.
*/
public class CopyTextFile {
public static void main( String[] args ) {
if (args.length < 2 ) {
System.out.println("Two command-line arguments are required!");
System.exit(1);
}
TextIO.readFile( args[0] ); // Open the original file for reading.
TextIO.writeFile( args[1] ); // Open the copy file for writing.
int lineCount; // Number of lines copied
lineCount = 0;
while ( TextIO.eof() == false ) {
// Read one line from the original file and write it to the copy.
String line;
line = TextIO.getln();
TextIO.putln(line);
lineCount++;
}
System.out.printf( "%d lines copied from %s to %s%n",
lineCount, args[0], args[1] );
}
}
Since most programs are run in a GUI environment these days, command-line arguments
aren’t as important as they used to be. But at least they provide a nice example of how array
parameters can be used.
uses the global variable to communicate with the rest of the program. This is a kind of sneaky,
back-door communication that is less visible than communication done through parameters,
and it risks violating the rule that the interface of a black box should be straightforward and
easy to understand. So before you use a global variable in a subroutine, you should consider
whether it’s really necessary.
I don’t advise you to take an absolute stand against using global variables inside subroutines.
There is at least one good reason to do it: If you think of the class as a whole as being a kind
of black box, it can be very reasonable to let the subroutines inside that box be a little sneaky
about communicating with each other, if that will make the class as a whole look simpler from
the outside.
else
answer = currentN / 2; // if not, this is the answer
return answer; // (Don’t forget to return the answer!)
}
Here is a subroutine that uses this nextN function. In this case, the improvement from the
version of the subroutine in Section 4.3 is not great, but if nextN() were a long function that
performed a complex computation, then it would make a lot of sense to hide that complexity
inside a function:
static void print3NSequence(int startingValue) {
int N; // One of the terms in the sequence.
int count; // The number of terms found.
N = startingValue; // Start the sequence with startingValue.
count = 1;
System.out.println("The 3N+1 sequence starting from " + N);
System.out.println();
System.out.println(N); // print initial term of sequence
while (N > 1) {
N = nextN( N ); // Compute next term, using the function nextN.
count++; // Count this term.
System.out.println(N); // Print this term.
}
System.out.println();
System.out.println("There were " + count + " terms in the sequence.");
}
∗ ∗ ∗
Here are a few more examples of functions. The first one computes a letter grade corre-
sponding to a given numerical grade, on a typical grading scale:
/**
* Returns the letter grade corresponding to the numerical
* grade that is passed to this function as a parameter.
*/
static char letterGrade(int numGrade) {
if (numGrade >= 90)
return ’A’; // 90 or above gets an A
else if (numGrade >= 80)
return ’B’; // 80 to 89 gets a B
else if (numGrade >= 65)
return ’C’; // 65 to 79 gets a C
else if (numGrade >= 50)
return ’D’; // 50 to 64 gets a D
else
return ’F’; // anything else gets an F
} // end of function letterGrade
CHAPTER 4. SUBROUTINES 164
The type of the return value of letterGrade() is char. Functions can return values of any
type at all. Here’s a function whose return value is of type boolean. It demonstrates some
interesting programming points, so you should read the comments:
/**
* This function returns true if N is a prime number. A prime number
* is an integer greater than 1 that is not divisible by any positive
* integer, except itself and 1. If N has any divisor, D, in the range
* 1 < D < N, then it has a divisor in the range 2 to Math.sqrt(N), namely
* either D itself or N/D. So we only test possible divisors from 2 to
* Math.sqrt(N).
*/
static boolean isPrime(int N) {
int divisor; // A number we will test to see whether it evenly divides N.
if (N <= 1)
return false; // No number <= 1 is a prime.
int maxToTry; // The largest divisor that we need to test.
maxToTry = (int)Math.sqrt(N);
// We will try to divide N by numbers between 2 and maxToTry.
// If N is not evenly divisible by any of these numbers, then
// N is prime. (Note that since Math.sqrt(N) is defined to
// return a value of type double, the value must be typecast
// to type int before it can be assigned to maxToTry.)
for (divisor = 2; divisor <= maxToTry; divisor++) {
if ( N % divisor == 0 ) // Test if divisor evenly divides N.
return false; // If so, we know N is not prime.
// No need to continue testing!
}
// If we get to this point, N must be prime. Otherwise,
// the function would already have been terminated by
// a return statement in the previous loop.
return true; // Yes, N is prime.
} // end of function isPrime
Finally, here is a function with return type String. This function has a String as parameter.
The returned value is a reversed copy of the parameter. For example, the reverse of “Hello
World” is “dlroW olleH”. The algorithm for computing the reverse of a string, str, is to
start with an empty string and then to append each character from str, starting from the last
character of str and working backwards to the first:
static String reverse(String str) {
String copy; // The reversed copy.
int i; // One of the positions in str,
// from str.length() - 1 down to 0.
copy = ""; // Start with an empty string.
for ( i = str.length() - 1; i >= 0; i-- ) {
// Append i-th char of str to copy.
copy = copy + str.charAt(i);
}
CHAPTER 4. SUBROUTINES 165
return copy;
}
A palindrome is a string that reads the same backwards and forwards, such as “radar”. The
reverse() function could be used to check whether a string, word, is a palindrome by testing
“if (word.equals(reverse(word)))”.
By the way, a typical beginner’s error in writing functions is to print out the answer, instead
of returning it. This represents a fundamental misunderstanding. The task of a function
is to compute a value and return it to the point in the program where the function was called.
That’s where the value is used. Maybe it will be printed out. Maybe it will be assigned to a
variable. Maybe it will be used in an expression. But it’s not for the function to decide.
/**
* print3NSequence prints a 3N+1 sequence to standard output, using
* startingValue as the initial value of N. It also prints the number
* of terms in the sequence. The value of the parameter, startingValue,
CHAPTER 4. SUBROUTINES 166
/**
* nextN computes and returns the next term in a 3N+1 sequence,
* given that the current term is currentN.
*/
static int nextN(int currentN) {
if (currentN % 2 == 1)
return 3 * currentN + 1;
else
return currentN / 2;
} // end of nextN()
You should read this program carefully and try to understand how it works.
and ones, a subroutine doesn’t seem all that different from a data value such as, for example,
as an integer, a string, or an array, which is also represented as a string of zeros and ones in
memory. We are used to thinking of subroutines and data as very different things, but inside
the computer, a subroutine is just another kind of data. Some programming languages make
it possible to work with a subroutine as a kind of data value. In Java 8, that ability was added
to Java in the form of something called lambda expressions.
Lambda expressions are becoming more and more common in Java programs. They are
especially useful for working with the JavaFX GUI toolkit, and it will be useful to know about
them before we cover GUI programming in Chapter 6. However, we won’t encounter them
again until near the end of the Chapter 5, so you can skip this section for now if you want.
typed language; to have a variable named sqr, we must declare that variable and give it a type.
But what sort of type would be appropriate for a value that is a function? The answer in Java
is something called a functional interface, which we turn to next.
But first one more note: Lambda expressions in Java can actually represent arbitrary sub-
routines, not just functions. Nevertheless, it is the term “function” that is usually associated
with them, rather than “subroutine” or “method.”
can be omitted, as long as they can be deduced from the context. For example, if the lambda
expression is known to be of type FunctionR2R, then the parameter type must be double, so it
is unnecessary to specify the parameter type in the lambda expression. Next, if there is exactly
one parameter and if its type is not specified, then the parentheses around the parameter list
can be omitted. On the right-hand side of the ->, if the only thing between the braces, { and },
is a single subroutine call statement, then the braces can be omitted. And if the right-hand side
has the form { return hexpressioni; }, then you can omit everything except the hexpressioni.
For example, suppose that we want a lambda expression to represent a function that com-
putes the square of a double value. The type of such a function can be the FunctionR2R interface
given above. If sqr is a variable of type FunctionR2R, then the value of the function can be a
lambda expression, which can be written in any of the following forms:
sqr = (double x) -> { return x*x; };
sqr = (x) -> { return x*x; };
sqr = x -> { return x*x; };
sqr = x -> x*x;
sqr = (double fred) -> fred*fred;
sqr = (z) -> z*z;
The last two statements are there to emphasize that the parameter names in a lambda expres-
sion are dummy parameters; their names are irrelevant. The six lambda expressions in these
statements all define exactly the same function. Note that the parameter type double can
be omitted because the compiler knows that sqr is of type FunctionR2R, and a FunctionR2R
requires a parameter of type double. A lambda expression can only be used in a context where
the compiler can deduce its type, and the parameter type has to be included only in a case
where leaving it out would make the type of the lambda expression ambiguous.
Now, in Java, the variable sqr as defined here is not quite a function. It is a value of
type FunctionR2R, which means that it contains a function named valueAt, as specified in
the definition of interface FunctionR2R. The full name of that function is sqr.valueAt, and we
must use that name to call the function. For example: sqr.valueAt(42) or sqr.valueAt(x) +
sqr.valueAt(y).
When a lambda expression has two parameters, the parentheses are not optional. Here is
an example of using the ArrayProcessor interface, which also demonstrates a lambda expression
with a multiline definition:
ArrayProcessor concat;
concat = (A,n) -> { // parentheses around (A,n) are required!
String str;
str = "";
for (int i = 0; i < n; i++)
str += A[i];
System.out.println(str);
}; // The semicolon marks the end of the assignment statement;
// it is not part of the lambda expression.
String[] nums;
nums = new String[4];
nums[0] = "One";
nums[1] = "Two";
nums[2] = "Three";
nums[3] = "Four";
for (int i = 1; i < nums.length; i++) {
CHAPTER 4. SUBROUTINES 170
concat.process( nums, i );
}
This will print out
One
OneTwo
OneTwoThree
OneTwoThreeFour
Things get more interesting when a lambda expression is used as an actual parameter, which
is the most common use in practice. For example, suppose that the following function is defined:
/**
* For a function f, compute f(start) + f(start+1) + ... + f(end).
* The value of end should be >= the value of start.
*/
static double sum( FunctionR2R f, int start, int end ) {
double total = 0;
for (int n = start; n <= end; n++) {
total = total + f.valueAt( n );
}
return total;
}
Note that since f is a value of type FunctionR2R, the value of f at n is actually written as
f.valueAt(n). When the function sum is called, the first parameter can be given as a lambda
expression. For example:
System.out.print("The sum of n squared for n from 1 to 100 is ");
System.out.println( sum( x -> x*x, 1, 100 ) );
System.out.print("The sum of 2 raised to the power n, for n from 1 to 10 is ");
System.out.println( sum( num -> Math.pow(2,num), 1, 10 ) );
As another example, suppose that we have a subroutine that performs a given task several
times. The task can be specified as a value of type Runnable:
static void doSeveralTimes( Runnable task, int repCount ) {
for (int i = 0; i < repCount; i++) {
task.run(); // Perform the task!
}
}
We could then say “Hello World” ten times by calling
doSeveralTimes( () -> System.out.println("Hello World"), 10 );
Note that for a lambda expression of type Runnable, the parameter list is given as an empty
pair of parentheses. Here is an example in which the syntax is getting rather complicated:
doSeveralTimes( () -> {
// count from 1 up to some random number between 5 and 25
int count = 5 + (int)(21*Math.random());
for (int i = 1; i <= count; i++) {
System.out.print(i + " ");
}
System.out.println();
}, 100);
CHAPTER 4. SUBROUTINES 171
This is a single subroutine call statement in which the first parameter is a lambda expression
that extends over multiple lines. The second parameter is 100, and the semicolon on the last
line ends the subroutine call statement.
We have seen examples of assigning a lambda expression to a variable and of using one as
an actual parameter. Here is an example in which a lambda expression is the return value of a
function:
static FunctionR2R makePowerFunction( int n ) {
return x -> Math.pow(x,n);
}
Then makePowerFunction(2) returns a FunctionR2R that computes the square of its parameter,
while makePowerFunction(10) returns a FunctionR2R that computes the 10-th power of its
parameter. This example also illustrates the fact that a lambda expression can use other
variables in addition to its parameter, such as n in this case (although there are some restrictions
on when that can be done).
read the user’s typed replies. A modern graphical user interface, with windows, buttons, scroll
bars, menus, text-input boxes, and so on, might make things easier for the user, but it forces
the programmer to cope with a hugely expanded array of possibilities. The programmer sees
this increased complexity in the form of great numbers of subroutines that are provided for
managing the user interface, as well as for other purposes.
4.6.1 Toolboxes
Someone who wanted to program for the original Macintosh computers—and to produce pro-
grams that look and behave the way users expected them to—had to deal with the “Macintosh
Toolbox,” a collection of well over a thousand different subroutines. There were routines for
opening and closing windows, for drawing geometric figures and text to windows, for adding
buttons to windows, and for responding to mouse clicks on the window. There were other
routines for creating menus and for reacting to user selections from menus. Aside from the
user interface, there were routines for opening files and reading data from them, for commu-
nicating over a network, for sending output to a printer, for handling communication between
programs, and in general for doing all the standard things that a computer has to do. Microsoft
Windows provides its own set of subroutines for programmers to use, and they are quite a bit
different from the subroutines used on the Mac. Linux has several different GUI toolboxes for
the programmer to choose from.
The analogy of a “toolbox” is a good one to keep in mind. Every programming project
involves a mixture of innovation and reuse of existing tools. A programmer is given a set of
tools to work with, starting with the set of basic tools that are built into the language: things
like variables, assignment statements, if statements, and loops. To these, the programmer can
add existing toolboxes full of routines that have already been written for performing certain
tasks. These tools, if they are well-designed, can be used as true black boxes: They can be called
to perform their assigned tasks without worrying about the particular steps they go through to
accomplish those tasks. The innovative part of programming is to take all these tools and apply
them to some particular project or problem (word-processing, keeping track of bank accounts,
processing image data from a space probe, Web browsing, computer games, . . . ). This is called
applications programming .
A software toolbox is a kind of black box, and it presents a certain interface to the pro-
grammer. This interface is a specification of what routines are in the toolbox, what parameters
they use, and what tasks they perform. This information constitutes the API , or Application
Programming Interface, associated with the toolbox. The Macintosh API is a specification
of all the routines available in the Macintosh Toolbox. A company that makes some hard-
ware device—say a card for connecting a computer to a network—might publish an API for
that device consisting of a list of routines that programmers can call in order to communicate
with and control the device. Scientists who write a set of routines for doing some kind of
complex computation—such as solving “differential equations,” say—would provide an API to
allow others to use those routines without understanding the details of the computations they
perform.
∗ ∗ ∗
The Java programming language is supplemented by a large, standard API. You’ve seen
part of this API already, in the form of mathematical subroutines such as Math.sqrt(), the
String data type and its associated routines, and the System.out.print() routines. The
standard Java API includes routines for working with graphical user interfaces, for network
CHAPTER 4. SUBROUTINES 173
communication, for reading and writing files, and more. It’s tempting to think of these routines
as being part of the Java language, but they are technically subroutines that have been written
and made available for use in Java programs.
Java is platform-independent. That is, the same program can run on platforms as diverse as
Mac OS, Windows, Linux, and others. The same Java API must work on all these platforms.
But notice that it is the interface that is platform-independent; the implementation of some
parts of the API varies from one platform to another. A Java system on a particular computer
includes implementations of all the standard API routines. A Java program includes only calls
to those routines. When the Java interpreter executes a program and encounters a call to one
of the standard routines, it will pull up and execute the implementation of that routine which
is appropriate for the particular platform on which it is running. This is a very powerful idea.
It means that you only need to learn one API to program for a wide variety of platforms.
j#$#
l!" u i
M S F+,-
s
nn
./01-2
r %&
'()*%
S T
I D
Similarly, the package javafx contains a package javafx.scene, which in turn contains
javafx.scene.control. This package contains classes that represent GUI components such as
buttons and input boxes. Another subpackage, javafx.scene.paint, contains class Color and
other classes that define ways to fill and stroke a shape.
The standard Java API includes thousands of classes in hundreds of packages. Many of
the classes are rather obscure or very specialized, but you might want to browse through the
documentation to see what is available. As I write this, the documentation for the complete
basic API for Java 8 can be found at
https://docs.oracle.com/javase/8/docs/api/
and for JavaFX at
https://docs.oracle.com/javase/8/javafx/api/toc.htm
See the subsection about “modules,” below, for a discussion of changes that were made the
language in Java 9 and for links to the documentation for Java 11. However, for the purposes
of this textbook, you will probably find that the Java 8 documentation is easier to use.
Even an expert programmer won’t be familiar with the entire Java API, or even a majority
of it. In this book, you’ll only encounter several dozen classes, and those will be sufficient for
writing a wide variety of programs.
at the beginning of a Java source code file, then, in the rest of the file, you can abbreviate the
full name javafx.scene.paint.Color to just the simple name of the class, which is Color.
Note that the import line comes at the start of a file (after the package statement, if there
is one) and is not inside any class. Although it is sometimes referred to as a statement, it is
more properly called an import directive since it is not a statement in the usual sense. The
import directive “import javafx.scene.paint.Color” would allow you to say
Color rectColor;
to declare the variable. Note that the only effect of the import directive is to allow you to use
simple class names instead of full “package.class” names. You aren’t really importing anything
substantial; if you leave out the import directive, you can still access the class—you just have
to use its full name. There is a shortcut for importing all the classes from a given package. For
example, you can import all the classes from java.util by saying
import java.util.*;
The “*” is a wildcard that matches every class in the package. (However, it does not match
sub-packages; for example, you cannot import the entire contents of all the sub-packages of
the javafx package by saying import javafx.*.)
Some programmers think that using a wildcard in an import statement is bad style, since
it can make a large number of class names available that you are not going to use and might
not even know about. They think it is better to explicitly import each individual class that
you want to use. In my own programming, I often use wildcards to import all the classes from
the most relevant packages, and use individual imports when I am using just one or two classes
from a given package.
A program that works with networking might include the line “import java.net.*;”, while
one that reads or writes files might use “import java.io.*;”. But when you start importing
lots of packages in this way, you have to be careful about one thing: It’s possible for two classes
that are in different packages to have the same name. For example, both the java.awt package
and the java.util package contain a class named List. If you import both java.awt.* and
java.util.*, the simple name List will be ambiguous. If you try to declare a variable of
type List, you will get a compiler error message about an ambiguous class name. You can
still use both classes in your program: Use the full name of the class, either java.awt.List or
java.util.List. Another solution, of course, is to use import to import the individual classes
you need, instead of importing entire packages.
Because the package java.lang is so fundamental, all the classes in java.lang are auto-
matically imported into every program. It’s as if every program began with the statement
“import java.lang.*;”. This is why we have been able to use the class name String instead
of java.lang.String, and Math.sqrt() instead of java.lang.Math.sqrt(). It would still,
however, be perfectly legal to use the longer forms of the names.
Programmers can create new packages. Suppose that you want some classes that you are
writing to be in a package named utilities. Then the source code files that defines those
classes must begin with the line
package utilities;
This would come even before any import directive in that file. Furthermore, the source code
file would be placed in a folder with the same name as the package, “utilities” in this example.
And a class that is in a subpackage must be in a subfolder. For example, a class in a package
named utilities.net would be in folder named “net” inside a folder named “utilities”. A
CHAPTER 4. SUBROUTINES 176
class that is in a package automatically has access to other classes in the same package; that
is, a class doesn’t have to import classes from the package in which it is defined.
In projects that define large numbers of classes, it makes sense to organize those classes
into packages. It also makes sense for programmers to create new packages as toolboxes that
provide functionality and APIs for dealing with areas not covered in the standard Java API.
(And in fact such “toolmaking” programmers often have more prestige than the applications
programmers who use their tools.)
However, with just a couple of exceptions such as class TextIO in package textio, the classes
written for this book are not in packages. For the purposes of this book, you need to know
about packages mainly so that you will be able to import TextIO and classes from the standard
packages. The standard packages are always available to the programs that you write. You
might wonder where the standard classes are actually located. Again, that can depend to some
extent on the version of Java that you are using. In Java 8 they are stored in jar files in a
subdirectory named lib inside the Java Runtime Environment installation directory. A .jar (or
“Java archive”) file is a single file that can contain many classes. Most of the classes used with
Java 8 can be found in a jar file named rt.jar. Things changed substantially in Java 9, as
discussed in the next subsection.
Although we won’t be creating packages explicitly, every class is actually part of a package.
If a class is not specifically placed in a package, then it is put in something called the default
package, which has no name. Almost all the examples that you see in this book are in the
default package.
have not installed a JDK on their computer. But just as for the JDK itself, different versions
of the custom runtime will be needed for Windows, for Mac OS, and for Linux. Furthermore,
when security updates are made to the JDK, they are not automatically applied to custom
runtimes, so the application developer takes on the responsibility of updating custom runtimes.
Once again, this is really only useful for fairly large applications.
In a JDK for Java 9 or later, compiled class files from the standard modules are stored
together in a file named modules inside a directory named lib in the main JDK directory. This
is a so-called “jimage file,” and there is a command-line tool named jimage for working with
such files. If you use the jlink tool to create a custom runtime, part of what it does is to
create a custom modules file containing just the modules that are needed by the runtime. In
the JDK 12 on my Linux computer, modules is a 130 megabyte file containing 30199 classes in
1000 packages in 70 modules. The JDK directory also has a subdirectory named jmods that
contains the modules in another form. However, it is not required for compiling and running
programs and, as far as I can tell, is meant mostly for use by jlink.
Modules in the JDK include, for example, java.base (which contains the basic modules such
as java.lang and java.util) and java.desktop (which include packages for the Swing GUI
toolkit). JavaFX packages include javafx.base, javafx.control, javafx.graphics, and a few that
are less generally useful. The API documentation for modular versions of Java is divided into
modules, then into packages, and finally into classes. This makes the documentation harder
to browse than in older versions of Java. However, the documentation web site does have an
effective search feature. As I write this, the documentation for Java 11 and for JavaFX 11 is
available at:
https://docs.oracle.com/en/java/javase/11/docs/api/index.html
https://openjfx.io/javadoc/11/
A class can be defined outside of any module, and it is possible for that class to use packages
from modules, provided that those packages are exported by the modules where they are defined.
In particular, a programmer can use classes from the JDK without ever thinking about modules
or knowing that they exist. This applies to all the command-line programs in this book.
However, when using Java 11 or later, things are different for GUI programs that use JavaFX,
which has been removed from the JDK and is distributed as a separate set of modules. As we
saw in Section 2.6, when you compile or run a JavaFX program, you need to specify a module
path that includes the JavaFX modules, and you need to provide an --add-modules option.
(In Section 2.6, the value for --add-modules was given as ALL-MODULE-PATH, which lets the
program access any modules that are found on the module path. An alternative is to specify a
list of names of just those modules that are actually used by the program.)
Aside from using modules with JavaFX and the basic background information in this section,
this textbook does not cover modules.
4.6.5 Javadoc
To use an API effectively, you need good documentation for it. The documentation for most
Java APIs is prepared using a system called Javadoc. For example, this system is used to
prepare the documentation for Java’s standard packages. And almost everyone who creates a
toolbox in Java publishes Javadoc documentation for it.
Javadoc documentation is prepared from special comments that are placed in the Java
source code file. Recall that one type of Java comment begins with /* and ends with */. A
Javadoc comment takes the same form, but it begins with /** rather than simply /*. You have
already seen comments of this form in many of the examples in this book.
CHAPTER 4. SUBROUTINES 178
Note that a Javadoc comment must be placed just before the subroutine that it is com-
menting on. This rule is always followed. You can have Javadoc comments for subroutines, for
member variables, and for classes. The Javadoc comment always immediately precedes the
thing it is commenting on.
Like any comment, a Javadoc comment is ignored by the computer when the file is compiled.
But there is a tool called javadoc that reads Java source code files, extracts any Javadoc
comments that it finds, and creates a set of Web pages containing the comments in a nicely
formatted, interlinked form. By default, javadoc will only collect information about public
classes, subroutines, and member variables, but it allows the option of creating documentation
for non-public things as well. If javadoc doesn’t find any Javadoc comment for something, it
will construct one, but the comment will contain only basic information such as the name and
type of a member variable or the name, return type, and parameter list of a subroutine. This
is syntactic information. To add information about semantics and pragmatics, you have to
write a Javadoc comment.
As an example, you can look at the documentation Web page for TextIO. The documentation
page was created by applying the javadoc tool to the source code file, TextIO.java. If you
have downloaded the on-line version of this book, the documentation can be found in the
TextIO Javadoc directory, or you can find a link to it in the on-line version of this section.
In a Javadoc comment, the *’s at the start of each line are optional. The javadoc tool
will remove them. In addition to normal text, the comment can contain certain special codes.
For one thing, the comment can contain HTML mark-up commands. HTML is the language
that is used to create web pages, and Javadoc comments are meant to be shown on web pages.
The javadoc tool will copy any HTML commands in the comments to the web pages that it
creates. The book will not teach you HTML, but as an example, you can add <p> to indicate
the start of a new paragraph. (Generally, in the absence of HTML commands, blank lines and
extra spaces in the comment are ignored. Furthermore, the characters & and < have special
meaning in HTML and should not be used in Javadoc comments except with those meanings;
they can be written as & and <.)
In addition to HTML commands, Javadoc comments can include doc tags, which are
processed as commands by the javadoc tool. A doc tag has a name that begins with the
character @. I will only discuss four tags: @author, @param, @return, and @throws. The
@author tag can be used only for a class, and should be followed by the name of the author.
The other three tags are used in Javadoc comments for a subroutine to provide information
about its parameters, its return value, and the exceptions that it might throw. These tags
must be placed at the end of the comment, after any description of the subroutine itself. The
syntax for using them is:
@param hparameter-name i hdescription-of-parameter i
@return hdescription-of-return-value i
@throws hexception-class-name i hdescription-of-exception i
The hdescriptionsi can extend over several lines. The description ends at the next doc tag or at
the end of the comment. You can include a @param tag for every parameter of the subroutine
and a @throws for as many types of exception as you want to document. You should have
a @return tag only for a non-void subroutine. These tags do not have to be given in any
particular order.
Here is an example that doesn’t do anything exciting but that does use all three types of
doc tag:
CHAPTER 4. SUBROUTINES 179
/**
* This subroutine computes the area of a rectangle, given its width
* and its height. The length and the width should be positive numbers.
* @param width the length of one side of the rectangle
* @param height the length the second side of the rectangle
* @return the area of the rectangle
* @throws IllegalArgumentException if either the width or the height
* is a negative number.
*/
public static double areaOfRectangle( double length, double width ) {
if ( width < 0 || height < 0 )
throw new IllegalArgumentException("Sides must have positive length.");
double area;
area = width * height;
return area;
}
I use Javadoc comments for many of my examples. I encourage you to use them in your
own code, even if you don’t plan to generate Web page documentation of your work, since it’s
a standard format that other Java programmers will be familiar with.
If you do want to create Web-page documentation, you need to run the javadoc tool. This
tool is available as a command in the Java Development Kit that was discussed in Section 2.6.
You can use the javadoc tool in a command line interface similarly to the way that the javac
and java commands are used. Javadoc can also be applied in the integrated development
environments that were also discussed in Section 2.6. I won’t go into any of the details here;
consult the documentation for your programming environment.
or give incorrect results, but you can only blame yourself, not the subroutine, because you
haven’t lived up to your side of the deal.
A postcondition of a subroutine represents the other side of the contract. It represents an
obligation of the subroutine. It is something that will be true after the subroutine has run
(assuming that its preconditions were met—and that there are no bugs in the subroutine).
The postcondition of the function Math.sqrt() is that the square of the value that is returned
by this function is equal to the parameter that is provided when the subroutine is called. Of
course, this will only be true if the precondition—that the parameter is greater than or equal
to zero—is met. A postcondition of the built-in subroutine System.out.print(x) is that the
value of the parameter has been displayed on the screen.
Preconditions most often give restrictions on the acceptable values of parameters, as in the
example of Math.sqrt(x). However, they can also refer to global variables that are used in the
subroutine. Or, if it only makes sense to call the subroutine at certain times, the precondition
might refer to the state that the program must be in when the subroutine is called.
The postcondition of a subroutine, on the other hand, specifies the task that it performs.
For a function, the postcondition should specify the value that the function returns.
Subroutines are sometimes described by comments that explicitly specify their preconditions
and postconditions. When you are given a pre-written subroutine, a statement of its precon-
ditions and postconditions tells you how to use it and what it does. When you are assigned
to write a subroutine, the preconditions and postconditions give you an exact specification of
what the subroutine is expected to do. I will use this approach in the example that constitutes
the rest of this section. The comments are given in the form of Javadoc comments, but I will
explicitly label the preconditions and postconditions. (Many computer scientists think that
new doc tags @precondition and @postcondition should be added to the Javadoc system for
explicit labeling of preconditions and postconditions, but that has not yet been done.)
/**
* Sets the color of one of the rectangles in the window.
*
* Precondition: row and col are in the valid range of row and column numbers,
* and r, g, and b are in the range 0 to 255, inclusive.
* Postcondition: The color of the rectangle in row number row and column
* number col has been set to the color specified by r, g,
* and b. r gives the amount of red in the color with 0
* representing no red and 255 representing the maximum
* possible amount of red. The larger the value of r, the
* more red in the color. g and b work similarly for the
* green and blue color components.
*/
public static void setColor(int row, int col, int r, int g, int b)
/**
* Gets the red component of the color of one of the rectangles.
*
* Precondition: row and col are in the valid range of row and column numbers.
* Postcondition: The red component of the color of the specified rectangle is
* returned as an integer in the range 0 to 255 inclusive.
*/
public static int getRed(int row, int col)
/**
* Like getRed, but returns the green component of the color.
*/
public static int getGreen(int row, int col)
/**
* Like getRed, but returns the blue component of the color.
*/
public static int getBlue(int row, int col)
/**
* Inserts a delay in the program (to regulate the speed at which the colors
* are changed, for example).
*
* Precondition: milliseconds is a positive integer.
* Postcondition: The program has paused for at least the specified number
* of milliseconds, where one second is equal to 1000
CHAPTER 4. SUBROUTINES 183
* milliseconds.
*/
public static void delay(int milliseconds)
Remember that these subroutines are members of the Mosaic class, so when they are called
from outside Mosaic, the name of the class must be included as part of the name of the routine.
For example, we’ll have to use the name Mosaic.isOpen() rather than simply isOpen().
You’ll notice that the comments on the subroutine don’t specify what happens when the
preconditions are not met. Although a subroutine is not really obligated by its contract to
do anything particular in that case, it would be good to know what happens. For example,
if the precondition, “row and col are in the valid range of row and column numbers,” on
the setColor() or getRed() routine is violated, an IllegalArgumentException will be thrown.
Knowing that fact would allow you to write programs that catch and handle the exception,
and it would be good to document it with a @throws doc tag in the Javadoc comment. Other
questions remain about the behavior of the subroutines. For example, what happens if you
call Mosaic.open() and there is already a mosaic window open on the screen? (In fact, the
second call will simply be ignored.) It’s difficult to fully document the behavior of a piece of
software—sometimes, you just have to experiment or look at the full source code.
∗ ∗ ∗
My idea for a program is to use the Mosaic class as the basis for a neat animation. I want
to fill the window with randomly colored squares, and then randomly change the colors in a
loop that continues as long as the window is open. “Randomly change the colors” could mean
a lot of different things, but after thinking for a while, I decide it would be interesting to have
a “disturbance” that wanders randomly around the window, changing the color of each square
that it encounters. Here’s a picture showing what the contents of the window might look like
at one point in time:
With basic routines for manipulating the window as a foundation, I can turn to the specific
problem at hand. A basic outline for my program is
Open a Mosaic window
Fill window with random colors
Move around, changing squares at random
Filling the window with random colors seems like a nice coherent task that I can work on
separately, so let’s decide to write a separate subroutine to do it. The third step can be
expanded a bit more, into the steps: Start in the middle of the window, then keep moving
CHAPTER 4. SUBROUTINES 184
to new squares and changing the color of those squares. This should continue as long as the
mosaic window is still open. Thus we can refine the algorithm to:
Open a Mosaic window
Fill window with random colors
Set the current position to the middle square in the window
As long as the mosaic window is open:
Randomly change color of the square at the current position
Move current position up, down, left, or right, at random
I need to represent the current position in some way. That can be done with two int variables
named currentRow and currentColumn that hold the row number and the column number of
the square where the disturbance is currently located. I’ll use 16 rows and 20 columns of squares
in my mosaic, so setting the current position to be in the center means setting currentRow to 8
and currentColumn to 10. I already have a subroutine, Mosaic.open(), to open the window.
To keep the main routine simple, I decide that I will write two more subroutines of my own to
carry out the two tasks in the while loop. The algorithm can then be written in Java as:
Mosaic.open(16,20,25,25)
fillWithRandomColors();
currentRow = 8; // Middle row, halfway down the window.
currentColumn = 10; // Middle column.
while ( true ) { // Program ends when user closes the window.
changeToRandomColor(currentRow, currentColumn);
randomMove();
}
With the proper wrapper, this is essentially the main() routine of my program. It turns out I
decided to make one small modification after running the completed program: To prevent the
animation from running too fast, I added the line “Mosaic.delay(10);” to the while loop.
The main() routine is taken care of, but to complete the program, I still have to write the
subroutines fillWithRandomColors(), changeToRandomColor(int,int), and randomMove().
Writing each of these subroutines is a separate, small task. The fillWithRandomColors()
routine is defined by the postcondition that “each of the rectangles in the mosaic has been
changed to a random color.” Pseudocode for an algorithm to accomplish this task can be given
as:
For each row:
For each column:
set the square in that row and column to a random color
“For each row” and “for each column” can be implemented as for loops. We’ve already planned
to write a subroutine changeToRandomColor that can be used to set the color. (The possi-
bility of reusing subroutines in several places is one of the big payoffs of using them!) So,
fillWithRandomColors() can be written in proper Java as:
static void fillWithRandomColors() {
int row, column;
for ( row = 0; row < 16; row++ )
for ( column = 0; column < 20; column++ )
changeToRandomColor(row,column);
}
CHAPTER 4. SUBROUTINES 185
A member variable can also be initialized at the point where it is declared, just as for a
local variable. For example:
public class Bank {
private static double interestRate = 0.05;
private static int maxWithdrawal = 200;
.
. // More variables and subroutines.
.
}
A static member variable is created as soon as the class is loaded by the Java interpreter, and
the initialization is also done at that time. In the case of member variables, this is not simply
an abbreviation for a declaration followed by an assignment statement. Declaration statements
are the only type of statement that can occur outside of a subroutine. Assignment statements
cannot, so the following is illegal:
public class Bank {
private static double interestRate;
interestRate = 0.05; // ILLEGAL:
. // Can’t be outside a subroutine!:
.
.
Because of this, declarations of member variables often include initial values. In fact, as
mentioned in Subsection 4.2.4, if no initial value is provided for a member variable, then a
default initial value is used. For example, when declaring an integer member variable, count,
“static int count;” is equivalent to “static int count = 0;”.
Even array variables can be initialized. An array contains several elements, not just a single
value. To initialize an array variable, you can provide a list of values, separated by commas,
and enclosed between a pair of braces. For example:
int[] smallPrimes = { 2, 3, 5, 7, 11, 13, 17, 19, 23, 29 };
In this statement, an array of int of length 10 is created and filled with the values in the list.
The length of the array is determined by the number of items in the list.
Note that this syntax for initializing arrays cannot be used in assignment statements. It
can only be used in a declaration statement at the time when the array variable is declared.
It is also possible to initialize an array variable with an array created using the new operator
(which can also be used in assignment statements). For example:
String[] nameList = new String[100];
but in that case, of course, all the array elements will have their default value.
always a good idea to run a program using several different values for any named constant, to
test that it works properly in all cases.
Here is the complete new program, RandomMosaicWalk2, with all modifications from the
previous version shown in italic. Note in particular how the constants ROWS and COLUMNS are
used in randomMove() when moving the disturbance from one edge of the mosaic to the opposite
edge. I’ve left out most of the comments to save space.
public class RandomMosaicWalk2 {
final static int ROWS = 20; // Number of rows in mosaic.
final static int COLUMNS = 30; // Number of columns in mosaic.
final static int SQUARE SIZE = 15; // Size of each square in mosaic.
static int currentRow; // Row currently containing the disturbance.
static int currentColumn; // Column currently containing the disturbance.
public static void main(String[] args) {
Mosaic.open( ROWS, COLUMNS, SQUARE SIZE, SQUARE SIZE );
fillWithRandomColors();
currentRow = ROWS / 2; // start at center of window
currentColumn = COLUMNS / 2;
while (true) {
changeToRandomColor(currentRow, currentColumn);
randomMove();
Mosaic.delay(5);
}
} // end main
static void fillWithRandomColors() {
for (int row=0; row < ROWS; row++) {
for (int column=0; column < COLUMNS; column++) {
changeToRandomColor(row, column);
}
}
} // end fillWithRandomColors
static void changeToRandomColor(int rowNum, int colNum) {
int red = (int)(256*Math.random()); // Choose random levels in range
int green = (int)(256*Math.random()); // 0 to 255 for red, green,
int blue = (int)(256*Math.random()); // and blue color components.
Mosaic.setColor(rowNum,colNum,red,green,blue);
} // end changeToRandomColor
static void randomMove() {
int directionNum; // Randomly set to 0, 1, 2, or 3 to choose direction.
directionNum = (int)(4*Math.random());
switch (directionNum) {
case 0: // move up
currentRow--;
if (currentRow < 0)
currentRow = ROWS - 1;
break;
case 1: // move right
currentColumn++;
if (currentColumn >= COLUMNS)
currentColumn = 0;
CHAPTER 4. SUBROUTINES 193
break;
case 2: // move down
currentRow++;
if (currentRow >= ROWS)
currentRow = 0;
break;
case 3: // move left
currentColumn--;
if (currentColumn < 0)
currentColumn = COLUMNS - 1;
break;
}
} // end randomMove
} // end class RandomMosaicWalk2
In the statements that make up the body of the playGame() subroutine, the name “count”
refers to the local variable. In the rest of the Game class, “count” refers to the member variable
(unless hidden by other local variables or parameters named count). However, the member vari-
able named count can also be referred to by the full name Game.count. Usually, the full name
is only used outside the class where count is defined. However, there is no rule against using
it inside the class. The full name, Game.count, can be used inside the playGame() subroutine
to refer to the member variable instead of the local variable. So, the full scope rule is that the
scope of a static member variable includes the entire class in which it is defined, but where the
simple name of the member variable is hidden by a local variable or formal parameter name, the
member variable must be referred to by its full name of the form hclassNamei.hvariableNamei.
(Scope rules for non-static members are similar to those for static members, except that, as we
shall see, non-static members cannot be used in static subroutines.)
The scope of a formal parameter of a subroutine is the block that makes up the body of the
subroutine. The scope of a local variable extends from the declaration statement that defines
the variable to the end of the block in which the declaration occurs. As noted above, it is
possible to declare a loop control variable of a for loop in the for statement, as in “for (int
i=0; i < 10; i++)”. The scope of such a declaration is considered as a special case: It is
valid only within the for statement and does not extend to the remainder of the block that
contains the for statement.
It is not legal to redefine the name of a formal parameter or local variable within its scope,
even in a nested block. For example, this is not allowed:
void badSub(int y) {
int x;
while (y > 0) {
int x; // ERROR: x is already defined.
.
.
.
}
}
In many languages, this would be legal; the declaration of x in the while loop would hide
the original declaration. It is not legal in Java; however, once the block in which a variable is
declared ends, its name does become available for reuse in Java. For example:
void goodSub(int y) {
while (y > 10) {
int x;
.
.
.
// The scope of x ends here.
}
while (y > 0) {
int x; // OK: Previous declaration of x has expired.
.
.
.
}
}
CHAPTER 4. SUBROUTINES 195
You might wonder whether local variable names can hide subroutine names. This can’t
happen, for a reason that might be surprising. There is no rule that variables and subroutines
have to have different names. The computer can always tell whether a name refers to a variable
or to a subroutine, because a subroutine name is always followed by a left parenthesis. It’s
perfectly legal to have a variable called count and a subroutine called count in the same class.
(This is one reason why I often write subroutine names with parentheses, as when I talk about
the main() routine. It’s a good idea to think of the parentheses as part of the name.) Even
more is true: It’s legal to reuse class names to name variables and subroutines. The syntax
rules of Java guarantee that the computer can always tell when a name is being used as a class
name. A class name is a type, and so it can be used to declare variables and formal parameters
and to specify the return type of a function. This means that you could legally have a class
called Insanity in which you declare a function
static Insanity Insanity( Insanity Insanity ) { ... }
The first Insanity is the return type of the function. The second is the function name, the
third is the type of the formal parameter, and the fourth is the name of the formal parameter.
However, please remember that not everything that is possible is a good idea!
Exercises 196
1. To “capitalize” a string means to change the first letter of each word in the string to upper (solution)
case (if it is not already upper case). For example, a capitalized version of “Now is the time
to act!” is “Now Is The Time To Act!”. Write a subroutine named printCapitalized
that will print a capitalized version of a string to standard output. The string to be printed
should be a parameter to the subroutine. Test your subroutine with a main() routine that
gets a line of input from the user and applies the subroutine to it.
Note that a letter is the first letter of a word if it is not immediately preceded in the
string by another letter. Recall from Exercise 3.4 that there is a standard boolean-valued
function Character.isLetter(char) that can be used to test whether its parameter is a
letter. There is another standard char-valued function, Character.toUpperCase(char),
that returns a capitalized version of the single character passed to it as a parameter. That
is, if the parameter is a letter, it returns the upper-case version. If the parameter is not a
letter, it just returns a copy of the parameter.
2. The hexadecimal digits are the ordinary, base-10 digits ’0’ through ’9’ plus the letters ’A’ (solution)
through ’F’. In the hexadecimal system, these digits represent the values 0 through 15,
respectively. Write a function named hexValue that uses a switch statement to find the
hexadecimal value of a given character. The character is a parameter to the function, and
its hexadecimal value is the return value of the function. You should count lower case
letters ’a’ through ’f’ as having the same value as the corresponding upper case letters.
If the parameter is not one of the legal hexadecimal digits, return -1 as the value of the
function.
A hexadecimal integer is a sequence of hexadecimal digits, such as 34A7, ff8, 174204, or
FADE. If str is a string containing a hexadecimal integer, then the corresponding base-10
integer can be computed as follows:
value = 0;
for ( i = 0; i < str.length(); i++ )
value = value*16 + hexValue( str.charAt(i) );
Of course, this is not valid if str contains any characters that are not hexadecimal digits.
Write a program that reads a string from the user. If all the characters in the string are
hexadecimal digits, print out the corresponding base-10 value. If not, print out an error
message.
3. Write a function that simulates rolling a pair of dice until the total on the dice comes up (solution)
to be a given number. The number that you are rolling for is a parameter to the function.
The number of times you have to roll the dice is the return value of the function. The
parameter should be one of the possible totals: 2, 3, . . . , 12. The function should throw
an IllegalArgumentException if this is not the case. Use your function in a program that
computes and prints the number of rolls it takes to get snake eyes. (Snake eyes means
that the total showing on the dice is 2.)
4. This exercise builds on Exercise 4.3. Every time you roll the dice repeatedly, trying to (solution)
get a given total, the number of rolls it takes can be different. The question naturally
arises, what’s the average number of rolls to get a given total? Write a function that
performs the experiment of rolling to get a given total 10000 times. The desired total is
Exercises 197
a parameter to the subroutine. The average number of rolls is the return value. Each
individual experiment should be done by calling the function you wrote for Exercise 4.3.
Now, write a main program that will call your function once for each of the possible totals
(2, 3, ..., 12). It should make a table of the results, something like:
Total On Dice Average Number of Rolls
------------- -----------------------
2 35.8382
3 18.0607
. .
. .
5. This exercise asks you to write a few lambda expressions and a function that returns a (solution)
lambda expression as its value. Suppose that a function interface ArrayProcessor is defined
as
public interface ArrayProcessor {
double apply( double[] array );
}
Write a class that defines four public static final variables of type ArrayProcessor
that process an array in the following ways: find the maximum value in the array, find the
minimum value in an array, find the sum of the values in the array, and find the average
of the values in the array. In each case, the value of the variable should be given by a
lambda expression. The class should also define a function
public static ArrayProcessor counter( double value ) { ...
This function should return an ArrayProcessor that counts the number of times that value
occurs in an array. The return value should be given as a lambda expression.
The class should have a main() routine that tests your work. The program that you
write for this exercise will need access to the file ArrayProcessor.java, which defines the
functional interface.
6. The sample program RandomMosaicWalk.java from Section 4.7 shows a “disturbance” (solution)
that wanders around a grid of colored squares. When the disturbance visits a square,
the color of that square is changed. Here’s an idea for a variation on that program. In
the new version, all the squares start out with the default color, black. Every time the
disturbance visits a square, a small amount is added to the green component of the color
of that square. The result will be a visually interesting effect, as the path followed by the
disturbance gradually turns a brighter and brighter green.
Write a subroutine that will add 25 to the green component of one of the squares in the
mosaic. (But don’t let the green component go over 255, since that’s the largest legal value
for a color component.) The row and column numbers of the square should be given as
parameters to the subroutine. Recall that you can discover the current green component
of the square in row r and column c with the function call Mosaic.getGreen(r,c). Use
your subroutine as a substitute for the changeToRandomColor() subroutine in the program
RandomMosaicWalk2.java. (This is the improved version of the program from Section 4.8
that uses named constants for the number of rows, number of columns, and square size.)
Set the number of rows and the number of columns to 80. Set the square size to 5.
By default, the rectangles in the mosaic have a “3D” appearance and a gray border
that makes them look nicer in the random walk program. But for this program, you
Exercises 198
want to turn off that effect. To do so, call Mosaic.setUse3DEffect(false) in the main
program.
Don’t forget that you will need Mosaic.java and MosaicCanvas.java to compile and run
your program, since they define non-standard classes that are required by the program.
7. For this exercise, you will do something even more interesting with the Mosaic class that (solution)
was discussed in Section 4.7. (Again, don’t forget that you will need Mosaic.java and
MosaicCanvas.java to compile and run your program.)
The program that you write for this exercise should start by filling a mosaic with
random colors. Then repeat the following until the user closes the mosaic window: Se-
lect one of the rectangles in the mosaic at random. Then select one of the neighboring
rectangles—above it, below it, to the left of it, or to the right of it. Copy the color of the
originally selected rectangle to the selected neighbor, so that the two rectangles now have
the same color.
As this process is repeated over and over, it becomes more and more likely that neigh-
boring squares will have the same color. The result is to build up larger color patches. On
the other hand, once the last square of a given color disappears, there is no way for that
color to ever reappear. (Extinction is forever!) If you let the program run long enough,
eventually the entire mosaic will be one uniform color.
8. Write a program that administers a basic addition quiz to the user. There should be (solution)
ten questions. Each question is a simple addition problem such as 17 + 42, where the
numbers in the problem are chosen at random (and are not too big). The program should
ask the user all ten questions and get the user’s answers. After asking all the questions, the
user should print each question again, with the user’s answer. If the user got the answer
right, the program should say so; if not, the program should give the correct answer. At
the end, tell the user their score on the quiz, where each correct answer counts for ten
points.
The program should use three subroutines, one to create the quiz, one to administer
the quiz, and one to grade the quiz. It can use three arrays, with three global variables
of type int[ ], to refer to the arrays. The first array holds the first number from every
question, the second holds the second number from every questions, and the third holds
the user’s answers.
Quiz 199
Quiz on Chapter 4
(answers)
1. A “black box” has an interface and an implementation. Explain what is meant by the
terms interface and implementation.
3. Briefly explain how subroutines can be useful in the top-down design of programs.
4. Discuss the concept of parameters. What are parameters for? What is the difference
between formal parameters and actual parameters?
5. Give two different reasons for using named constants (declared with the final modifier).
8. Suppose that SupplyInt is a functional interface that defines the method public int get().
Write a lambda expression of type SupplyInt that gets a random integer in the range 1 to
6 inclusive. Write another lambda expression of type SupplyInt that gets an int by asking
the user to enter an integer and then returning the user’s response.
9. Write a subroutine named “stars” that will output a line of stars to standard output. (A
star is the character “*”.) The number of stars should be given as a parameter to the
subroutine. Use a for loop. For example, the command “stars(20)” would output
********************
10. Write a main() routine that uses the subroutine that you wrote for Question 7 to output
10 lines of stars with 1 star in the first line, 2 stars in the second line, and so on, as shown
below.
*
**
***
****
*****
******
*******
********
*********
**********
Quiz 200
11. Write a function named countChars that has a String and a char as parameters. The
function should count the number of times the character occurs in the string, and it should
return the result as the value of the function.
12. Write a subroutine with three parameters of type int. The subroutine should determine
which of its parameters is smallest. The value of the smallest parameter should be returned
as the value of the subroutine.
13. Write a function that finds the average of the first N elements of an array of type double.
The array and N are parameters to the subroutine.
14. Explain the purpose of the following function, and explain how it works:
static int[] stripZeros( int[] list ) {
int count = 0;
for (int i = 0; i < list.length; i++) {
if ( list[i] != 0 )
count++;
}
int[] newList;
newList = new int[count];
int j = 0;
for (int i = 0; i < list.length; i++) {
if ( list[i] != 0 ) {
newList[j] = list[i];
j++;
}
}
return newList;
}
Chapter 5
Whereas a subroutine represents a single task, an object can encapsulate both data (in
the form of instance variables) and a number of different tasks or “behaviors” related to that
data (in the form of instance methods). Therefore objects provide another, more sophisticated
type of structure that can be used to help manage the complexity of large programs.
The first four sections of this chapter introduce the basic things you need to know to work
with objects and to define simple classes. The remaining sections cover more advanced topics;
you might not understand them fully the first time through. In particular, Section 5.5 covers the
most central ideas of object-oriented programming: inheritance and polymorphism. However,
in this textbook, we will generally use these ideas in a limited form, by creating independent
classes and building on existing classes rather than by designing entire hierarchies of classes
from scratch.
201
CHAPTER 5. OBJECTS AND CLASSES 202
often referred to as methods. Now that we are starting to use objects, I will be using the term
“method” more often than “subroutine.”
class 34567ata
name:
age:
An important point is that the static member variables are part of the representation of
the class in memory. Their full names, UserData.name and UserData.age, use the name of
the class, since they are part of the class. When we use class UserData to represent the user
of the program, there can only be one user, since we only have memory space to store data
about one user. Note that the class, UserData, and the variables it contains exist as long as
the program runs. (That is essentially what it means to be “static.”) Now, consider a similar
class that includes some non-static variables:
CHAPTER 5. OBJECTS AND CLASSES 203
class PlayerData {
static int playerCount;
String name;
int age;
}
I’ve also included a static variable in the PlayerData class. Here, the static variable
playerCount is stored as part of the representation of the class in memory. Its full name
is PlayerData.playerCount, and there is only one of it, which exists as long as the program
runs. However, the other two variables in the class definition are non-static. There is no such
variable as PlayerData.name or PlayerData.age, since non-static variables do not become
part of the class itself. But the PlayerData class can be used to create objects. There can be
many objects created using the class, and each one will have its own variables called name and
age. This is what it means for the non-static parts of the class to be a template for objects:
Every object gets its own copy of the non-static part of the class. We can visualize the situation
in the computer’s memory after several objects have been created like this:
class PlayerData
playerCount: 3
(constructor)
instanceof PlayerData
name: age:
age:
instanceof PlayerData
name:
age:
Note that the static variable playerCount is part of the class, and there is only one copy.
On the other hand, every object contains a name and an age. An object that is created from
a class is called an instance of that class, and as the picture shows, every object “knows”
which class was used to create it. I’ve shown class PlayerData as containing something called a
“constructor;” the constructor is a subroutine that creates objects.
Now there can be many “players,” because we can make new objects to represent new
players on demand. A program might use the PlayerData class to store information about
multiple players in a game. Each player has a name and an age. When a player joins the game,
a new PlayerData object can be created to represent that player. If a player leaves the game,
the PlayerData object that represents that player can be destroyed. A system of objects in
the program is being used to dynamically model what is happening in the game. You can’t
do this with static variables! “Dynamic” is the opposite of “static.”
∗ ∗ ∗
An object that is created using a class is said to be an instance of that class. We will
sometimes say that the object belongs to the class. The variables that the object contains are
CHAPTER 5. OBJECTS AND CLASSES 204
called instance variables. The methods (that is, subroutines) that the object contains are
called instance methods. For example, if the PlayerData class, as defined above, is used to
create an object, then that object is an instance of the PlayerData class, and name and age
are instance variables in the object.
My examples here don’t include any methods, but methods work similarly to variables.
Static methods are part of the class; non-static, or instance, methods become part of objects
created from the class. It’s not literally true that each object contains its own copy of the actual
compiled code for an instance method. But logically an instance method is part of the object,
and I will continue to say that the object “contains” the instance method.
Note that you should distinguish between the source code for the class, and the class itself
(in memory). The source code determines both the class and the objects that are created from
that class. The “static” definitions in the source code specify the things that are part of the
class itself (in the computer’s memory), whereas the non-static definitions in the source code
specify things that will become part of every instance object that is created from the class.
By the way, static member variables and static member subroutines in a class are sometimes
called class variables and class methods, since they belong to the class itself, rather than
to instances of that class.
As you can see, the static and the non-static portions of a class are very different things and
serve very different purposes. Many classes contain only static members, or only non-static,
and we will see only a few examples of classes that contain a mixture of the two.
Student std;
However, declaring a variable does not create an object! This is an important point, which is
related to this Very Important Fact:
You should think of objects as floating around independently in the computer’s memory. In
fact, there is a special portion of memory called the heap where objects live. Instead of holding
an object itself, a variable holds the information necessary to find the object in memory. This
information is called a reference or pointer to the object. In effect, a reference to an object
is the address of the memory location where the object is stored. When you use a variable of
object type, the computer uses the reference in the variable to find the actual object.
In a program, objects are created using an operator called new, which creates an object and
returns a reference to that object. (In fact, the new operator calls a special subroutine called
a “constructor” in the class.) For example, assuming that std is a variable of type Student,
declared as above, the assignment statement
std = new Student();
would create a new object which is an instance of the class Student, and it would store a
reference to that object in the variable std. The value of the variable is a reference, or pointer,
to the object. The object itself is somewhere in the heap. It is not quite true, then, to say that
the object is the “value of the variable std” (though sometimes it is hard to avoid using this
terminology). It is certainly not at all true to say that the object is “stored in the variable
std.” The proper terminology is that “the variable std refers to or points to the object,”
and I will try to stick to that terminology as much as possible. If I ever say something like “std
is an object,” you should read it as meaning “std is a variable that refers to an object.”
So, suppose that the variable std refers to an object that is an instance of class Student.
That object contains instance variables name, test1, test2, and test3. These instance vari-
ables can be referred to as std.name, std.test1, std.test2, and std.test3. This follows the
usual naming convention that when B is part of A, then the full name of B is A.B. For example,
a program might include the lines
System.out.println("Hello, " + std.name + ". Your test grades are:");
System.out.println(std.test1);
System.out.println(std.test2);
System.out.println(std.test3);
This would output the name and test grades from the object to which std refers. Simi-
larly, std can be used to call the getAverage() instance method in the object by saying
std.getAverage(). To print out the student’s average, you could say:
System.out.println( "Your average is " + std.getAverage() );
More generally, you could use std.name any place where a variable of type String is legal.
You can use it in expressions. You can assign a value to it. You can even use it to call subroutines
from the String class. For example, std.name.length() is the number of characters in the
student’s name.
It is possible for a variable like std, whose type is given by a class, to refer to no object at
all. We say in this case that std holds a null pointer or null reference. The null pointer is
written in Java as “null”. You can store a null reference in the variable std by saying
CHAPTER 5. OBJECTS AND CLASSES 206
std = null;
null is an actual value that is stored in the variable, not a pointer to something else. It is not
correct to say that the variable “points to null”; in fact, the variable is null. For example, you
can test whether the value of std is null by testing
if (std == null) . . .
If the value of a variable is null, then it is, of course, illegal to refer to instance variables
or instance methods through that variable—since there is no object, and hence no instance
variables to refer to! For example, if the value of the variable std is null, then it would be
illegal to refer to std.test1. If your program attempts to use a null pointer illegally in this
way, the result is an error called a null pointer exception. When this happens while the
program is running, an exception of type NullPointerException is thrown.
Let’s look at a sequence of statements that work with objects:
Student std, std1, // Declare four variables of
std2, std3; // type Student.
std = new Student(); // Create a new object belonging
// to the class Student, and
// store a reference to that
// object in the variable std.
std1 = new Student(); // Create a second Student object
// and store a reference to
// it in the variable std1.
std2 = std1; // Copy the reference value in std1
// into the variable std2.
std3 = null; // Store a null reference in the
// variable std3.
std.name = "John Smith"; // Set values of some instance variables.
std1.name = "Mary Jones";
// (Other instance variables have default
// initial values of zero.)
After the computer executes these statements, the situation in the computer’s memory looks
like this:
CHAPTER 5. OBJECTS AND CLASSES 207
std:
std1:
std2:
std3: null
name: name:
9<>9?@ 0 test1: 0
9<>9A@ 0 test2: 0
9<>9B@ 0 test3: 0
g<9Cverage() getAverage()
In this picture, when a variable contains a reference to an object, the value of that variable is
shown as an arrow pointing to the object. Note, by the way, that the Strings are objects! The
variable std3, with a value of null, doesn’t point anywhere. The arrows from std1 and std2
both point to the same object. This illustrates a Very Important Point:
When the assignment “std2 = std1;” was executed, no new object was created. Instead, std2
was set to refer to the very same object that std1 refers to. This is to be expected, since the
assignment statement just copies the value that is stored in std1 into std2, and that value
is a pointer, not an object. But this has some consequences that might be surprising. For
example, std1.name and std2.name are two different names for the same variable, namely
the instance variable in the object that both std1 and std2 refer to. After the string "Mary
Jones" is assigned to the variable std1.name, it is also true that the value of std2.name is
"Mary Jones". There is a potential for a lot of confusion here, but you can help protect yourself
from it if you keep telling yourself, “The object is not in the variable. The variable just holds
a pointer to the object.”
You can test objects for equality and inequality using the operators == and !=, but
here again, the semantics are different from what you are used to. When you make a test
“if (std1 == std2)”, you are testing whether the values stored in std1 and std2 are the
same. But the values that you are comparing are references to objects; they are not objects.
So, you are testing whether std1 and std2 refer to the same object, that is, whether they point
to the same location in memory. This is fine, if it’s what you want to do. But sometimes,
what you want to check is whether the instance variables in the objects have the same values.
To do that, you would need to ask whether “std1.test1 == std2.test1 && std1.test2 ==
std2.test2 && std1.test3 == std2.test3 && std1.name.equals(std2.name)”.
CHAPTER 5. OBJECTS AND CLASSES 208
I’ve remarked previously that Strings are objects, and I’ve shown the strings "Mary Jones"
and "John Smith" as objects in the above illustration. (Strings are special objects, treated
by Java in a special way, and I haven’t attempted to show the actual internal structure of the
String objects.) Since strings are objects, a variable of type String can only hold a reference to a
string, not the string itself. This explains why using the == operator to test strings for equality
is not a good idea. Suppose that greeting is a variable of type String, and that it refers to
the string "Hello". Then would the test greeting == "Hello" be true? Well, maybe, maybe
not. The variable greeting and the String literal "Hello" each refer to a string that contains
the characters H-e-l-l-o. But the strings could still be different objects, that just happen to
contain the same characters; in that case, greeting == "Hello" would be false. The function
greeting.equals("Hello") tests whether greeting and "Hello" contain the same characters,
which is almost certainly the question you want to ask. The expression greeting == "Hello"
tests whether greeting and "Hello" contain the same characters stored in the same mem-
ory location. (Of course, a String variable such as greeting can also contain the special value
null, and it would make sense to use the == operator to test whether “greeting == null”.)
∗ ∗ ∗
The fact that variables hold references to objects, not objects themselves, has a couple of
other consequences that you should be aware of. They follow logically, if you just keep in mind
the basic fact that the object is not stored in the variable. The object is somewhere else; the
variable points to it.
Suppose that a variable that refers to an object is declared to be final. This means that
the value stored in the variable can never be changed, once the variable has been initialized.
The value stored in the variable is a reference to the object. So the variable will continue to
refer to the same object as long as the variable exists. However, this does not prevent the data
in the object from changing. The variable is final, not the object. It’s perfectly legal to say
final Student stu = new Student();
stu.name = "John Doe"; // Change data in the object;
// The value stored in stu is not changed!
// It still refers to the same object.
Next, suppose that obj is a variable that refers to an object. Let’s consider what happens
when obj is passed as an actual parameter to a subroutine. The value of obj is assigned to
a formal parameter in the subroutine, and the subroutine is executed. The subroutine has no
power to change the value stored in the variable, obj. It only has a copy of that value. However,
the value is a reference to an object. Since the subroutine has a reference to the object, it can
change the data stored in the object. After the subroutine ends, obj still points to the same
object, but the data stored in the object might have changed. Suppose x is a variable of type
int and stu is a variable of type Student. Compare:
void dontChange(int z) { void change(Student s) {
z = 42; s.name = "Fred";
} }
The lines: The lines:
x = 17; stu.name = "Jane";
dontChange(x); change(stu);
System.out.println(x); System.out.println(stu.name);
output the value 17. output the value "Fred".
CHAPTER 5. OBJECTS AND CLASSES 209
and a setter method might check that the value that is being assigned to the variable is legal:
public void setTitle( String newTitle ) {
if ( newTitle == null ) // Don’t allow null strings as titles!
title = "(Untitled)"; // Use an appropriate default value instead.
else
title = newTitle;
}
Even if you can’t think of any extra chores to do in a getter or setter method, you might change
your mind in the future when you redesign and improve your class. If you’ve used a getter and
setter from the beginning, you can make the modification to your class without affecting any of
the classes that use your class. The private member variable is not part of the public interface
of your class; only the public getter and setter methods are, and you are free to change their
implementations without changing the public interface of your class. If you haven’t used get
and set from the beginning, you’ll have to contact everyone who uses your class and tell them,
“Sorry people, you’ll have to track down every use that you’ve made of this variable and change
your code to use my new get and set methods instead.”
A couple of final notes: Some advanced aspects of Java rely on the naming convention
for getter and setter methods, so it’s a good idea to follow the convention rigorously. And
though I’ve been talking about using getter and setter methods for a variable, you can define
get and set methods even if there is no variable. A getter and/or setter method defines a
property of the class, that might or might not correspond to a variable. For example, if a class
includes a public void instance method with signature setValue(double), then the class has
a “property” named value of type double, and it has this property whether or not the class
has a member variable named value.
would be 42, not 17, since list[1] and newlist[1] are just different names for the same
element in the array. All this is very natural, once you understand that arrays are objects and
array variables hold pointers to arrays.
This fact also comes into play when an array is passed as a parameter to a subroutine. The
value that is copied into the subroutine is a pointer to the array. The array is not copied. Since
the subroutine has a reference to the original array, any changes that it makes to elements of
the array are being made to the original and will persist after the subroutine returns.
∗ ∗ ∗
Arrays are objects. They can also hold objects. The base type of an array can be a class.
We have already seen this when we used arrays of type String[ ], but any class can be used as
the base type. For example, suppose Student is the class defined earlier in this section. Then
we can have arrays of type Student[ ]. For an array of type Student[ ], each element of the array
is a variable of type Student. To store information about 30 students, we could use an array:
Student[] classlist; // Declare a variable of type Student[].
classlist = new Student[30]; // The variable now points to an array.
The array has 30 elements, classlist[0], classlist[1], . . . classlist[29]. When the array
is created, it is filled with the default initial value, which for an object type is null. So, although
we have 30 array elements of type Student, we don’t yet have any actual Student objects! All
we have is 30 nulls. If we want student objects, we have to create them:
Student[] classlist;
classlist = new Student[30];
for ( int i = 0; i < 30; i++ ) {
classlist[i] = new Student();
}
Once we have done this, each classlist[i] points to an object of type Student. If we want
to talk about the name of student number 3, we can use classlist[3].name. The average
for student number i can be computed by calling classlist[i].getAverage(). You can do
anything with classlist[i] that you could do with any other variable of type Student.
5.2.2 Constructors
Objects are created with the operator, new. For example, a program that wants to use a
PairOfDice object could say:
PairOfDice dice; // Declare a variable of type PairOfDice.
dice = new PairOfDice(); // Construct a new object and store a
// reference to it in the variable.
CHAPTER 5. OBJECTS AND CLASSES 213
In this example, “new PairOfDice()” is an expression that allocates memory for the object,
initializes the object’s instance variables, and then returns a reference to the object. This
reference is the value of the expression, and that value is stored by the assignment statement in
the variable, dice, so that after the assignment statement is executed, dice refers to the newly
created object. Part of this expression, “PairOfDice()”, looks like a subroutine call, and that
is no accident. It is, in fact, a call to a special type of subroutine called a constructor . This
might puzzle you, since there is no such subroutine in the class definition. However, every class
has at least one constructor. If the programmer doesn’t write a constructor definition in a class,
then the system will provide a default constructor for that class. This default constructor
does nothing beyond the basics: allocate memory and initialize instance variables. If you want
more than that to happen when an object is created, you can include one or more constructors
in the class definition.
The definition of a constructor looks much like the definition of any other subroutine, with
three exceptions. A constructor does not have any return type (not even void). The name
of the constructor must be the same as the name of the class in which it is defined. And the
only modifiers that can be used on a constructor definition are the access modifiers public,
private, and protected. (In particular, a constructor can’t be declared static.)
However, a constructor does have a subroutine body of the usual form, a block of statements.
There are no restrictions on what statements can be used. And a constructor can have a list
of formal parameters. In fact, the ability to include parameters is one of the main reasons for
using constructors. The parameters can provide data to be used in the construction of the
object. For example, a constructor for the PairOfDice class could provide the values that are
initially showing on the dice. Here is what the class would look like in that case:
public class PairOfDice {
public int die1; // Number showing on the first die.
public int die2; // Number showing on the second die.
public PairOfDice(int val1, int val2) {
// Constructor. Creates a pair of dice that
// are initially showing the values val1 and val2.
die1 = val1; // Assign specified values
die2 = val2; // to the instance variables.
}
public void roll() {
// Roll the dice by setting each of the dice to be
// a random number between 1 and 6.
die1 = (int)(Math.random()*6) + 1;
die2 = (int)(Math.random()*6) + 1;
}
} // end class PairOfDice
The constructor is declared as “public PairOfDice(int val1, int val2) ...”, with no
return type and with the same name as the name of the class. This is how the Java com-
piler recognizes a constructor. The constructor has two parameters, and values for these
parameters must be provided when the constructor is called. For example, the expression
“new PairOfDice(3,4)” would create a PairOfDice object in which the values of the instance
variables die1 and die2 are initially 3 and 4. Of course, in a program, the value returned by
the constructor should be used in some way, as in
CHAPTER 5. OBJECTS AND CLASSES 214
∗ ∗ ∗
Constructors are subroutines, but they are subroutines of a special type. They are certainly
not instance methods, since they don’t belong to objects. Since they are responsible for creating
objects, they exist before any objects have been created. They are more like static member
subroutines, but they are not and cannot be declared to be static. In fact, according to the
Java language specification, they are technically not members of the class at all! In particular,
constructors are not referred to as “methods.”
Unlike other subroutines, a constructor can only be called using the new operator, in an
expression that has the form
new hclass-name i ( hparameter-list i )
where the hparameter-listi is possibly empty. I call this an expression because it computes and
returns a value, namely a reference to the object that is constructed. Most often, you will store
the returned reference in a variable, but it is also legal to use a constructor call in other ways,
for example as a parameter in a subroutine call or as part of a more complex expression. Of
course, if you don’t save the reference in a variable, you won’t have any way of referring to the
object that was just created.
A constructor call is more complicated than an ordinary subroutine or function call. It is
helpful to understand the exact steps that the computer goes through to execute a constructor
call:
CHAPTER 5. OBJECTS AND CLASSES 216
1. First, the computer gets a block of unused memory in the heap, large enough to hold an
object of the specified type.
2. It initializes the instance variables of the object. If the declaration of an instance variable
specifies an initial value, then that value is computed and stored in the instance variable.
Otherwise, the default initial value is used.
3. The actual parameters in the constructor, if any, are evaluated, and the values are assigned
to the formal parameters of the constructor.
4. The statements in the body of the constructor, if any, are executed.
5. A reference to the object is returned as the value of the constructor call.
The end result of this is that you have a reference to a newly constructed object.
∗ ∗ ∗
For another example, let’s rewrite the Student class that was used in Section 1. I’ll add a
constructor, and I’ll also take the opportunity to make the instance variable, name, private.
public class Student {
private String name; // Student’s name.
public double test1, test2, test3; // Grades on three tests.
public Student(String theName) {
// Constructor for Student objects;
// provides a name for the Student.
// The name can’t be null.
if ( theName == null )
throw new IllegalArgumentException("name can’t be null");
name = theName;
}
public String getName() {
// Getter method for reading the value of the private
// instance variable, name.
return name;
}
public double getAverage() {
// Compute average test grade.
return (test1 + test2 + test3) / 3;
}
} // end of class Student
An object of type Student contains information about some particular student. The con-
structor in this class has a parameter of type String, which specifies the name of that student.
Objects of type Student can be created with statements such as:
std = new Student("John Smith");
std1 = new Student("Mary Jones");
In the original version of this class, the value of name had to be assigned by a program after it
created the object of type Student. There was no guarantee that the programmer would always
remember to set the name properly. In the new version of the class, there is no way to create
a Student object except by calling the constructor, and that constructor automatically sets the
name. Furthermore, the constructor makes it impossible to have a student object whose name is
CHAPTER 5. OBJECTS AND CLASSES 217
null. The programmer’s life is made easier, and whole hordes of frustrating bugs are squashed
before they even have a chance to be born.
Another type of guarantee is provided by the private modifier. Since the instance variable,
name, is private, there is no way for any part of the program outside the Student class to get at
the name directly. The program sets the value of name, indirectly, when it calls the constructor.
I’ve provided a getter function, getName(), that can be used from outside the class to find out
the name of the student. But I haven’t provided any setter method or other way to change the
name. Once a student object is created, it keeps the same name as long as it exists.
Note that it would be legal, and good style, to declare the variable name to be “final”
in this class. An instance variable can be final provided it is either assigned a value in its
declaration or is assigned a value in every constructor in the class. It is illegal to assign a value
to a final instance variable, except inside a constructor.
∗ ∗ ∗
Let’s take this example a little farther to illustrate one more aspect of classes: What happens
when you mix static and non-static in the same class? In that case, it’s legal for an instance
method in the class to use static member variables or call static member subroutines. An object
knows what class it belongs to, and it can refer to static members of that class. But there is
only one copy of the static member, in the class itself. Effectively, all the objects share one
copy of the static member.
As an example, consider a version of the Student class to which I’ve added an ID for each
student and a static member called nextUniqueID. Although there is an ID variable in each
student object, there is only one nextUniqueID variable.
public class Student {
private String name; // Student’s name.
public double test1, test2, test3; // Grades on three tests.
private int ID; // Unique ID number for this student.
private static int nextUniqueID = 0;
// keep track of next available unique ID number
Student(String theName) {
// Constructor for Student objects; provides a name for the Student,
// and assigns the student a unique ID number.
name = theName;
nextUniqueID++;
ID = nextUniqueID;
}
public String getName() {
// Getter method for reading the value of the private
// instance variable, name.
return name;
}
public int getID() {
// Getter method for reading the value of ID.
return ID;
}
public double getAverage() {
// Compute average test grade.
CHAPTER 5. OBJECTS AND CLASSES 218
program bugs are caused by such errors. A programmer might accidently delete an object even
though there are still references to that object. This is called a dangling pointer error , and
it leads to problems when the program tries to access an object that is no longer there. Another
type of error is a memory leak , where a programmer neglects to delete objects that are no
longer in use. This can lead to filling memory with objects that are completely inaccessible,
and the program might run out of memory even though, in fact, large amounts of memory are
being wasted.
Because Java uses garbage collection, such errors are simply impossible. Garbage collection
is an old idea and has been used in some programming languages since the 1960s. You might
wonder why all languages don’t use garbage collection. In the past, it was considered too slow
and wasteful. However, research into garbage collection techniques combined with the incredible
speed of modern computers have combined to make garbage collection feasible. Programmers
should rejoice.
color is. When you draw with a partially transparent color, the background shows through the
color to some extent. A larger value of the parameter t gives a color that is less transparent
and more opaque.
A Color object has only a few instance methods that you are likely to use. Mainly, there
are functions like getRed() to get the individual color components of the color. There are no
“setter” methods to change the color components. In fact, a Color is an immutable object,
meaning that all of its instance variables are final and cannot be changed after the object is
created. Strings are another example of immutable objects, and we will make some of our own
later in this section.
The main point of all this, again, is that many problems have already been solved, and the
solutions are available in Java’s standard classes. If you are faced with a task that looks like
it should be fairly common, it might be worth looking through a Java reference to see whether
someone has already written a class that you can use.
If dice refers to a PairOfDice object, then dice.toString() will return strings such as
“3 and 6”, “5 and 1”, and “double 2”, depending on the numbers showing on the dice. This
method would be used automatically to convert dice to type String in a statement such as
System.out.println( "The dice came up " + dice );
so this statement might output, “The dice came up 5 and 1” or “The dice came up double 2”.
You’ll see another example of a toString() method in the next section.
A disk in this program can be represented as an object. A disk has properties—color, location,
and size—that can be instance variables in the object. As for instance methods, we need to think
about what we might want to do with a disk. An obvious candidate is that we need to be able to
draw it, so we can include an instance method draw(g), where g is a graphics context that will
be used to do the drawing. The class can also include one or more constructors. A constructor
initializes the object. It’s not always clear what data should be provided as parameters to the
constructor. In this case, as an example, the constructor’s parameters specify the location and
size for the circle, but the constructor makes up a color using random values for the red, green,
and blue components. Here’s the complete class:
import javafx.scene.paint.Color;
import javafx.scene.canvas.GraphicsContext;
/**
* A simple class that holds the size, color, and location of a colored disk,
* with a method for drawing the circle in a graphics context. The circle
* is drawn as a filled oval, with a black outline.
*/
public class CircleInfo {
public int radius; // The radius of the circle.
public int x,y; // The location of the center of the circle.
CHAPTER 5. OBJECTS AND CLASSES 223
(int)(height*Math.random()),
(int)(100*Math.random()) );
}
In each frame, the radius of the disk is increased and the disk is drawn using the code
circleData[i].radius++;
circleData[i].draw(g);
These statements look complicated, so let’s unpack them. Now, circleData[i] is an ele-
ment of the array circleData. That means that it is a variable of type CircleInfo. This
variable refers to an object of type CircleInfo, which contains a public instance variable named
radius. This means that circleData[i].radius is the full name for that variable. Since it
is a variable of type int, we can use the ++ operator to increment its value. So the effect of
circleData[i].radius++ is to increase the radius of the circle by one. The second line of code
is similar, but in that statement, circleData[i].draw is an instance method in the CircleInfo
object. The statement circleData[i].draw(g) calls that instance method with a parameter g
that represents the GraphicsContext that is being used for drawing.
The source code example GrowingCircleAnimation.java contains the full source code for the
program, if you are interested. Since the program uses class CircleInfo, you will also need a copy
of CircleInfo.java in order to compile and run the program.
On the other hand, the Student class from the previous section is not very reusable. It
seems to be crafted to represent students in a particular course where the grade will be based
on three tests. If there are more tests or quizzes or papers, it’s useless. If there are two people
in the class who have the same name, we are in trouble (one reason why numerical student ID’s
are often used). Admittedly, it’s much more difficult to develop a general-purpose student class
than a general-purpose pair-of-dice class. But this particular Student class is good only as an
example in a programming textbook.
∗ ∗ ∗
A large programming project goes through a number of stages, starting with specification
of the problem to be solved, followed by analysis of the problem and design of a program
to solve it. Then comes coding , in which the program’s design is expressed in some actual
programming language. This is followed by testing and debugging of the program. After that
comes a long period of maintenance, which means fixing any new problems that are found
in the program and modifying it to adapt it to changing requirements. Together, these stages
form what is called the software life cycle. (In the real world, the ideal of consecutive stages
is seldom if ever achieved. During the analysis stage, it might turn out that the specifications
are incomplete or inconsistent. A problem found during testing requires at least a brief return
to the coding stage. If the problem is serious enough, it might even require a new design.
Maintenance usually involves redoing some of the work from previous stages. . . .)
Large, complex programming projects are only likely to succeed if a careful, systematic
approach is adopted during all stages of the software life cycle. The systematic approach to
programming, using accepted principles of good design, is called software engineering . The
software engineer tries to efficiently construct programs that verifiably meet their specifications
and that are easy to modify if necessary. There is a wide range of “methodologies” that can
be applied to help in the systematic design of programs. (Most of these methodologies seem to
involve drawing little boxes to represent program components, with labeled arrows to represent
relationships among the boxes.)
We have been discussing object orientation in programming languages, which is relevant to
the coding stage of program development. But there are also object-oriented methodologies for
analysis and design. The question in this stage of the software life cycle is, How can one discover
or invent the overall structure of a program? As an example of a rather simple object-oriented
approach to analysis and design, consider this advice: Write down a description of the problem.
Underline all the nouns in that description. The nouns should be considered as candidates for
becoming classes or objects in the program design. Similarly, underline all the verbs. These
are candidates for methods. This is your starting point. Further analysis might uncover the
need for more classes and methods, and it might reveal that subclassing can be used to take
advantage of similarities among classes.
This is perhaps a bit simple-minded, but the idea is clear and the general approach can be
effective: Analyze the problem to discover the concepts that are involved, and create classes to
represent those concepts. The design should arise from the problem itself, and you should end
up with a program whose structure reflects the structure of the problem in a natural way.
Consider card games that are played with a standard deck of playing cards (a so-called “poker”
deck, since it is used in the game of poker).
values of aces and face cards. (When you read the Card class, you’ll see that I’ve also added
support for Jokers.)
A Card object can be constructed knowing the value and the suit of the card. For example,
we can call the constructor with statements such as:
card1 = new Card( Card.ACE, Card.SPADES ); // Construct ace of spades.
card2 = new Card( 10, Card.DIAMONDS ); // Construct 10 of diamonds.
card3 = new Card( v, s ); // This is OK, as long as v and s
// are integer expressions.
A Card object needs instance variables to represent its value and suit. I’ve made these
private so that they cannot be changed from outside the class, and I’ve provided getter methods
getSuit() and getValue() so that it will be possible to discover the suit and value from outside
the class. The instance variables are initialized in the constructor, and are never changed after
that. In fact, I’ve declared the instance variables suit and value to be final, since they are
never changed after they are initialized. An instance variable can be declared final provided
it is either given an initial value in its declaration or is initialized in every constructor in the
class. Since all its instance variables are final, a Card is an immutable object.
Finally, I’ve added a few convenience methods to the class to make it easier to print out
cards in a human-readable form. For example, I want to be able to print out the suit of a
card as the word “Diamonds”, rather than as the meaningless code number 2, which is used
in the class to represent diamonds. Since this is something that I’ll probably have to do in
many programs, it makes sense to include support for it in the class. So, I’ve provided instance
methods getSuitAsString() and getValueAsString() to return string representations of the
suit and value of a card. Finally, I’ve defined the instance method toString() to return a
string with both the value and suit, such as “Queen of Hearts”. Recall that this method will
be used automatically whenever a Card needs to be converted into a String, such as when the
card is concatenated onto a string with the + operator. Thus, the statement
System.out.println( "Your card is the " + card );
is equivalent to
System.out.println( "Your card is the " + card.toString() );
If the card is the queen of hearts, either of these will print out “Your card is the Queen of
Hearts”.
Here is the complete Card class, which can also be found in Card.java. This class is general
enough to be highly reusable, so the work that went into designing, writing, and testing it pays
off handsomely in the long run.
/**
* An object of type Card represents a playing card from a
* standard Poker deck, including Jokers. The card has a suit, which
* can be spades, hearts, diamonds, clubs, or joker. A spade, heart,
* diamond, or club has one of the 13 values: ace, 2, 3, 4, 5, 6, 7,
* 8, 9, 10, jack, queen, or king. Note that "ace" is considered to be
* the smallest value. A joker can also have an associated value;
* this value can be anything and can be used to keep track of several
* different jokers.
*/
public class Card {
public final static int SPADES = 0; // Codes for the 4 suits, plus Joker.
CHAPTER 5. OBJECTS AND CLASSES 230
/**
* Lets the user play one game of HighLow, and returns the
* user’s score in that game. The score is the number of
* correct guesses that the user makes.
*/
private static int play() {
Deck deck = new Deck(); // Get a new deck of cards, and
// store a reference to it in
// the variable, deck.
Card currentCard; // The current card, which the user sees.
Card nextCard; // The next card in the deck. The user tries
// to predict whether this is higher or lower
// than the current card.
int correctGuesses ; // The number of correct predictions the
// user has made. At the end of the game,
// this will be the user’s score.
char guess; // The user’s guess. ’H’ if the user predicts that
CHAPTER 5. OBJECTS AND CLASSES 234
set to the next card in the deck after the user makes
his prediction. */
currentCard = nextCard;
System.out.println();
System.out.println("The card is " + currentCard);
} // end of while loop
System.out.println();
System.out.println("The game is over.");
System.out.println("You made " + correctGuesses
+ " correct predictions.");
System.out.println();
return correctGuesses;
} // end play()
to compute the “value” of a Blackjack hand according to the rules of the game. The rules are
as follows: The value of a hand is obtained by adding up the values of the cards in the hand.
The value of a numeric card such as a three or a ten is its numerical value. The value of a Jack,
Queen, or King is 10. The value of an Ace can be either 1 or 11. An Ace should be counted
as 11 unless doing so would put the total value of the hand over 21. Note that this means that
the second, third, or fourth Ace in the hand will always be counted as 1.
One way to handle this is to extend the existing Hand class by adding a method that
computes the Blackjack value of the hand. Here’s the definition of such a class:
public class BlackjackHand extends Hand {
/**
* Computes and returns the value of this hand in the game
* of Blackjack.
*/
public int getBlackjackValue() {
int val; // The value computed for the hand.
boolean ace; // This will be set to true if the
// hand contains an ace.
int cards; // Number of cards in the hand.
val = 0;
ace = false;
cards = getCardCount(); // (method defined in class Hand.)
for ( int i = 0; i < cards; i++ ) {
// Add the value of the i-th card in the hand.
Card card; // The i-th card;
int cardVal; // The blackjack value of the i-th card.
card = getCard(i);
cardVal = card.getValue(); // The normal value, 1 to 13.
if (cardVal > 10) {
cardVal = 10; // For a Jack, Queen, or King.
}
if (cardVal == 1) {
ace = true; // There is at least one ace.
}
val = val + cardVal;
}
// Now, val is the value of the hand, counting any ace as 1.
// If there is an ace, and if changing its value from 1 to
// 11 would leave the score less than or equal to 21,
// then do so by adding the extra 10 points to val.
if ( ace == true && val + 10 <= 21 )
val = val + 10;
return val;
} // end getBlackjackValue()
} // end class BlackjackHand
CHAPTER 5. OBJECTS AND CLASSES 237
not inherited structure). The relationship between subclass and superclass is sometimes shown
by a diagram in which the subclass is shown below, and connected to, its superclass, as shown
on the left below:
class A OPQRR U
(superclass)
cEGHH J
KHLNcEGHHa
OPQRR Y
In Java, to create a class named “B” as a subclass of a class named “A”, you would write
class B extends A {
.
. // additions to, and modifications of,
. // stuff inherited from class A
.
}
Several classes can be declared as subclasses of the same superclass. The subclasses, which
might be referred to as “sibling classes,” share some structures and behaviors—namely, the ones
they inherit from their common superclass. The superclass expresses these shared structures
and behaviors. In the diagram shown on the right above, classes B, C, and D are sibling classes.
Inheritance can also extend over several “generations” of classes. This is shown in the diagram,
where class E is a subclass of class D which is itself a subclass of class A. In this case, class E
is considered to be a subclass of class A, even though it is not a direct subclass. This whole set
of classes forms a small class hierarchy .
Zehicle
The Vehicle class would include instance variables such as registrationNumber and owner and
instance methods such as transferOwnership(). These are variables and methods common
to all vehicles. The three subclasses of Vehicle—Car, Truck, and Motorcycle—could then be
used to hold variables and methods specific to particular types of vehicles. The Car class
might add an instance variable numberOfDoors, the Truck class might have numberOfAxles,
CHAPTER 5. OBJECTS AND CLASSES 239
and the Motorcycle class could have a boolean variable hasSidecar. (Well, it could in theory
at least, even if it might give a chuckle to the people at the Department of Motor Vehicles.)
The declarations of these classes in a Java program would look, in outline, like this (although
they are likely to be defined in separate files and declared as public classes):
class Vehicle {
int registrationNumber;
Person owner; // (Assuming that a Person class has been defined!)
void transferOwnership(Person newOwner) {
. . .
}
. . .
}
class Car extends Vehicle {
int numberOfDoors;
. . .
}
class Truck extends Vehicle {
int numberOfAxles;
. . .
}
class Motorcycle extends Vehicle {
boolean hasSidecar;
. . .
}
Suppose that myCar is a variable of type Car that has been declared and initialized with the
statement
Car myCar = new Car();
Given this declaration, a program could refer to myCar.numberOfDoors, since numberOfDoors
is an instance variable in the class Car. But since class Car extends class Vehicle, a car also
has all the structure and behavior of a vehicle. This means that myCar.registrationNumber,
myCar.owner, and myCar.transferOwnership() also exist.
Now, in the real world, cars, trucks, and motorcycles are in fact vehicles. The same is true
in a program. That is, an object of type Car or Truck or Motorcycle is automatically an object
of type Vehicle too. This brings us to the following Important Fact:
The practical effect of this in our example is that an object of type Car can be assigned to a
variable of type Vehicle. That is, it would be legal to say
Vehicle myVehicle = myCar;
or even
Vehicle myVehicle = new Car();
CHAPTER 5. OBJECTS AND CLASSES 240
After either of these statements, the variable myVehicle holds a reference to a Vehicle object
that happens to be an instance of the subclass, Car. The object “remembers” that it is in fact
a Car, and not just a Vehicle. Information about the actual class of an object is stored as part
of that object. It is even possible to test whether a given object belongs to a given class, using
the instanceof operator. The test:
if (myVehicle instanceof Car) ...
determines whether the object referred to by myVehicle is in fact a car.
On the other hand, the assignment statement
myCar = myVehicle; // ERROR!
would be illegal because myVehicle could potentially refer to other types of vehicles that are
not cars. This is similar to a problem we saw previously in Subsection 2.5.6: The computer
will not allow you to assign an int value to a variable of type short, because not every int is a
short. Similarly, it will not allow you to assign a value of type Vehicle to a variable of type Car
because not every vehicle is a car. As in the case of ints and shorts, the solution here is to use
type-casting. If, for some reason, you happen to know that myVehicle does in fact refer to a
Car, you can use the type cast (Car)myVehicle to tell the computer to treat myVehicle as if
it were actually of type Car. So, you could say
myCar = (Car)myVehicle;
and you could even refer to ((Car)myVehicle).numberOfDoors. (The parentheses are
necessary because of precedence. The “.” has higher precedence than the type-cast, so
(Car)myVehicle.numberOfDoors would be read as (Car)(myVehicle.numberOfDoors), an at-
tempt to type-cast the int myVehicle.numberOfDoors into a Vehicle, which is impossible.)
As an example of how this could be used in a program, suppose that you want to print out
relevant data about the Vehicle referred to by myVehicle. If it’s a Car, you will want to print
out the car’s numberOfDoors, but you can’t say myVehicle.numberOfDoors, since there is no
numberOfDoors in the Vehicle class. But you could say:
System.out.println("Vehicle Data:");
System.out.println("Registration number: "
+ myVehicle.registrationNumber);
if (myVehicle instanceof Car) {
System.out.println("Type of vehicle: Car");
Car c;
c = (Car)myVehicle; // Type-cast to get access to numberOfDoors!
System.out.println("Number of doors: " + c.numberOfDoors);
}
else if (myVehicle instanceof Truck) {
System.out.println("Type of vehicle: Truck");
Truck t;
t = (Truck)myVehicle; // Type-cast to get access to numberOfAxles!
System.out.println("Number of axles: " + t.numberOfAxles);
}
else if (myVehicle instanceof Motorcycle) {
System.out.println("Type of vehicle: Motorcycle");
Motorcycle m;
m = (Motorcycle)myVehicle; // Type-cast to get access to hasSidecar!
System.out.println("Has a sidecar: " + m.hasSidecar);
}
CHAPTER 5. OBJECTS AND CLASSES 241
Note that for object types, when the computer executes a program, it checks whether
type-casts are valid. So, for example, if myVehicle refers to an object of type Truck, then
the type cast (Car)myVehicle would be an error. When this happens, an exception of type
ClassCastException is thrown. This check is done at run time, not compile time, because the
actual type of the object referred to by myVehicle is not known when the program is compiled.
The code above avoids ClassCastExceptions by using instanceof to test the type of the variable
before doing a type cast.
5.5.4 Polymorphism
As another example, consider a program that deals with shapes drawn on the screen. Let’s say
that the shapes include rectangles, ovals, and roundrects of various colors. (A “roundrect” is
just a rectangle with rounded corners.)
Three classes, Rectangle, Oval, and RoundRect, could be used to represent the three types of
shapes. These three classes would have a common superclass, Shape, to represent features that
all three shapes have in common. The Shape class could include instance variables to represent
the color, position, and size of a shape, and it could include instance methods for changing the
values of those properties. Changing the color, for example, might involve changing the value
of an instance variable, and then redrawing the shape in its new color:
class Shape {
Color color; // (must be imported from package javafx.scene.paint)
void setColor(Color newColor) {
// Method to change the color of the shape.
color = newColor; // change value of instance variable
redraw(); // redraw shape, which will appear in new color
}
void redraw() {
// method for drawing the shape
? ? ? // what commands should go here?
}
. . . // more instance variables and methods
} // end of class Shape
Now, you might see a problem here with the method redraw(). The problem is that each
different type of shape is drawn differently. The method setColor() can be called for any type
of shape. How does the computer know which shape to draw when it executes the redraw()?
Informally, we can answer the question like this: The computer executes redraw() by asking
the shape to redraw itself. Every shape object knows what it has to do to redraw itself.
In practice, this means that each of the specific shape classes has its own redraw() method:
CHAPTER 5. OBJECTS AND CLASSES 242
BeveledRects
To implement beveled rectangles, I can write a new subclass, BeveledRect, of class Shape
and give it its own redraw() method. Automatically, code that I wrote previously—such as
the statement someShape.redraw()—can now suddenly start drawing beveled rectangles, even
though the beveled rectangle class didn’t exist when I wrote the statement!
∗ ∗ ∗
In the statement “someShape.redraw();”, the redraw message is sent to the object
someShape. Look back at the method in the Shape class for changing the color of a shape:
void setColor(Color newColor) {
color = newColor; // change value of instance variable
redraw(); // redraw shape, which will appear in new color
}
A redraw message is sent here, but which object is it sent to? Well, the setColor method is
itself a message that was sent to some object. The answer is that the redraw message is sent to
that same object, the one that received the setColor message. If that object is a rectangle,
then it contains a redraw() method for drawing rectangles, and that is the one that is executed.
If the object is an oval, then the redraw() method from the Oval class is executed. This is what
you should expect, but it means that the “redraw();” statement in the setColor() method
does not necessarily call the redraw() method in the Shape class! The redraw() method that
is executed could be in any subclass of Shape. This is just another case of polymorphism.
The answer may be surprising: We should leave it blank! The fact is that the class Shape
represents the abstract idea of a shape, and there is no way to draw such a thing. Only
particular, concrete shapes like rectangles and ovals can be drawn. So, why should there
even be a redraw() method in the Shape class? Well, it has to be there, or it would be
illegal to call it in the setColor() method of the Shape class, and it would be illegal to write
“someShape.redraw();”. The compiler would complain that someShape is a variable of type
Shape and there’s no redraw() method in the Shape class.
Nevertheless the version of redraw() in the Shape class itself will never actually be called.
In fact, if you think about it, there can never be any reason to construct an actual object of
type Shape! You can have variables of type Shape, but the objects they refer to will always
belong to one of the subclasses of Shape. We say that Shape is an abstract class. An abstract
class is one that is not used to construct objects, but only as a basis for making subclasses. An
abstract class exists only to express the common properties of all its subclasses. A class that
is not abstract is said to be concrete. You can create objects belonging to a concrete class,
but not to an abstract class. A variable whose type is given by an abstract class can only refer
to objects that belong to concrete subclasses of the abstract class.
Similarly, we say that the redraw() method in class Shape is an abstract method , since
it is never meant to be called. In fact, there is nothing for it to do—any actual redrawing is
done by redraw() methods in the subclasses of Shape. The redraw() method in Shape has
to be there. But it is there only to tell the computer that all Shapes understand the redraw
message. As an abstract method, it exists merely to specify the common interface of all the
actual, concrete versions of redraw() in the subclasses. There is no reason for the abstract
redraw() in class Shape to contain any code at all.
Shape and its redraw() method are semantically abstract. You can also tell the computer,
syntactically, that they are abstract by adding the modifier “abstract” to their definitions.
For an abstract method, the block of code that gives the implementation of an ordinary method
is replaced by a semicolon. An implementation must then be provided for the abstract method
in any concrete subclass of the abstract class. Here’s what the Shape class would look like as
an abstract class:
public abstract class Shape {
Color color; // color of shape.
void setColor(Color newColor) {
// method to change the color of the shape
color = newColor; // change value of instance variable
redraw(); // redraw shape, which will appear in new color
}
abstract void redraw();
// abstract method---must be defined in
// concrete subclasses
. . . // more instance variables and methods
} // end of class Shape
Once you have declared the class to be abstract, it becomes illegal to try to create actual
objects of type Shape, and the computer will report a syntax error if you try to do so.
Note, by the way, that the Vehicle class discussed above would probably also be an abstract
class. There is no way to own a vehicle as such—the actual vehicle has to be a car or a truck
CHAPTER 5. OBJECTS AND CLASSES 245
If you run the ShapeDraw program, you can click one of the buttons along the bottom to
add a shape to the picture. The new shape will appear in the upper left corner of the drawing
area. The color of the shape is given by the “pop-up menu” of colors below the drawing area.
Once a shape is on the screen, you can drag it around with the mouse. A shape will maintain
CHAPTER 5. OBJECTS AND CLASSES 246
the same front-to-back order with respect to other shapes on the screen, even while you are
dragging it. However, you can move a shape out in front of all the other shapes if you hold
down the shift key as you click on it.
In the program, the only time when the actual class of a shape is used is when that shape is
added to the screen. Once the shape has been created, it is manipulated entirely as an abstract
shape. The routine that implements dragging, for example, works with variables of type Shape
and makes no reference to any of its subclasses. As the shape is being dragged, the dragging
routine just calls the shape’s draw method each time the shape has to be drawn, so it doesn’t
have to know how to draw the shape or even what type of shape it is. The object is responsible
for drawing itself. If I wanted to add a new type of shape to the program, I would define a
new subclass of Shape, add another button, and program the button to add the correct type of
shape to the screen. No other changes in the programming would be necessary.
Suppose that a reference to “test1” occurs in the definition of some instance method. The
method is part of some particular object of type Student. When that method gets executed,
the occurrence of the name “test1” refers to the test1 variable in that same object. (This
is why simple names of instance members cannot be used in static methods—when a static
method is executed, it is not part of an object, and hence there are no instance members in
sight!)
This leaves open the question of full names for instance members inside the same class
where they are defined. We need a way to refer to “the object that contains this method.” Java
defines a special variable named this for just this purpose. The variable this can be used in
the source code of an instance method to refer to the object that contains the method. This
intent of the name, “this,” is to refer to “this object,” the one right here that this very method
is in. If var is an instance variable in the same object as the method, then “this.var” is a
full name for that variable. If otherMethod() is an instance method in the same object, then
this.otherMethod() could be used to call that method. Whenever the computer executes an
instance method, it automatically sets the variable this to refer to the object that contains the
method.
(Some object oriented languages use the name “self” instead of “this.” Here, an object is seen
as an entity that receives messages and responds by performing some action. From the point
of view of that entity, an instance variable such as self.name refers to the entity’s own name,
something that is part of the entity itself. Calling an instance method such as self.redraw()
is like saying “message to self: redraw!”)
One common use of this is in constructors. For example:
public class Student {
private String name; // Name of the student.
public Student(String name) {
// Constructor. Create a student with specified name.
this.name = name;
}
.
. // More variables and methods.
.
}
In the constructor, the instance variable called name is hidden by a formal parameter that is also
called “name.” However, the instance variable can still be referred to by its full name, which
is this.name. In the assignment statement “this.name = name”, the “name” on the right is
the formal parameter, and the value of that formal parameter is being assigned to the instance
variable, this.name. This is considered to be acceptable style: There is no need to dream up
cute new names for formal parameters that are just used to initialize instance variables. You
can use the same name for the parameter as for the instance variable.
There are other uses for this. Sometimes, when you are writing an instance method, you
need to pass the object that contains the method to a subroutine, as an actual parameter. In
that case, you can use this as the actual parameter. For example, if you wanted to print out
a string representation of the object, you could say “System.out.println(this);”. Or you
could assign the value of this to another variable in an assignment statement. You can store it
in an array. In fact, you can do anything with this that you could do with any other variable,
except change its value. (Consider it to be a final variable.)
CHAPTER 5. OBJECTS AND CLASSES 248
Note that this allows you to extend the behavior of the roll() method even if you don’t know
how the method is implemented in the superclass!
This constructor provides several options and does a lot of initialization. I wanted to provide
easier-to-use constructors with fewer options, but all the initialization still has to be done. The
class also contains these constructors:
public MosaicCanvas() {
this(42,42);
}
public MosaicCanvas(int rows, int columns) {
this(rows,columns,16,16);
}
Each of these constructors exists just to call another constructor, while providing constant
values for some of the parameters. For example, this(42,42) calls the second constructor
listed here, while that constructor in turn calls the main, four-parameter constructor. That
main constructor is eventually called in all cases, so that all the essential initialization gets
done in every case.
5.7 Interfaces
Some object-oriented programming languages, such as C++, allow a class to extend two or
more superclasses. This is called multiple inheritance. In the illustration below, for example,
class E is shown as having both class A and class B as direct superclasses, while class F has
three direct superclasses.
class A efhkk o
class C
class D efhkk p
efhkk m
Such multiple inheritance is not allowed in Java. The designers of Java wanted to keep the
language reasonably simple, and felt that the benefits of multiple inheritance were not worth the
cost in increased complexity. However, Java does have a feature that can be used to accomplish
many of the same goals as multiple inheritance: interfaces. We have already encountered
“functional interfaces” in Section 4.5 in connection with lambda expressions. A functional
interface specifies a single method. However, interfaces can be much more complicated than
that, and they have many other uses.
You are not likely to need to write your own interfaces until you get to the point of writing
fairly complex programs. However, there are several interfaces that are used in important ways
in Java’s standard packages, and you will need to learn how to use them.
CHAPTER 5. OBJECTS AND CLASSES 251
The main difference is that a class that extends AbstractStrokeable cannot extend any other
class, while a class that implements Strokeable can also extend some class, as well as implement
other interfaces. Of course, an abstract class can contain non-abstract methods as well as
abstract methods. An interface is like a “pure” abstract class, which contains only abstract
methods.
Note that the methods declared in an interface must be public and abstract. In fact, since
that is the only option, it is not necessary to specify either of these modifiers in the declaration.
In addition to method declarations, an interface can also include variable declarations. The
variables must be "public static final" and effectively become public static final variables
in every class that implements the interface. In fact, since the variables can only be public and
static and final, specifying the modifiers is optional. For example,
public interface ConversionFactors {
int INCHES PER FOOT = 12;
int FEET PER YARD = 3;
int YARDS PER MILE = 1760;
}
This is a convenient way to define named constants that can be used in several classes. A class
that implements ConversionFactors can use the constants defined in the interface as if they were
defined in the class.
Note in particular that any variable that is defined in an interface is a constant. It’s not
really a variable at all. An interface cannot add instance variables to the classes that implement
it.
An interface can extend one or more other interfaces. For example, if Strokeable is the
interface given above and Fillable is an interface that defines a method fill(g), then we could
define
public interface Drawable extends Strokeable, Fillable {
// (more methods/constants could be defined here)
}
A concrete class that implements Drawable must then provide implementations for the stroke()
method from Strokeable and the draw() method from Fillable, as well as for any abstract
methods specified directly in the Drawable interface.
An interface is usually defined in its own .java file, whose name must match the name of
the interface. For example, Strokeable would be defined in a file named Strokeable.java.
Just like a class, an interface can be in a package and can import things from other packages.
A default method in an interface must be marked with the modifier default. It can op-
tionally be marked public but, as for everything else in interfaces, default methods are auto-
matically public and the public modifier can be omitted. Here is an example:
public interface Readable { // represents a source of input
public char readChar(); // read the next character from the input
default public String readLine() { // read up to the next line feed
StringBuilder line = new StringBuilder();
char ch = readChar();
while (ch != ’\n’) {
line.append(ch);
ch = readChar();
}
return line.toString();
}
}
A concrete class that implements this interface must provide an implementation for readChar().
It will inherit a definition for readLine() from the interface, but can provide a new definition if
necessary. When a class includes an implementation for a default method, the implementation
given in the class overrides the default method from the interface.
Note that the default readLine() calls the abstract method readChar(), whose definition
will only be provided in an implementing class. The reference to readChar() in the definition
is polymorphic. The default implementation of readLine() is one that would make sense
in almost any class that implements Readable. Here’s a rather silly example of a class that
implements Readable, including a main() routine that tests the class. Can you figure out what
it does?
public class Stars implements Readable {
public char readChar() {
if (Math.random() > 0.02)
return ’*’;
else
return ’\n’;
}
public static void main(String[] args) {
Stars stars = new Stars();
for (int i = 0 ; i < 10; i++ ) {
String line = stars.readLine(); // Calls the default method!
System.out.println( line );
}
}
}
Default methods provide Java with a capability similar to something called a “mixin” in
other programming languages, namely the ability to mix functionality from another source into
a class. Since a class can implement several interfaces, it is possible to mix in functionality
from several different sources.
CHAPTER 5. OBJECTS AND CLASSES 254
the work of a larger class. And, more seriously, there are other good reasons for nesting the
definition of one class inside another class.
In Java, a nested class is any class whose definition is inside the definition of another
class. (In fact, a class can even be nested inside a method, which must, of course, itself be
inside a class.) Nested classes can be either named or anonymous. I will come back to the
topic of anonymous classes later in this section. A named nested class, like most other things
that occur in classes, can be either static or non-static. Interfaces, like classes, can be nested
inside class definitions and can be either static or non-static. (In fact, interface definitions can
contain static nested classes and interfaces, but that is not something that you will see in this
textbook.)
When you compile the above class definition, two class files will be created. Even though
the definition of Line is nested inside WireFrameModel, the compiled Line class is stored in a
separate file. The name of the class file for Line will be WireFrameModel$Line.class.
If game is a variable of type PokerGame, then, conceptually, game contains its own copy of
the Player class. In an instance method of a PokerGame object, a new Player object would
be created by saying “new Player()”, just as for any other class. (A Player object could be
created outside the PokerGame class with an expression such as “game.new Player()”. Again,
however, this is rare.) The Player object will have access to the deck and pot instance variables
in the PokerGame object. Each PokerGame object has its own deck and pot and Players.
Players of that poker game use the deck and pot for that game; players of another poker game
use the other game’s deck and pot. That’s the effect of making the Player class non-static.
This is the most natural way for players to behave. A Player object represents a player of one
particular poker game. If Player were an independent class or a static nested class, on the
other hand, it would represent the general idea of a poker player, independent of a particular
poker game.
This constructor defines a new class, without giving it a name. At run time, it creates an
object that belongs to that class. This form of the new operator can be used in any statement
where a regular “new” could be used. The intention of this expression is to create: “a new object
belonging to a class that is the same as hsuperclass-or-interfacei but with these hmethods-and-
variablesi added.” The effect is to create a uniquely customized object, just at the point in
the program where you need it. Note that it is possible to base an anonymous class on an
interface, rather than a class. In this case, the anonymous class must implement the interface
by defining all the methods that are declared in the interface. If an interface is used as a base,
the hparameter-listi must be empty. Otherwise, it can contain parameters for a constructor in
the hsuperclassi.
For now, we will look at one not-very-plausible example. Suppose that Drawable is an
interface defined as:
public interface Drawable {
public void draw(GraphicsContext g);
}
Suppose that we want a Drawable object that draws a filled, red, 100-pixel square. Rather than
defining a new, separate class and then using that class to create the object, we can use an
anonymous class to create the object in one statement:
Drawable redSquare = new Drawable() {
public void draw(GraphicsContext g) {
g.setFill(Color.RED);
g.fillRect(10,10,100,100);
}
};
CHAPTER 5. OBJECTS AND CLASSES 258
Then redSquare refers to an object that implements Drawable and that draws a red square
when its draw() method is called. By the way, the semicolon at the end of the statement is not
part of the class definition; it’s the semicolon that is required at the end of every declaration
statement.
Anonymous classes are often used for actual parameters. For example, consider the following
simple method, which draws a Drawable in two different graphics contexts:
void drawTwice( GraphicsContext g1, GraphicsContext g2, Drawable figure ) {
figure.draw(g1);
figure.draw(g2);
}
When calling this method, the third parameter can be created using an anonymous inner class.
For example:
drawTwice( firstG, secondG, new Drawable() {
void draw(GraphicsContext g) {
g.fillOval(10,10,100,100);
}
} );
When a Java class is compiled, each anonymous nested class will produce a separate
class file. If the name of the main class is MainClass, for example, then the names of the
class files for the anonymous nested classes will be MainClass$1.class, MainClass$2.class,
MainClass$3.class, and so on.
Of course, in this example, Drawable is a functional interface, and we could use lambda
expressions (Section 4.5) instead of anonymous classes. The last example could then be written
simply
drawTwice( firstG, secondG, g -> g.fillOval(10,10,100,100) );
and redSquare could be defined as
Drawable redSquare = g -> {
g.setFill(Color.RED);
g.fillRect(10,10,100,100);
};
This approach has the advantage that it does not create an extra .class file. However, lambda
expressions can only be used with functional interfaces, while anonymous classes can be used
with any interface or class. Before Java 8, anonymous classes were often used for handling
events in GUI programs. With Java 8 and JavaFX, they can be mostly replaced in that context
by lambda expressions.
defined above, then any object that implements Drawable can be passed as a parameter to that
subroutine. And that object can be defined by a local class.
In an example earlier in this section, we passed a customized object of type Drawable to the
drawTwice() method, which takes a parameter of type Drawable. In that example, the class
was an anonymous inner class. Local classes are often anonymous, but that is not required. It
is also true that anonymous classes are often local classes, but that is also not required. For
example, an anonymous class could be used to define the initial value of a global variable. In
that case, the anonymous class is not enclosed in any subroutine and therefore is not local.
The definition of a local class can use local variables from the subroutine where it is defined.
It can also use parameters to that subroutine. However, there is a restriction on the use of such
variables and parameters in a local class: The local variable or parameter must be declared
to be final or, if it is not explicitly declared final, then it must be “effectively final.” A
parameter is effectively final if its value is not changed inside the subroutine (including in any
local class that references the parameter). A local variable is effectively final if its value is never
changed after it is initialized. Note that there is no such restriction on global variables that are
used in the definition of a local class.
The same restriction on the use of local variables also applies to lambda expressions, which
are very similar to anonymous classes. Here is an example using the stream API and the
Runnable interface, which are discussed in Section 4.5. This subroutine will print out the
numbers 1 to 10 in some indeterminate order (since it uses a parallel stream):
static void print1to10() {
ArrayList<Runnable> printers = new ArrayList<>();
for (int i = 1; i <= 10; i++) {
int x = i;
printers.add( () -> System.out.println(x) );
}
printers.parallelStream().forEach( r -> r.run() );
}
The local variable x is effectively final and therefore can be used in the lambda expression.
On the other hand, it would have been illegal to use the variable i directly in the lambda
expression, since i is not effectively final; its value is changed when i++ is executed.
Exercises 260
1. In all versions of the PairOfDice class in Section 5.2, the instance variables die1 and die2 (solution)
are declared to be public. They really should be private, so that they would be protected
from being changed from outside the class. Write another version of the PairOfDice class
in which the instance variables die1 and die2 are private. Your class will need “getter”
methods that can be used to find out the values of die1 and die2. (The idea is to protect
their values from being changed from outside the class, but still to allow the values to be
read.) Include other improvements in the class, including at least a toString() method.
Test your class with a short program that counts how many times a pair of dice is rolled,
before the total of the two dice is equal to two.
numbers have been entered, print out each of the six statistics that are available from
calc.
3. This problem uses the PairOfDice class from Exercise 5.1 and the StatCalc class from (solution)
Exercise 5.2.
The program in Exercise 4.4 performs the experiment of counting how many times a
pair of dice is rolled before a given total comes up. It repeats this experiment 10000 times
and then reports the average number of rolls. It does this whole process for each possible
total (2, 3, . . . , 12).
Redo that exercise. But instead of just reporting the average number of rolls, you
should also report the standard deviation and the maximum number of rolls. Use a
PairOfDice object to represent the dice. Use a StatCalc object to compute the statistics.
(You’ll need a new StatCalc object for each possible total, 2, 3, . . . , 12. You can use a
new pair of dice if you want, but it’s not required.)
4. The BlackjackHand class from Subsection 5.5.1 is an extension of the Hand class from Sec- (solution)
tion 5.4. The instance methods in the Hand class are discussed in that section. In addition
to those methods, BlackjackHand includes an instance method, getBlackjackValue(),
which returns the value of the hand for the game of Blackjack. For this exercise, you will
also need the Deck and Card classes from Section 5.4.
A Blackjack hand typically contains from two to six cards. Write a program to test the
BlackjackHand class. You should create a BlackjackHand object and a Deck object. Pick
a random number between 2 and 6. Deal that many cards from the deck and add them to
the hand. Print out all the cards in the hand, and then print out the value computed for
the hand by getBlackjackValue(). Repeat this as long as the user wants to continue.
In addition to TextIO.java, your program will depend on Card.java, Deck.java,
Hand.java, and BlackjackHand.java.
5. Write a program that lets the user play Blackjack. The game will be a simplified version (solution)
of Blackjack as it is played in a casino. The computer will act as the dealer. As in
the previous exercise, your program will need the classes defined in Card.java, Deck.java,
Hand.java, and BlackjackHand.java. (This is the longest and most complex program that
has come up so far in the exercises.)
You should first write a subroutine in which the user plays one game. The subroutine
should return a boolean value to indicate whether the user wins the game or not. Return
true if the user wins, false if the dealer wins. The program needs an object of class
Deck and two objects of type BlackjackHand, one for the dealer and one for the user.
The general object in Blackjack is to get a hand of cards whose value is as close to 21 as
possible, without going over. The game goes like this.
• First, two cards are dealt into each player’s hand. If the dealer’s hand has a value of
21 at this point, then the dealer wins. Otherwise, if the user has 21, then the user
wins. (This is called a “Blackjack”.) Note that the dealer wins on a tie, so if both
players have Blackjack, then the dealer wins.
• Now, if the game has not ended, the user gets a chance to add some cards to her
hand. In this phase, the user sees her own cards and sees one of the dealer’s two
cards. (In a casino, the dealer deals himself one card face up and one card face down.
All the user’s cards are dealt face up.) The user makes a decision whether to “Hit”,
Exercises 262
which means to add another card to her hand, or to “Stand”, which means to stop
taking cards.
• If the user Hits, there is a possibility that the user will go over 21. In that case, the
game is over and the user loses. If not, then the process continues. The user gets to
decide again whether to Hit or Stand.
• If the user Stands, the game will end, but first the dealer gets a chance to draw cards.
The dealer only follows rules, without any choice. The rule is that as long as the
value of the dealer’s hand is less than or equal to 16, the dealer Hits (that is, takes
another card). The user should see all the dealer’s cards at this point. Now, the
winner can be determined: If the dealer has gone over 21, the user wins. Otherwise,
if the dealer’s total is greater than or equal to the user’s total, then the dealer wins.
Otherwise, the user wins.
Two notes on programming: At any point in the subroutine, as soon as you know who
the winner is, you can say “return true;” or “return false;” to end the subroutine
and return to the main program. To avoid having an overabundance of variables in your
subroutine, remember that a function call such as userHand.getBlackjackValue() can
be used anywhere that a number could be used, including in an output statement or in
the condition of an if statement.
Write a main program that lets the user play several games of Blackjack. To make
things interesting, give the user 100 dollars, and let the user make bets on the game. If
the user loses, subtract the bet from the user’s money. If the user wins, add an amount
equal to the bet to the user’s money. End the program when the user wants to quit or
when she runs out of money.
6. Exercise 4.8 asked you to write a program that administers a 10-question addition quiz. (solution)
Rewrite that program so that it uses the following class to represent addition questions:
public class AdditionQuestion {
private int a, b; // The numbers in the problem.
public AdditionQuestion() { // constructor
a = (int)(Math.random() * 50 + 1);
b = (int)(Math.random() * 50);
}
public String getQuestion() {
return "What is " + a + " + " + b + " ?";
}
public int getCorrectAnswer() {
return a + b;
}
}
7. Rewrite the program from the previous exercise so that it administers a quiz with several (solution)
different kinds of questions. In the previous exercise, you used a class to represent addition
questions. For this exercise, you will use the following interface, or an equivalent abstract
class, to represent the more general idea of a question that has an integer as its answer:
Exercises 263
Quiz on Chapter 5
(answers)
1. Object-oriented programming uses classes and objects. What are classes and what are
objects? What is the relationship between classes and objects?
2. Explain carefully what null means in Java, and why this special value is necessary.
4. Suppose that Kumquat is the name of a class and that fruit is a variable of type Kumquat.
What is the meaning of the statement “fruit = new Kumquat();”? That is, what does
the computer do when it executes this statement? (Try to give a complete answer. The
computer does several things.)
7. Modify the following class so that the two instance variables are private and there is a
getter method and a setter method for each instance variable:
public class Player {
String name;
int score;
}
8. Explain why the class Player that is defined in the previous question has an instance
method named toString(), even though no definition of this method appears in the
definition of the class.
10. Java uses “garbage collection” for memory management. Explain what is meant here by
garbage collection. What is the alternative to garbage collection?
11. What is an abstract class, and how can you recognize an abstract class in Java?
13. For this problem, you should write a very simple but complete class. The class represents
a counter that counts 0, 1, 2, 3, 4, . . . . The name of the class should be Counter. It has
one private instance variable representing the value of the counter. It has two instance
methods: increment() adds one to the counter value, and getValue() returns the current
counter value. Write a complete definition for the class, Counter.
14. This problem uses the Counter class from the previous question. The following program
segment is meant to simulate tossing a coin 100 times. It should use two Counter objects,
headCount and tailCount, to count the number of heads and the number of tails. Fill in
the blanks so that it will do so:
Quiz 265
15. Explain why it can never make sense to test “if (obj.equals(null))”.
Chapter 7
Computers get a lot of their power from working with data structures. A data
structure is an organized collection of related data. An object is a data structure, but this
type of data structure—consisting of a fairly small number of named instance variables—is just
the beginning. In many cases, programmers build complicated data structures by hand, by
linking objects together. We’ll look at these custom-built data structures in Chapter 9. But
there is one type of data structure that is so important and so basic that it is built into every
programming language: the array.
You have already encountered arrays in Section 3.8 and Subsection 5.1.4. We continue the
study of arrays in this chapter, including some new details of their use and some additional
array-processing techniques. In particular, we will look at the important topic of algorithms
for searching and sorting an array.
An array has a fixed size that can’t be changed after the array is created. But in many cases,
it is useful to have a data structure that can grow and shrink as necessary. In this chapter, we
will look at a standard class, ArrayList, that represents such a data structure.
333
CHAPTER 7. ARRAYS AND ARRAYLISTS 334
For example, if A is an array of type int[ ], then we could print all the values from A with
the for-each loop:
for ( int item : A )
System.out.println( item );
and we could add up all the positive integers in A with:
int sum = 0; // This will be the sum of all the positive numbers in A.
for ( int item : A ) {
if (item > 0)
sum = sum + item;
}
The for-each loop is not always appropriate. For example, there is no simple way to use
it to process the items in just a part of an array, or to process the elements in reverse order.
However, it does make the code a little simpler when you do want to process all the values, in
order. since it eliminates any need to use array indices.
It’s important to note that a for-each loop processes the values in the array, not the
elements (where an element means the actual memory location that is part of the array). For
example, consider the following incorrect attempt to fill an array of integers with 17’s:
int[] intList = new int[10];
for ( int item : intList ) { // INCORRECT! DOES NOT MODIFY THE ARRAY!
item = 17;
}
The assignment statement item = 17 assigns the value 17 to the loop control variable, item.
However, this has nothing to do with the array. When the body of the loop is executed, the
value from one of the elements of the array is copied into item. The statement item = 17
replaces that copied value but has no effect on the array element from which it was copied; the
value in the array is not changed. The loop is equivalent to
int[] intList = new int[10];
for ( int i = 0; i < intList.length; i++ ) {
int item = intList[i];
item = 17;
}
which certainly does not change the value of any element in the array.
When this method is called, the subroutine call statement must have one actual parameter of
type GraphicsContext, which can be followed by any number of actual parameters of type Point.
For a final example, let’s look at a method that strings together all of the values in a list of
strings into a single, long string. This example uses a for-each loop to process the array:
public static String concat( String... values ) {
StringBuilder buffer; // Use a StringBuilder for more efficient concatenation.
buffer = new StringBuilder(); // Start with an empty StringBuilder.
for ( String str : values ) { // A "for each" loop for processing the values.
buffer.append(str); // Add string to the buffer.
}
return buffer.toString(); // return the contents of the buffer
}
Given this method definition, the method call concat("Hello", "World") would return the
string “HelloWorld”, and concat() would return an empty string. Since a variable arity method
can also accept an array as actual parameter, we could also call concat(lines) where lines
is of type String[ ]. This would concatenate all the elements of the array into a single string.
The second parameter in a call to createMenu is an array of strings. The array that is passed
as an actual parameter could be created in place, using the new operator. For example, we can
use the following statement to create an entire File menu:
Menu fileMenu = createMenu( "File",
new String[] { "New", "Open", "Close", null, "Quit" } );
This should convince you that being able to create and use an array “in place” in this way can
be very convenient, in the same way that anonymous inner classes are convenient. (However,
this example could have been done even more conveniently if createMenu() had been written
as a variable arity method!)
By the way, it is perfectly legal to use the “new BaseType[] { ... }” syntax instead of
the array initializer syntax in the declaration of an array variable. For example, instead of
saying:
int[] primes = { 2, 3, 5, 7, 11, 13, 17, 19 };
you can say, equivalently,
int[] primes = new int[] { 2, 3, 5, 7, 11, 17, 19 };
In fact, rather than use a special notation that works only in the context of declaration state-
ments, I sometimes prefer to use the second form.
∗ ∗ ∗
One final note: For historical reasons, an array declaration such as
int[] list;
can also be written as
int list[];
which is a syntax used in the languages C and C++. However, this alternative syntax does not
really make much sense in the context of Java, and it is probably best avoided. After all, the
intent is to declare a variable of a certain type, and the name of that type is “int[ ]”. It makes
sense to follow the “htype-namei hvariable-namei;” syntax for such declarations.
CHAPTER 7. ARRAYS AND ARRAYLISTS 339
game could be stored in an array, playerList, of type Player[ ]. Since the number of players can
change, you will follow the partially full array pattern, and you will need a variable, playerCt,
to record the number of players currently in the game. Assuming that there will never be more
than 10 players in the game, you could declare the variables as:
Player[] playerList = new Player[10]; // Up to 10 players.
int playerCt = 0; // At the start, there are no players.
After some players have joined the game, playerCt will be greater than 0, and
the player objects representing the players will be stored in the array elements
playerList[0], playerList[1], . . . , playerList[playerCt-1]. Note that the array ele-
ment playerList[playerCt] is not in use: Besides being the number of items in the array,
playerCt is also the index of the next open spot in the array. The procedure for adding a new
player, newPlayer, to the game is simple:
playerList[playerCt] = newPlayer; // Put new player in next
// available spot.
playerCt++; // And increment playerCt to count the new player.
But deleting a player from the game is a little harder, since you don’t want to leave a “hole”
in the array where the deleted player used to be. Suppose you want to delete the player at
index k in playerList. The number of players goes down by one, so one fewer space is used
in the array. If you are not worried about keeping the players in any particular order, then
one way to delete player number k is to move the player from the last occupied position in the
array into position k and then to decrement the value of playerCt:
playerList[k] = playerList[playerCt - 1];
playerCt--;
The player previously in position k is no longer in the array, so we have deleted that player
from the list. The player previously in position playerCt - 1 is now in the array twice. But
it’s only in the occupied or valid part of the array once, since playerCt has decreased by one.
Remember that every element of the array has to hold some value, but only the values in
positions 0 through playerCt - 1 will be looked at or processed in any way. (By the way, you
should think about what happens if the player that is being deleted is in the last position in
the list. The code does still work in this case. What exactly happens?)
Suppose that when deleting the player in position k, you’d like to keep the remaining players
in the same order. (Maybe because they take turns in the order in which they are stored in the
array.) To do this, all the players in positions k+1 and following must move up one position in
the array. Player k+1 replaces player k, who is out of the game. Player k+2 fills the spot left
open when player k+1 is moved. And so on. The code for this is
for (int i = k+1; i < playerCt; i++) {
playerList[i-1] = playerList[i];
}
playerCt--;
Here is an illustration of the two ways of deleting an item from a partially full array. Here,
player “C” is being deleted:
CHAPTER 7. ARRAYS AND ARRAYLISTS 341
playerList playerList
playerCt ½ 5 A playerCt 6 Þ
â
B B
F
C ÚÆÊÉÁÏÇÊ×velyÄ ÇÆÆ ß D
To delete "¾¿ ÀÁÂÃ
the arrayÄ ÅÆÇÈÉÁ¾Ê D ÊÐÉ ×ÊÉÃÌ ÊÐÇÊ D à
ËÁÂÅÌ ÀÁÂÃ Í Ê ΠE ÀÂÆÆÂÔ ¾ ×Ï the à á
ÇÏË ÊÐÉ Ñ ÃÂÒÉÌ ÇÁÁÇÈ ÓÇÏ ÃÂÒÉ
F ÕÅ ÂÏÉ ÌÅace. á
into the ÌÅÇÓÉ
ÔÐÉÁÉ ¾ ÕÌed to be. D replÇÓÉÌ ¾Û
ÖÐÉ ÌÅace that .. Ü ÁÉÅÆÇÓÉÌ ÝÛ ..
ÕÌÉË ÊÂ ÐÂÆË ÊÐÉ Ñ . Ñ ÁÉÅÆÇÓÉÌ Ü. .
×Ì Ï ÆÂÏØer in uÌÉÙ
∗ ∗ ∗
This leaves open the question of what happens when a partially full array becomes full, but
you still want to add more items to it? We can’t change the size of the array—but we can make
a new, bigger array and copy the data from the old array into the new array. But what does it
mean to copy an array in the first place?
Suppose that A and B are array variables, with the same base type, and that A already refers
to an array. Suppose that we want B to refer to a copy of A. The first thing to note is that the
assignment statement
B = A;
does not make a copy of A. Arrays are objects, and an array variable can only hold a pointer
to an array. The assignment statement copies the pointer from A into B, and the result is that
A and B now point to the same array. For example, A[0] and B[0] are just different names for
exactly the same array element. To make B refer to a copy of A, we need to make an entirely
new array and copy all the items from A into B. Let’s say that A and B are of type double[ ].
Then to make a copy of A, we can say
double[] B;
B = new double[A.length]; // Make a new array with the same length as A.
for ( int i = 0; i < A.length; i++ ) {
B[i] = A[i];
}
To solve the problem of adding to a partially full array that has become full, we just need
to make a new array that is bigger than the existing array. The usual choice is to make a
new array twice as big as the old. We need to meet one more requirement: At the end, the
variable that referred to the old array must now point to the new array. That variable is what
gives us access to the data, and in the end, the data is in the new array. Fortunately, a simple
assignment statement will make the variable point to the correct array. Let’s suppose that we
are using playerList and playerCt to store the players in a game, as in the example above,
and we want to add newPlayer to the game. Here is how we can do that even if the playerList
array is full:
if ( playerCt == playerList.length ) {
// The number of players is already equal to the size of the array.
// The array is full. Make a new array that has more space.
Player[] temp; // A variable to point to the new array.
temp = new Player[ 2*playerList.length ]; // Twice as big as the old array.
for ( int i = 0; i < playerList.length; i++ ) {
CHAPTER 7. ARRAYS AND ARRAYLISTS 342
temp[i] = playerList[i]; // Copy item from old array into new array.
}
playerList = temp; // playerList now points to the new, bigger array.
}
// At this point, we know that there is room in the array for newPlayer.
playerList[playerCt] = newPlayer;
playerCt++;
After the new array has been created, there is no longer any variable that points to the old
array, so it will be garbage collected.
∗ ∗ ∗
The Arrays class contains other useful methods. I’ll mention a few of them. As with
Arrays.copyOf, there are actually multiple versions of all of these methods, for different array
types.
• Arrays.fill( array, value ) — Fill an entire array with a specified value. The type
of value must be compatible with the base type of the array. For example, assuming
that numlist is an array of type double[ ], then Arrays.fill(numlist,17) will set every
element of numlist to have the value 17.
• Arrays.fill( array, fromIndex, toIndex, value ) — Fills part of the array with
value, starting at index number fromIndex and ending with index number toIndex-1.
Note that toIndex itself is not included.
• Arrays.toString( array ) — A function that returns a String containing all the values
from array, separated by commas and enclosed between square brackets. The values in
the array are converted into strings in the same way they would be if they were printed
out.
• Arrays.sort( array ) — Sorts the entire array. To sort an array means to rearrange
the values in the array so that they are in increasing order. This method works for arrays
of String and arrays of primitive type values (except for boolean, which would be kind of
silly). But it does not work for all arrays, since it must be meaningful to compare any two
values in the array, to see which is “smaller.” We will discuss array-sorting algorithms in
Section 7.4.
• Arrays.sort( array, fromIndex, toIndex ) — Sorts just the elements from
array[fromIndex] up to array[toIndex-1]
• Arrays.binarySearch( array, value ) — Searches for value in the array. The array
must already be sorted into increasing order. This is a function that returns an int. If
the value is found in the array, the return value is the index of an element that contains
that value. If the value does not occur in the array, the return value is -1. We will discuss
the binary search algorithm in Section 7.4.
case 1:
randomFont = font2;
break;
case 2:
randomFont = font3;
break;
case 3:
randomFont = font4;
break;
case 4:
randomFont = font5;
break;
}
In the new version of the program, the five fonts are stored in an array, which is named fonts.
This array is declared as an instance variable of type Font[ ]
Font[] fonts;
The array is created in the constructor, using the array literal syntax from Subsection 7.1.3:
fonts= new Font[] {
Font.font("Times New Roman", FontWeight.BOLD, 20),
Font.font("Arial", FontWeight.BOLD, FontPosture .ITALIC, 28),
Font.font("Verdana", 32),
Font.font(40),
Font.font("Times New Roman", FontWeight.BOLD, FontPosture .ITALIC, 60)
};
This makes it much easier to select one of the fonts at random. It can be done with the
statements
Font randomFont; // One of the 5 fonts, chosen at random.
int fontIndex; // A random number in the range 0 to 4.
fontIndex = (int)(Math.random() * 5);
randomFont = fonts[ fontIndex ];
The switch statement has been replaced by a few lines of code. In fact, the preceding four
lines can be replaced by the single line
Font randomFont = fonts[ (int)(Math.random() * 5) ];
This is a very typical application of arrays. Note that this example uses the random access
property: We can pick an array index at random and go directly to the array element at that
index.
Here is another example of the same sort of thing. Months are often stored as numbers 1, 2,
3, . . . , 12. Sometimes, however, these numbers have to be translated into the names January,
February, . . . , December. The translation can be done very easily with an array. The array
can be declared and initialized as
static String[] monthName = { "January", "February", "March",
"April", "May", "June",
"July", "August", "September",
"October", "November", "December" };
If mnth is a variable that holds one of the integers 1 through 12, then monthName[mnth-1] is the
name of the corresponding month. We need the “-1” because months are numbered starting
from 1, while array elements are numbered starting from 0. Simple array indexing does the
translation for us!
CHAPTER 7. ARRAYS AND ARRAYLISTS 347
/**
* Returns the number of items currently in the array.
*/
public int size() {
return itemCt;
}
/**
* Adds a new item to the end of the array. The size increases by one.
*/
public void add(int item) {
if (itemCt == items.length)
items = Arrays.copyOf( items, 2*items.length );
items[itemCt] = item;
itemCt++;
}
/**
* Removes the item at a given index in the array. The size of the array
* decreases by one. Items following the removed item are moved up one
* space in the array.
* Throws ArrayIndexOutOfBoundsException if the index is not valid.
*/
public void remove(int index) {
if ( index < 0 || index >= itemCt )
throw new ArrayIndexOutOfBoundsException("Illegal index, " + index);
for (int j = index+1; j < itemCt; j++)
items[j-1] = items[j];
itemCt--;
}
} // end class DynamicArrayOfInt
Everything here should be clear, except possibly why the original size of the items array is 8. In
fact, the number 8 is arbitrary and has no effect on the functionality of the class. Any positive
integer would work, but it doesn’t make sense for the array to start off very big. The array will
grow as needed if the number of items turns out to be large.
The example ReverseInputNumbers.java used a partially full array of int to print a list
of input numbers in the reverse of the order in which they are input. In that program, an
ordinary array of length 100 was used to hold the numbers. In any given run of the program,
the size of the array could be much too large, or it could be too small, resulting in an exception.
The program can now be written using a DynamicArrayOfInt, which will adapt itself to any
reasonable number of inputs. For the program, see ReverseWithDynamicArray.java. It’s a
silly program, but the principle holds in any application where the amount of data cannot be
predicted in advance: The size of a dynamic data structure can adapt itself to any amount of
data, limited only by the amount of memory available to the program.
This is a nice example, but there is a real problem with it. Suppose that we want to have a
dynamic array of String. We can’t use a DynamicArrayOfInt object to hold strings, so it looks
like we need to write a whole new class, DynamicArrayOfString. If we want a dynamic array to
store players in a game, we would need a class DynamicArrayOfPlayer. And so on. It looks like
we have to write a dynamic array class for every possible type of data! That can’t be right! In
fact, Java has a solution to this problem, a standard class that implements dynamic arrays and
CHAPTER 7. ARRAYS AND ARRAYLISTS 349
can work with any type of data. The class is called ArrayList, and we’ll see how it works in the
next section.
7.3 ArrayList
As we have just seen in Subsection 7.2.4, we can easily encode the dynamic array pattern
into a class, but it looks like we need a different class for each data type. In fact, Java has a
feature called “parameterized types” that makes it possible to avoid the multitude of classes,
and Java has a single class named ArrayList that implements the dynamic array pattern for all
data types.
error, and the program will not compile. However, note that objects belonging to a subclass of
T can be added to the list, since objects belonging to a subclass of T are still considered to be
of type T. Thus, for example, a variable of type ArrayList<Pane> can be used to hold objects
of type BorderPane, TilePane, GridPane, or any other subclass of Pane. (Of course, this is the
same way arrays work: An object of type T[ ] can hold objects belonging to any subclass of T.)
Similarly, if T is an interface, then any object that implements interface T can be added to the
list.
An object of type ArrayList<T> has all of the instance methods that you would expect in
a dynamic array implementation. Here are some of the most useful. Suppose that list is a
variable of type ArrayList<T>. Then we have:
• list.size() — This function returns the current size of the list, that is, the number of
items currently in the list. The only valid positions in the list are numbers in the range
0 to list.size()-1. Note that the size can be zero. A call to the default constructor
new ArrayList<T>() creates a list of size zero.
• list.add(obj) — Adds an object onto the end of the list, increasing the size by 1. The
parameter, obj, can refer to an object of type T, or it can be null.
• list.get(N) — This function returns the value stored at position N in the list. The
return type of this function is T. N must be an integer in the range 0 to list.size()-1.
If N is outside this range, an error of type IndexOutOfBoundsException occurs. Calling
this function is similar to referring to A[N] for an array, A, except that you can’t use
list.get(N) on the left side of an assignment statement.
• list.set(N, obj) — Assigns the object, obj, to position N in the ArrayList, replacing the
item previously stored at position N. The parameter obj must be of type T. The integer
N must be in the range from 0 to list.size()-1. A call to this function is equivalent to
the command A[N] = obj for an array A.
• list.clear() — Removes all items from the list, setting its size to zero.
• list.remove(N) — For an integer, N, this removes the N-th item in the ArrayList. N must
be in the range 0 to list.size()-1. Any items in the list that come after the removed
item are moved up one position. The size of the list decreases by 1.
• list.remove(obj) — If the specified object occurs somewhere in the list, it is removed
from the list. Any items in the list that come after the removed item are moved up one
position. The size of the ArrayList decreases by 1. If obj occurs more than once in the
list, only the first copy is removed. If obj does not occur in the list, nothing happens;
this is not an error.
• list.indexOf(obj) — A function that searches for the object, obj, in the list. If the
object is found in the list, then the first position number where it is found is returned. If
the object is not found, then -1 is returned.
For the last two methods listed here, obj is compared to an item in the list by calling
obj.equals(item), unless obj is null. This means, for example, that strings are tested for
equality by checking the contents of the strings, not their location in memory.
Java comes with several parameterized classes representing different data structures. Those
classes make up the Java Collection Framework . Here we consider only ArrayList, but we
will return to this important topic in much more detail in Chapter 10.
By the way, ArrayList can also be used as a non-parametrized type. This means that you
can declare variables and create objects of type ArrayList such as
CHAPTER 7. ARRAYS AND ARRAYLISTS 351
import textio.TextIO;
import java.util.ArrayList;
/**
* Reads a list of non-zero numbers from the user, then prints
* out the input numbers in the reverse of the order in which
* the were entered. There is no limit on the number of inputs.
*/
public class ReverseWithArrayList {
public static void main(String[] args) {
ArrayList<Integer> list;
list = new ArrayList<Integer>();
System.out.println("Enter some non-zero integers. Enter 0 to end.");
while (true) {
System.out.print("? ");
int number = TextIO.getlnInt();
if (number == 0)
break;
list.add(number);
}
System.out.println();
System.out.println("Your numbers in reverse are:");
for (int i = list.size() - 1; i >= 0; i--) {
System.out.printf("%10d%n", list.get(i));
}
}
}
As illustrated in this example, ArrayLists are commonly processed using for loops, in much
the same way that arrays are processed. for example, the following loop prints out all the items
for a variable namelist of type ArrayList<String>:
for ( int i = 0; i < namelist.size(); i++ ) {
String item = namelist.get(i);
System.out.println(item);
}
You can also use for-each loops with ArrayLists, so this example could also be written
for ( String item : namelist ) {
System.out.println(item);
}
When working with wrapper classes, the loop control variable in the for-each loop can be
a primitive type variable. This works because of unboxing. For example, if numbers is of type
ArrayList<Double>, then the following loop can be used to add up all the values in the list:
double sum = 0;
for ( double num : numbers ) {
sum = sum + num;
}
This will work as long as none of the items in the list are null. If there is a possibility of null
values, then you will want to use a loop control variable of type Double and test for nulls. For
example, to add up all the non-null values in the list:
CHAPTER 7. ARRAYS AND ARRAYLISTS 354
double sum;
for ( Double num : numbers ) {
if ( num != null ) {
sum = sum + num; // Here, num is SAFELY unboxed to get a double.
}
}
∗ ∗ ∗
For a more complete and useful example, we will look at the program SimplePaint2.java.
This is a much improved version of SimplePaint.java from Subsection 6.3.3. In the new program,
the user can sketch curves in a drawing area by clicking and dragging with the mouse. The user
can select the drawing color using a menu. The background color of the drawing area can also
be selected using a menu. And there is a “Control” menu that contains several commands: An
“Undo” command, which removes the most recently drawn curve from the screen, a “Clear”
command that removes all the curves, and a “Use Symmetry” checkbox that turns a symmetry
feature on and off. Curves that are drawn by the user when the symmetry option is on are
reflected horizontally and vertically to produce a symmetric pattern. (Symmetry is there just
to look pretty.)
Unlike the original SimplePaint program, this new version uses a data structure to store
information about the picture that has been drawn by the user. When the user selects a new
background color, the canvas is filled with the new background color, and all of the curves that
were there previously are redrawn on the new background. To do that, we need to store enough
data to redraw all of the curves. Similarly, the Undo command is implemented by deleting the
data for most recently drawn curve, and then redrawing the entire picture using the remaining
data.
The data structure that we need is implemented using ArrayLists. The main data for a
curve consists of a list of the points on the curve. This data is stored in an object of type
ArrayList<Point2D>. (Point2D is standard class in package javafx.geometry. A Point2D can
be constructed from two double values, giving the (x,y) coordinates of a point. A Point2D
object, pt, has getter methods pt.getX() and pt.getY() that return the x and y coordinates.)
But in addition to a list of points on a curve, to redraw the curve, we also need to know its
color, and we need to know whether the symmetry option should be applied to the curve. All
the data that is needed to redraw the curve is grouped into an object of type CurveData that
is defined as a nested class in the program:
private static class CurveData {
Color color; // The color of the curve.
boolean symmetric; // Are horizontal and vertical reflections also drawn?
ArrayList<Point2D> points; // The points on the curve.
}
However, a picture can contain many curves, not just one, so to store all the data necessary to
redraw the entire picture, we need a list of objects of type CurveData. For this list, the program
uses an ArrayList, curves, declared as
ArrayList<CurveData> curves = new ArrayList<CurveData>();
Here we have a list of objects, where each object contains a list of points as part of its data!
Let’s look at a few examples of processing this data structure. When the user clicks the mouse
on the drawing surface, it’s the start of a new curve, and a new CurveData object must be
created. The instance variables in the new CurveData object must also be initialized. Here
CHAPTER 7. ARRAYS AND ARRAYLISTS 355
is the code from the mousePressed() routine that does this, where currentCurve is a global
variable of type CurveData:
currentCurve = new CurveData(); // Create a new CurveData object.
currentCurve.color = currentColor; // The color of a curve is taken from an
// instance variable that represents the
// currently selected drawing color.
currentCurve.symmetric = useSymmetry; // The "symmetric" property of the curve
// is also copied from the current value
// of an instance variable, useSymmetry.
currentCurve.points = new ArrayList<Point2D>(); // A new point list object.
As the user drags the mouse, new points are added to currentCurve, and line segments of the
curve are drawn between points as they are added. When the user releases the mouse, the curve
is complete, and it is added to the list of curves by calling
curves.add( currentCurve );
When the user changes the background color or selects the “Undo” command, the picture
has to be redrawn. The program has a redraw() method that completely redraws the picture.
That method uses the data in curves to draw all the curves. The basic structure is a for-each
loop that processes the data for each individual curve in turn. This has the form:
for ( CurveData curve : curves ) {
.
. // Draw the curve represented by the object, curve, of type CurveData.
.
}
In the body of this loop, curve.points is a variable of type ArrayList<Point2D> that holds the
list of points on the curve. The i-th point on the curve can be obtained by calling the get()
method of this list: curve.points.get(i). This returns a value of type Point2D which has
getter methods named getX() and getY(). We can refer directly to the x-coordinate of the
i-th point as:
curve.points.get(i).getX()
This might seem rather complicated, but it’s a nice example of a complex name that specifies
a path to a desired piece of data: Go to the object, curve. Inside curve, go to points. Inside
points, get the i-th item. And from that item, get the x coordinate by calling its getX()
method. Here is the complete definition of the method that redraws the picture:
private void redraw() {
g.setFill(backgroundColor);
g.fillRect(0,0,canvas.getWidth(),canvas.getHeight());
for ( CurveData curve : curves ) {
g.setStroke(curve.color);
for (int i = 1; i < curve.points.size(); i++) {
// Draw a line segment from point number i-1 to point number i.
double x1 = curve.points.get(i-1).getX();
double y1 = curve.points.get(i-1).getY();
double x2 = curve.points.get(i).getX();
double y2 = curve.points.get(i).getY();
drawSegment(curve.symmetric,x1,y1,x2,y2);
CHAPTER 7. ARRAYS AND ARRAYLISTS 356
}
}
}
drawSegment() is a method that strokes the line segment from (x1,y1) to (x2,y2). If the
first parameter is true, it also draws the horizontal and vertical reflections of that segment.
I have mostly been interested here in discussing how the program uses ArrayList, but I
encourage you to read the full source code, SimplePaint2.java, and to try out the program.
In addition to serving as an example of using parameterized types, it also serves as another
example of creating and using menus. You should be able to understand the entire program.
7.4.1 Searching
There is an obvious algorithm for searching for a particular item in an array: Look at each
item in the array in turn, and check whether that item is the one you are looking for. If so,
the search is finished. If you look at every item without finding the one you want, then you
can be sure that the item is not in the array. It’s easy to write a subroutine to implement this
algorithm. Let’s say the array that you want to search is an array of ints. Here is a method
that will search the array for a specified integer. If the integer is found, the method returns
the index of the location in the array where it is found. If the integer is not in the array, the
method returns the value -1 as a signal that the integer could not be found:
/**
* Searches the array A for the integer N. If N is not in the array,
* then -1 is returned. If N is in the array, then the return value is
* the first integer i that satisfies A[i] == N.
*/
static int find(int[] A, int N) {
for (int index = 0; index < A.length; index++) {
if ( A[index] == N )
return index; // N has been found at this index!
}
// If we get this far, then N has not been found
// anywhere in the array. Return a value of -1.
return -1;
}
This method of searching an array by looking at each item in turn is called linear search.
If nothing is known about the order of the items in the array, then there is really no better
alternative algorithm. But if the elements in the array are known to be in increasing or decreas-
ing order, then a much faster search algorithm can be used. An array in which the elements are
in order is said to be sorted . Of course, it takes some work to sort an array, but if the array
is to be searched many times, then the work done in sorting it can really pay off.
Binary search is a method for searching for a given item in a sorted array. Although
the implementation is not trivial, the basic idea is simple: If you are searching for an item in
a sorted list, then it is possible to eliminate half of the items in the list by inspecting a single
item. For example, suppose that you are looking for the number 42 in a sorted array of 1000
integers. Let’s assume that the array is sorted into increasing order. Suppose you check item
number 500 in the array, and find that the item is 93. Since 42 is less than 93, and since the
elements in the array are in increasing order, we can conclude that if 42 occurs in the array
at all, then it must occur somewhere before location 500. All the locations numbered 500 or
above contain values that are greater than or equal to 93. These locations can be eliminated
as possible locations of the number 42.
The next obvious step is to check location 250. If the number at that location is, say, -21,
then you can eliminate locations before 250 and limit further search to locations between 251
and 499. The next test will limit the search to about 125 locations, and the one after that to
about 62. After just 10 steps, there is only one location left. This is a whole lot better than
looking through every element in the array. If there were a million items, it would still take
only 20 steps for binary search to search the array! (Mathematically, the number of steps is
approximately equal to the logarithm, in the base 2, of the number of items in the array.)
CHAPTER 7. ARRAYS AND ARRAYLISTS 358
In order to make binary search into a Java subroutine that searches an array A for an item
N, we just have to keep track of the range of locations that could possibly contain N. At each
step, as we eliminate possibilities, we reduce the size of this range. The basic operation is to
look at the item in the middle of the range. If this item is greater than N, then the second
half of the range can be eliminated. If it is less than N, then the first half of the range can
be eliminated. If the number in the middle just happens to be N exactly, then the search is
finished. If the size of the range decreases to zero, then the number N does not occur in the
array. Here is a subroutine that implements this idea:
/**
* Searches the array A for the integer N.
* Precondition: A must be sorted into increasing order.
* Postcondition: If N is in the array, then the return value, i,
* satisfies A[i] == N. If N is not in the array, then the
* return value is -1.
*/
static int binarySearch(int[] A, int N) {
int lowestPossibleLoc = 0;
int highestPossibleLoc = A.length - 1;
while (highestPossibleLoc >= lowestPossibleLoc) {
int middle = (lowestPossibleLoc + highestPossibleLoc) / 2;
if (A[middle] == N) {
// N has been found at this index!
return middle;
}
else if (A[middle] > N) {
// eliminate locations >= middle
highestPossibleLoc = middle - 1;
}
else {
// eliminate locations <= middle
lowestPossibleLoc = middle + 1;
}
}
// At this point, highestPossibleLoc < LowestPossibleLoc,
// which means that N is known to be not in the array. Return
// a -1 to indicate that N could not be found in the array.
return -1;
}
with k, if any. And given a key, k, and a value v, add the pair (k,v) to the association list
(replacing the pair, if any, that had the same key value). The two operations are usually called
get and put.
Association lists are very widely used in computer science. For example, a compiler has
to keep track of the location in memory associated with each variable. It can do this with an
association list in which each key is a variable name and the associated value is the address of
that variable in memory. Another example would be a mailing list, if we think of it as associating
an address to each name on the list. As a related example, consider a phone directory that
associates a phone number to each name. We’ll look at a highly simplified version of this
example. (This is not meant to be a realistic way to implement a phone directory!)
The items in the phone directory’s association list could be objects belonging to the class:
class PhoneEntry {
String name;
String phoneNum;
}
The data for a phone directory consists of an array of type PhoneEntry[ ] and an integer
variable to keep track of how many entries are actually stored in the directory. The technique
of dynamic arrays (Subsection 7.2.4) can be used in order to avoid putting an arbitrary limit on
the number of entries that the phone directory can hold. Using an ArrayList would be another
possibility. A PhoneDirectory class should include instance methods that implement the “get”
and “put” operations. Here is one possible simple definition of the class:
/**
* A PhoneDirectory holds a list of names with a phone number for
* each name. It is possible to find the number associated with
* a given name, and to specify the phone number for a given name.
*/
public class PhoneDirectory {
/**
* An object of type PhoneEntry holds one name/number pair.
*/
private static class PhoneEntry {
String name; // The name.
String number; // The associated phone number.
}
private PhoneEntry[] data; // Array that holds the name/number pairs.
private int dataCount; // The number of pairs stored in the array.
/**
* Constructor creates an initially empty directory.
*/
public PhoneDirectory() {
data = new PhoneEntry[1];
dataCount = 0;
}
/**
* Looks for a name/number pair with a given name. If found, the index
* of the pair in the data array is returned. If no pair contains the
* given name, then the return value is -1. This private method is
* used internally in getNumber() and putNumber().
CHAPTER 7. ARRAYS AND ARRAYLISTS 360
*/
private int find( String name ) {
for (int i = 0; i < dataCount; i++) {
if (data[i].name.equals(name))
return i; // The name has been found in position i.
}
return -1; // The name does not exist in the array.
}
/**
* Finds the phone number, if any, for a given name.
* @return The phone number associated with the name; if the name does
* not occur in the phone directory, then the return value is null.
*/
public String getNumber( String name ) {
int position = find(name);
if (position == -1)
return null; // There is no phone entry for the given name.
else
return data[position].number;
}
/**
* Associates a given name with a given phone number. If the name
* already exists in the phone directory, then the new number replaces
* the old one. Otherwise, a new name/number pair is added. The
* name and number should both be non-null. An IllegalArgumentException
* is thrown if this is not the case.
*/
public void putNumber( String name, String number ) {
if (name == null || number == null)
throw new IllegalArgumentException("name and number cannot be null");
int i = find(name);
if (i >= 0) {
// The name already exists, in position i in the array.
// Just replace the old number at that position with the new.
data[i].number = number;
}
else {
// Add a new name/number pair to the array. If the array is
// already full, first create a new, larger array.
if (dataCount == data.length) {
data = Arrays.copyOf( data, 2*data.length );
}
PhoneEntry newEntry = new PhoneEntry(); // Create a new pair.
newEntry.name = name;
newEntry.number = number;
data[dataCount] = newEntry; // Add the new pair to the array.
dataCount++;
}
}
} // end class PhoneDirectory
CHAPTER 7. ARRAYS AND ARRAYLISTS 361
The class defines a private instance method, find(), that uses linear search to find the
position of a given name in the array of name/number pairs. The find() method is used
both in the getNumber() method and in the putNumber() method. Note in particular that
putNumber(name,number) has to check whether the name is in the phone directory. If so, it
just changes the number in the existing entry; if not, it has to create a new phone entry and
add it to the array.
This class could use a lot of improvement. For one thing, it would be nice to use binary
search instead of simple linear search in the getNumber method. However, we could only do
that if the list of PhoneEntries were sorted into alphabetical order according to name. In fact,
it’s really not all that hard to keep the list of entries in sorted order, as you’ll see in the next
subsection.
I will mention that association lists are also called “maps,” and Java has a standard pa-
rameterized type named Map that implements association lists for keys and values of any type.
The implementation is much more efficient than anything you can do with basic arrays. You
will encounter this class in Chapter 10.
35
∗ ∗ ∗
A variation of selection sort is used in the Hand class that was introduced in Subsection 5.4.1.
(By the way, you are finally in a position to fully understand the source code for the Hand class
from that section; note that it uses ArrayList. The source file is Hand.java.)
In the Hand class, a hand of playing cards is represented by an ArrayList<Card>. The objects
stored in the list are of type Card. A Card object contains instance methods getSuit() and
getValue() that can be used to determine the suit and value of the card. In my sorting method,
I actually create a new list and move the cards one-by-one from the old list to the new list.
The cards are selected from the old list in increasing order. In the end, the new list becomes
the hand and the old list is discarded. This is not the most efficient procedure. But hands of
cards are so small that the inefficiency is negligible. Here is the code for sorting cards by suit:
/**
* Sorts the cards in the hand so that cards of the same suit are
* grouped together, and within a suit the cards are sorted by value.
* Note that aces are considered to have the lowest value, 1.
*/
public void sortBySuit() {
ArrayList<Card> newHand = new ArrayList<Card>();
CHAPTER 7. ARRAYS AND ARRAYLISTS 364
7.4.5 Unsorting
I can’t resist ending this section on sorting with a related problem that is much less common,
but is a bit more fun. That is the problem of putting the elements of an array into a random
CHAPTER 7. ARRAYS AND ARRAYLISTS 365
order. The typical case of this problem is shuffling a deck of cards. A good algorithm for
shuffling is similar to selection sort, except that instead of moving the biggest item to the end
of the list, an item is selected at random and moved to the end of the list. Here is a subroutine
to shuffle an array of ints:
/**
* Postcondition: The items in A have been rearranged into a random order.
*/
static void shuffle(int[] A) {
for (int lastPlace = A.length-1; lastPlace > 0; lastPlace--) {
// Choose a random location from among 0,1,...,lastPlace.
int randLoc = (int)(Math.random()*(lastPlace+1));
// Swap items in locations randLoc and lastPlace.
int temp = A[randLoc];
A[randLoc] = A[lastPlace];
A[lastPlace] = temp;
}
}
For the most part, you can ignore the reality and keep the picture of a grid in mind.
Sometimes, though, you will need to remember that each row in the grid is really an array in
itself. These arrays can be referred to as A[0], A[1], and A[2]. Each row is in fact a value of
type int[ ]. It could, for example, be passed to a subroutine that asks for a parameter of type
int[ ].
Some of the consequences of this structure are a little subtle. For example, thinking of a 2D
array, A, as an array of arrays, we see that A.length makes sense and is equal to the number
of rows of A. If A has the usual shape for a 2D array, then the number of columns in A would
be the same as the number of elements in the first row, that is, A[0].length. But there is no
rule that says that all of the rows of A must have the same length (although an array created
with new BaseType[rows][columns] will always have that form). Each row in a 2D array is
a separate one-dimensional array, and each of those arrays can have a different length. In fact,
it’s even possible for a row to be null. For example, the statement
A = new int[3][];
with no number in the second set of brackets, creates an array of 3 elements where all the
elements are null. There are places for three rows, but no actual rows have been created. You
can then create the rows A[0], A[1], and A[2] individually.
CHAPTER 7. ARRAYS AND ARRAYLISTS 367
3 -7 12 13 0 17 21 In a symmetric matrix, 3
the elements above the
-7 -1
-7 -1 5 -2 9 11 2 ÿd shown in red)
12 5 -3 12 22 15 30 duplicate elements 12 5 -3
below the diagonal
6 -2 12 15 13 4 4 (blue). So a symmetric 6 -2 12 15
matrix can be stored 0 9 22 13 35
0 9 22 13 35 1 24
as a "triangular matrix"
17 11 15 4 1 8 -5 with rows of dierent 17 11 15 4 1 8
lengths.
21 2 30 -4 24 -5 16 21 2 30 -4 24 -5 16
It’s easy enough to make a triangular array, if we create each row separately. To create a 7-by-7
triangular array of double, we can use the code segment
double[][] matrix = new double[7][]; // rows have not yet been created!
for (int i = 0; i < 7; i++) {
matrix[i] = new double[i+1]; // Create row i with i + 1 elements.
}
We just have to remember that if we want to know the value of the matrix at (i,j), and if
i < j, then we actually have to get the value of matrix[j][i] in the triangular matrix. And
similarly for setting values. It’s easy to write a class to represent symmetric matrices:
/**
* Represents symmetric n-by-n matrices of real numbers.
*/
public class SymmetricMatrix {
private double[][] matrix; // A triangular matrix to hold the data.
/**
* Creates an n-by-n symmetric matrix in which all entries are 0.
*/
public SymmetricMatrix(int n) {
matrix = new double[n][];
for (int i = 0; i < n; i++)
matrix[i] = new double[i+1];
}
/**
* Returns the matrix entry at position (row,col). (If row < col,
* the value is actually stored at position (col,row).)
*/
public double get( int row, int col ) {
if (row >= col)
return matrix[row][col];
else
return matrix[col][row];
}
/**
* Sets the value of the matrix entry at (row,col). (If row < col,
* the value is actually stored at position (col,row).)
CHAPTER 7. ARRAYS AND ARRAYLISTS 368
*/
public void set( int row, int col, double value ) {
if (row >= col)
matrix[row][col] = value;
else
matrix[col][row] = value;
}
/**
* Returns the number of rows and columns in the matrix.
*/
public int size() {
return matrix.length; // The size is the number of rows.
}
} // end class SymmetricMatrix
This class is in the file SymmetricMatrix.java, and a small program to test it can be found in
TestSymmetricMatrix.java.
By the way, the standard function Arrays.copyOf() can’t make a full copy of a 2D array
in a single step. To do that, you need to copy each row separately. To make a copy of a
two-dimensional array of int, for example:
int[][] B = new int[A.length][]; // B has as many rows as A.
for (int i = 0; i < A.length; i++) {
B[i] = Arrays.copyOf(A[i], A[i].length)); // Copy row i.
}
Here’s a picture of part of a Life board, showing the same board before and after the rules have
been applied. The rules are applied to every cell in the grid. The picture shows how they apply
to four of the cells:
Dead cell with 3 l
neighbors comes to life
L
ll
l
neighbor dies.
L
ll with 5 l
neighbors dies.
L
ll with 3 l
S neighbors l S
The Game of Life is interesting because it gives rise to many interesting and surprising
patterns. (Look it up on Wikipedia.) Here, we are just interested in writing a program to
simulate the game. The complete program can be found in the file Life.java. In the program,
the life grid is shown as a grid of squares in which dead squares are black and living squares
are white. (The program uses MosaicCanvas.java from Section 4.7 to represent the grid, so you
will also need that file to compile and run the program.) In the program, you can fill the life
board randomly with dead and alive cells, or you can use the mouse to set up the game board.
There is a “Step” button that will compute one time-step of the game, and a “Start” button
that will run time steps as an animation.
We’ll look at some of the array processing involved in implementing the Game of Life for
this program. Since a cell can only be alive or dead, it is natural to use a two-dimensional
array of boolean[ ][ ] to represent the states of all the cells. The array is named alive, and
alive[r][c] is true when the cell in row r, column c is alive. The number of rows and the
number of columns are equal and are given by a constant, GRID SIZE. So, for example, to fill
the Life grid with random values, the program uses simple nested for loops:
for (int r = 0; r < GRID SIZE; r++) {
for (int c = 0; c < GRID SIZE; c++) {
// Use a 25% probability that the cell is alive.
alive[r][c] = (Math.random() < 0.25);
}
}
Note that the expression (Math.random() < 0.25) is a true/false value that can be assigned to
a boolean array element. The array is also used to set the color of the cells on the screen. Since
the grid of cells is displayed on screen as a MosaicCanvas, setting the colors is done using the
MosaicCanvas API. Note that the actual drawing is done in the MosaicCanvas class (which has
its own 2D array of type Color[ ][ ] to keep track of the colors of each cell). The Life program just
has to set the colors in the mosaic, using the MosaicCanvas API. This is done in the program
in a method named showBoard() that is called each time the board changes. Again, simple
nested for loops are used to set the color of each square in the grid:
for (int r = 0; r < GRID SIZE; r++) {
for (int c = 0; c < GRID SIZE; c++) {
if (alive[r][c])
display.setColor(r,c,Color.WHITE);
else
display.setColor(r,c,null); // Shows the background color, black.
CHAPTER 7. ARRAYS AND ARRAYLISTS 370
}
}
Of course, the most interesting part of the program is computing the new state of the board
by applying the rules to the current state. The rules apply to each individual cell, so again we
can use nested for loops to work through all the cells on the board, but this time the processing
is more complicated. Note first that we can’t make changes to the values in the array as we work
through it, since we will need to know the old state of a cell when processing its neighboring
cells. In fact, the program uses a second array to hold the new board as it is being created.
When the new board is finished, it can be substituted for the old board. The algorithm goes
like this in pseudocode:
let newboard be a new boolean[][] array
for each row r:
for each column c:
Let N be the number of neighbors of cell (r,c) in the alive array
if ((N is 3) or (N is 2 and alive[r][c]))
newboard[r][c] = true;
else
newboard[r][c] = false;
alive = newboard
Note that at the end of the process, alive is pointing to a new array. This doesn’t matter
as long as the contents of the array represent the new state of the game. The old array will
be garbage collected. The test for whether newboard[r][c] should be true or false might
not be obvious, but it implements the rules correctly. We still need to work on counting the
neighbors. Consider the cell in row r and column c. If it’s not at an edge of the board, then
it’s clear where its neighbors are:
Column c-1
Column c
Column c+1
R
R
R !
The row above row number r is row number r-1, and the row below is r+1. Similarly
for the columns. We just have to look at the values of alive[r-1][c-1], alive[r-1][c],
alive[r-1][c+1], alive[r][c-1], alive[r][c+1], alive[r+1][c-1], alive[r+1][c], and
alive[r+1][c+1], and count the number that are true. (You should make sure that you
understand how the array indexing works here.)
But there is a problem when the cell is along one of the edges of the grid. In that case, some
of the array elements in the list don’t exist, and an attempt to use them will cause an exception.
To avoid the exception, we have to give special consideration to cells along the edges. One idea
is that before referencing any array element, check that the array element actually exists. In
that case, the code for neighbor counting becomes
if (r-1 >= 0 && c-1 >= 0 && alive[r-1][c-1])
N++; // A cell at position (r-1,c-1) exists and is alive.
if (r-1 >= 0 && alive[r-1][c])
N++; // A cell at position (r-1,c) exists and is alive.
CHAPTER 7. ARRAYS AND ARRAYLISTS 371
7.5.3 Checkers
As a final example for this chapter, we’ll look at a more substantial example of using a 2D
array. This is the longest program that we have encountered so far, with 745 lines of code.
CHAPTER 7. ARRAYS AND ARRAYLISTS 372
The program lets two users play checkers against each other. The checkers game is played on
an eight-by-eight board, which is based on an example from Subsection 6.5.1. The players are
called “red” and “black,” after the color of their checkers. I’m not going to explain the rules of
checkers here; possibly you can learn them by trying out the program.
In the program, a player moves by clicking on the piece that they want to move, and then
clicking on the empty square to which it is to be moved. As an aid to the players, any square
that the current player can legally click at a given time is highlighted with a brightly colored
border. The square containing a piece that has been selected to be moved, if any, is surrounded
by a yellow border. Other pieces that can legally be moved are surrounded by a cyan-colored
border. If a piece has already been selected to be moved, each empty square that it can legally
move to is highlighted with a green border. The game enforces the rule that if the current
player can jump one of the opponent’s pieces, then the player must jump. When a player’s
piece becomes a king, by reaching the opposite end of the board, a big white “K” is drawn on
the piece. Here is a picture of the program early in a game. It is black’s turn to move. Black
has selected the piece in the yellow-outlined square to be moved. Black can click one of the
squares outlined in green to complete the move, or can click one of the squares outlined in cyan
to select a different piece to be moved.
I will only cover a part of the programming for this example. I encourage you to read the
complete source code, Checkers.java. It’s long and complex, but with some study, you should
understand all the techniques that it uses. The program is a good example of state-based,
event-driven, object-oriented programming.
∗ ∗ ∗
The data about the pieces on the board are stored in a two-dimensional array. Because of
the complexity of the program, I wanted to divide it into several classes. In addition to the main
class, there are several nested classes. One of these classes is CheckersData, which handles the
data for the board. It is not directly responsible for any part of the graphics or event-handling,
but it provides methods that can be called by other classes that handle graphics and events. It
is mainly this class that I want to talk about.
The CheckersData class has an instance variable named board of type int[][]. The value
of board is set to “new int[8][8]”, an 8-by-8 grid of integers. The values stored in the grid
are defined as constants representing the possible contents of a square on a checkerboard:
CHAPTER 7. ARRAYS AND ARRAYLISTS 373
0 1 2 3 4 5 6 7
A regular black piece can only move “down” the grid. That is, the row number of the square
it moves to must be greater than the row number of the square it comes from. A regular red
piece can only move up the grid. Kings of either color can move in both directions.
One function of the CheckersData class is to take care of changes to the data structures
that need to be made when one of the users moves a checker. An instance method named
makeMove() is provided to do this. When a player moves a piece from one square to another,
the values of two elements in the array are changed. But that’s not all. If the move is a jump,
then the piece that was jumped is removed from the board. (The method checks whether the
move is a jump by checking if the square to which the piece is moving is two rows away from
the square where it starts.) Furthermore, a RED piece that moves to row 0 or a BLACK piece
that moves to row 7 becomes a king. Putting all that into a subroutine is good programming:
the rest of the program doesn’t have to worry about any of these details. It just calls this
makeMove() method:
/**
* Make the move from (fromRow,fromCol) to (toRow,toCol). It is
* ASSUMED that this move is legal! If the move is a jump, the
* jumped piece is removed from the board. If a piece moves
* to the last row on the opponent’s side of the board, the
* piece becomes a king.
*/
void makeMove(int fromRow, int fromCol, int toRow, int toCol) {
board[toRow][toCol] = board[fromRow][fromCol]; // Move the piece.
board[fromRow][fromCol] = EMPTY; // The square it moved from is now empty.
if (fromRow - toRow == 2 || fromRow - toRow == -2) {
// The move is a jump. Remove the jumped piece from the board.
int jumpRow = (fromRow + toRow) / 2; // Row of the jumped piece.
int jumpCol = (fromCol + toCol) / 2; // Column of the jumped piece.
board[jumpRow][jumpCol] = EMPTY;
CHAPTER 7. ARRAYS AND ARRAYLISTS 374
}
if (toRow == 0 && board[toRow][toCol] == RED)
board[toRow][toCol] = RED KING; // Red piece becomes a king.
if (toRow == 7 && board[toRow][toCol] == BLACK)
board[toRow][toCol] = BLACK KING; // Black piece becomes a king.
} // end makeMove()
An even more important function of the CheckersData class is to find legal moves on the
board. In my program, a move in a Checkers game is represented by an object belonging to
the following class:
/**
* A CheckersMove object represents a move in the game of
* Checkers. It holds the row and column of the piece that is
* to be moved and the row and column of the square to which
* it is to be moved. (This class makes no guarantee that
* the move is legal.)
*/
private static class CheckersMove {
int fromRow, fromCol; // Position of piece to be moved.
int toRow, toCol; // Square it is to move to.
CheckersMove(int r1, int c1, int r2, int c2) {
// Constructor. Set the values of the instance variables.
fromRow = r1;
fromCol = c1;
toRow = r2;
toCol = c2;
}
boolean isJump() {
// Test whether this move is a jump. It is assumed that
// the move is legal. In a jump, the piece moves two
// rows. (In a regular move, it only moves one row.)
return (fromRow - toRow == 2 || fromRow - toRow == -2);
}
} // end class CheckersMove.
The CheckersData class has an instance method which finds all the legal moves that are
currently available for a specified player. This method is a function that returns an array
of type CheckersMove[ ]. The array contains all the legal moves, represented as CheckersMove
objects. The specification for this method reads
/**
* Return an array containing all the legal CheckersMoves
* for the specified player on the current board. If the player
* has no legal moves, null is returned. The value of player
* should be one of the constants RED or BLACK; if not, null
* is returned. If the returned value is non-null, it consists
* entirely of jump moves or entirely of regular moves, since
* if the player can jump, only jumps are legal moves.
*/
CheckersMove[] getLegalMoves(int player)
CHAPTER 7. ARRAYS AND ARRAYLISTS 375
another given square is itself non-trivial. The square the player is moving to must actually be
on the board, and it must be empty. Furthermore, regular red and black pieces can only move
in one direction. I wrote the following utility method to check whether a player can make a
given non-jump move:
/**
* This is called by the getLegalMoves() method to determine
* whether the player can legally move from (r1,c1) to (r2,c2).
* It is ASSUMED that (r1,c1) contains one of the player’s
* pieces and that (r2,c2) is a neighboring square.
*/
private boolean canMove(int player, int r1, int c1, int r2, int c2) {
if (r2 < 0 || r2 >= 8 || c2 < 0 || c2 >= 8)
return false; // (r2,c2) is off the board.
if (board[r2][c2] != EMPTY)
return false; // (r2,c2) already contains a piece.
if (player == RED) {
if (board[r1][c1] == RED && r2 > r1)
return false; // Regular red piece can only move down.
return true; // The move is legal.
}
else {
if (board[r1][c1] == BLACK && r2 < r1)
return false; // Regular black piece can only move up.
return true; // The move is legal.
}
} // end canMove()
This method is called by my getLegalMoves() method to check whether one of the possible
moves that it has found is actually legal. I have a similar method that is called to check whether
a jump is legal. In this case, I pass to the method the square containing the player’s piece,
the square that the player might move to, and the square between those two, which the player
would be jumping over. The square that is being jumped must contain one of the opponent’s
pieces. This method has the specification:
/**
* This is called by other methods to check whether
* the player can legally jump from (r1,c1) to (r3,c3).
* It is assumed that the player has a piece at (r1,c1), that
* (r3,c3) is a position that is 2 rows and 2 columns distant
* from (r1,c1) and that (r2,c2) is the square between (r1,c1)
* and (r3,c3).
*/
private boolean canJump(int player, int r1, int c1,
int r2, int c2, int r3, int c3) { . . .
Given all this, you should be in a position to understand the complete getLegalMoves()
method. It’s a nice way to finish off this chapter, since it combines several topics that we’ve
looked at: one-dimensional arrays, ArrayLists, and two-dimensional arrays:
CHAPTER 7. ARRAYS AND ARRAYLISTS 377
moves.add(new CheckersMove(row,col,row-1,col-1));
}
}
}
}
/* If no legal moves have been found, return null. Otherwise, create
an array just big enough to hold all the legal moves, copy the
legal moves from the ArrayList into the array, and return the array.
*/
if (moves.size() == 0)
return null;
else {
CheckersMove[] moveArray = new CheckersMove[moves.size()];
for (int i = 0; i < moves.size(); i++)
moveArray[i] = moves.get(i);
return moveArray;
}
} // end getLegalMoves
The checkers program is complex, and you can be sure that it didn’t just fall together.
It took a good deal of design work to decide what classes and objects would be used, what
methods should be written, and what algorithms the methods should use. The complete source
code is in the file Checkers.java. Take a look!
Exercises 379
1. Write a subroutine that creates an ArrayList containing several different random integers (solution)
in the range from 1 up to some specified maximum. The number of integers and the
maximum allowed value for the integers should be parameters to the subroutine. Write a
main() routine to test your subroutine.
2. Suppose that M is a two-dimensional array that has R rows and C columns. The trans- (solution)
pose of M is defined to be an array T that has C rows and R columns such that
T[i][j] = M[j][i] for each i and j. Write a function that takes an array of type int[ ][ ]
as a parameter, and returns the transpose of that array. (Assume that the parameter is a
typical 2D array in which all the rows have the same length.) Also write a subroutine to
print a 2D array of integers in neat rows and columns, and include a main() routine to
test your work.
3. In Subsection 7.4.4, it is mentioned that the standard sorting method Arrays.sort() is (solution)
much faster and efficient than selection sort. Write a program to test this claim. To be
specific, your program should create a large array filled with random real numbers. It
should use both Arrays.sort() and selectionSort() to sort the array, and it should
time how long it takes to perform each sort. Furthermore, it should do the same thing for
a large array of random Strings. To find the times, you can use System.nanoTime() (see
Subsection 2.3.1 and the example TimedComputation.java).
4. In Exercise 6.2, you wrote a program DragTwoSquares that allows the user to drag a red (solution)
square and a blue square around on a canvas. Write an much improved version where
the user can add squares to a canvas and drag them around. In particular: If the user
shift-clicks or right-clicks the canvas, then the user is trying to drag a square; find the
square that contains the mouse position, if any, and move it as the user drags the mouse.
Other clicks should add squares. You can place the center of the new square at the current
mouse position. To make the picture more visually appealing, give each square a random
color, and when you draw the squares, draw a black outline around each square. (My
program also gives the square a random alpha value between 0.5 and 1.0).
Write a class to represent the data needed for drawing one square, and use an ArrayList
to store the data for all the squares in the picture. If the user drags a square completely
off the canvas, delete it from the list.
5. Write a program that will read a sequence of positive real numbers entered by the user (solution)
and will print the same numbers in sorted order from smallest to largest. The user will
input a zero to mark the end of the input. Assume that at most 100 positive numbers
will be entered. Do not use any built-in function such as Arrays.sort(). Do the sorting
yourself.
6. Write a program that will read a text file selected by the user, and will make an alphabetical (solution)
list of all the different words in that file. All words should be converted to lower case, and
duplicates should be eliminated from the list. The list should be written to an output file
selected by the user. As discussed in Subsection 2.4.4, you can use TextIO to read and
write files. Use a variable of type ArrayList<String> to store the words. It is not easy to
separate a file into words as you are reading it, especially if you want to allow apostrophes
in the middle of a word. You can use the following method in your program:
Exercises 380
/**
* Read the next word from TextIO, if there is one. First, skip past
* any non-letters in the input. If an end-of-file is encountered before
* a word is found, return null. Otherwise, read and return the word.
* A word is defined as a sequence of letters. Also, a word can include
* an apostrophe if the apostrophe is surrounded by letters on each side.
* @return the next word from TextIO, or null if an end-of-file is
* encountered
*/
private static String readNextWord() {
char ch = TextIO.peek(); // Look at next character in input.
while (ch != TextIO.EOF && ! Character.isLetter(ch)) {
// Skip past non-letters.
TextIO.getAnyChar(); // Read the character.
ch = TextIO.peek(); // Look at the next character.
}
if (ch == TextIO.EOF) // Encountered end-of-file
return null;
// At this point, we know the next character is a letter, so read a word.
String word = ""; // This will be the word that is read.
while (true) {
word += TextIO.getAnyChar(); // Append the letter onto word.
ch = TextIO.peek(); // Look at next character.
if ( ch == ’\’’ ) {
// The next character is an apostrophe. Read it, and
// if the following character is a letter, add both the
// apostrophe and the letter onto the word and continue
// reading the word. If the character after the apostrophe
// is not a letter, the word is done, so break out of the loop.
TextIO.getAnyChar(); // Read the apostrophe.
ch = TextIO.peek(); // Look at char that follows apostrophe.
if (Character.isLetter(ch)) {
word += "\’" + TextIO.getAnyChar();
ch = TextIO.peek(); // Look at next char.
}
else
break;
}
if ( ! Character.isLetter(ch) ) {
// If the next character is not a letter, the word is
// finished, so break out of the loop.
break;
}
// If we haven’t broken out of the loop, next char is a letter.
}
return word; // Return the word that has been read.
}
Note that this method will return null when the file has been entirely read. You can use
this as a signal to stop processing the input file.
7. The game of Go Moku (also known as Pente or Five Stones) is similar to Tic-Tac-Toe, (solution)
except that it is played on a much larger board and the object is to get five squares in a
Exercises 381
row rather than three. The board should have 13 rows and 13 columns of squares. Players
take turns placing pieces on a board. A piece can be placed in any empty square. The
first player to get five pieces in a row—horizontally, vertically, or diagonally—wins. If all
squares are filled before either player wins, then the game is a draw. Write a program
that lets two players play Go Moku against each other.
Your program will be simpler than the Checkers program from Subsection 7.5.3. Play
alternates strictly between the two players, and there is no need to highlight the legal
moves. You will only need one nested subclass, a subclass of Canvas to draw the board
and do all the work of the game, like the nested CheckersBoard in the Chekers program.
You will probably want to look at the source code for the checkers program, Checkers.java,
for ideas about the general outline of the program.
The hardest part of the program is checking whether the move that a player makes is
a winning move. To do this, you have to look in each of the four possible directions from
the square where the user has placed a piece. You have to count how many pieces that
player has in a row in that direction. If the number is five or more in any direction, then
that player wins. As a hint, here is part of the code from my program. This code counts
the number of pieces that the user has in a row in a specified direction. The direction is
specified by two integers, dirX and dirY. The values of these variables are 0, 1, or -1, and
at least one of them is non-zero. For example, to look in the horizontal direction, dirX is
1 and dirY is 0.
int ct = 1; // Number of pieces in a row belonging to the player.
int r, c; // A row and column to be examined
r = row + dirX; // Look at square in specified direction.
c = col + dirY;
while ( r >= 0 && r < 13 && c >= 0 && c < 13
&& board[r][c] == player ) {
// Square is on the board, and it
// contains one of the player’s pieces.
ct++;
r += dirX; // Go on to next square in this direction.
c += dirY;
}
r = row - dirX; // Now, look in the opposite direction.
c = col - dirY;
while ( r >= 0 && r < 13 && c >= 0 && c < 13
&& board[r][c] == player ) {
ct++;
r -= dirX; // Go on to next square in this direction.
c -= dirY;
}
Here is a picture of my program, just after black has won the game.
Exercises 382
Quiz 383
Quiz on Chapter 7
(answers)
2. What is the purpose of the following variable-arity method? What are the values of
same(1,2,3), same(17,17,17,17), and same(2)? Why?
static double same( int... value ) {
for (int i = 1; i < value.length; i++) {
if ( value[i-1] != value[i] )
return false;
}
return true;
}
4. What is the main advantage of binary search over linear search? What is the main
disadvantage?
5. What is meant by a dynamic array? What is the advantage of a dynamic array over a
regular array?
9. Write a complete static method that finds the largest value in an array of ints. The
method should have one parameter, which is an array of type int[]. The largest number
in the array should be returned as the value of the method.
10. Suppose that temperature measurements were made on each day of 2018 in each of 100
cities. The measurements have been stored in an array
int[][] temps = new int[100][365];
where temps[c][d] holds the measurement for city number c on the dth day of the year.
Write a code segment that will print out the average temperature, over the course of the
whole year, for each city. The average temperature for a city can be obtained by adding
up all 365 measurements for that city and dividing the answer by 365.0.
12. What is the purpose of the following subroutine? What is the meaning of the value that
it returns, in terms of the value of its parameter?
static double[] sums( double[][] data ) {
double[] answers = new double[ data.length ];
for (int i = 0; i < data.length; i++) {
double sum = 0;
for (int j = 0; j < data[i].length; i++)
sum = sum + data[i][j];
answers[i] = sum;
}
return answers;
}
Chapter 8
In previous chapters, we have covered the fundamentals of programming. The chapters that
follow this one will cover more advanced aspects of programming. The ideas that are presented
will generally be more complex and the programs that use them a little more complicated. This
relatively short chapter is a kind of turning point in which we look at the problem of getting
such complex programs right.
Computer programs that fail are much too common. Programs are fragile. A tiny error
can cause a program to misbehave or crash. Most of us are familiar with this from our own
experience with computers. And we’ve all heard stories about software glitches that cause
spacecraft to crash, web sites to go offline, telephone service to fail, and, in a few cases, people
to die.
Programs don’t have to be as bad as they are. It might well be impossible to guarantee
that programs are problem-free, but careful programming and well-designed programming tools
can help keep the problems to a minimum. This chapter will look at issues of correctness and
robustness of programs. Section 8.2 discusses how you can think about and analyze programs
to make the programs that you write more likely to be correct—possibly even provably correct.
Section 8.3 looks more closely at exceptions and the try..catch statement, and Section 8.4
introduces assertions, another of the tools that Java provides as an aid in writing correct
programs.
In Section 8.5, we look at another issue that is important for programs in the real world:
efficiency. Even a completely correct program is not very useful if it takes an unreasonable
amount of time to run. The last section of this chapter introduces techniques for analyzing the
run time efficiency of algorithms.
Some of the topics in this chapter are the topics of advanced courses in computer science,
and only a brief overview can be given here. But what you do learn here will be useful in the
rest of the book.
385
CHAPTER 8. CORRECTNESS, ROBUSTNESS, EFFICIENCY 386
discovered in simulation. The Mariner 18 space probe was lost because of an error in one
line of a program. The Gemini V space capsule missed its scheduled landing target by
a hundred miles, because a programmer forgot to take into account the rotation of the
Earth.
• In 1990, AT&T’s long-distance telephone service was disrupted throughout the United
States when a newly loaded computer program proved to contain a bug.
Of course, there have been more recent problems. For example, computer software error
contributed to the Northeast Blackout of 2003, one of the largest power outages in history. In
2006, the Airbus A380 was delayed by software incompatibility problems, at a cost of perhaps
billions of dollars. In 2007, a software problem grounded thousands of planes at the Los Angeles
International Airport. On May 6, 2010, a flaw in an automatic trading program apparently
resulted in a 1000-point drop in the Dow Jones Industrial Average.
These are just a few examples. Software problems are all too common. As programmers,
we need to understand why that is true and what can be done about it.
other languages, such as C and C++, it’s up to the programmer to make sure that the index
is within the legal range. Suppose that an array, A, has three locations, A[0], A[1], and A[2].
Then A[3], A[4], and so on refer to memory locations beyond the end of the array. In Java,
an attempt to store data in A[3] will be detected. The program will be terminated (unless the
error is “caught”, as discussed in Section 3.7). In C or C++, the computer will just go ahead
and store the data in memory that is not part of the array. Since there is no telling what that
memory location is being used for, the result will be unpredictable. The consequences could
be much more serious than a terminated program. (See, for example, the discussion of buffer
overflow errors later in this section.)
Pointers are a notorious source of programming errors. In Java, a variable of object type
holds either a pointer to an object or the special value null. Any attempt to use a null
value as if it were a pointer to an actual object will be detected by the system. In some other
languages, again, it’s up to the programmer to avoid such null pointer errors. In my first
Macintosh computer, a long time ago, a null pointer was actually implemented as if it were a
pointer to memory location zero. A program could use a null pointer to change values stored
in memory near location zero. Unfortunately, the Macintosh stored important system data in
those locations. Changing that data could cause the whole system to crash, a consequence more
severe than a single failed program.
Another type of pointer error occurs when a pointer value is pointing to an object of the
wrong type or to a segment of memory that does not even hold a valid object at all. These
types of errors are impossible in Java, which does not allow programmers to manipulate pointers
directly. In other languages, it is possible to set a pointer to point, essentially, to any location
in memory. If this is done incorrectly, then using the pointer can have unpredictable results.
Another type of error that cannot occur in Java is a memory leak. In Java, once there are
no longer any pointers that refer to an object, that object is “garbage collected” so that the
memory that it occupied can be reused. In other languages, it is the programmer’s responsibility
to return unused memory to the system. If the programmer fails to do this, unused memory
can build up, leaving less memory for programs and data. There is a story that many common
programs for older Windows computers had so many memory leaks that the computer would
run out of memory after a few days of use and would have to be restarted.
Many programs have been found to suffer from buffer overflow errors. Buffer overflow
errors often make the news because they are responsible for many network security problems.
When one computer receives data from another computer over a network, that data is stored in
a buffer. The buffer is just a segment of memory that has been allocated by a program to hold
data that it expects to receive. A buffer overflow occurs when more data is received than will
fit in the buffer. The question is, what happens then? If the error is detected by the program
or by the networking software, then the only thing that has happened is a failed network
data transmission. The real problem occurs when the software does not properly detect buffer
overflows. In that case, the software continues to store data in memory even after the buffer is
filled, and the extra data goes into some part of memory that was not allocated by the program
as part of the buffer. That memory might be in use for some other purpose. It might contain
important data. It might even contain part of the program itself. This is where the real security
issues come in. Suppose that a buffer overflow causes part of a program to be replaced with
extra data received over a network. When the computer goes to execute the part of the program
that was replaced, it’s actually executing data that was received from another computer. That
data could be anything. It could be a program that crashes the computer or takes it over. A
malicious programmer who finds a convenient buffer overflow error in networking software can
CHAPTER 8. CORRECTNESS, ROBUSTNESS, EFFICIENCY 389
try to exploit that error to trick other computers into executing his programs.
For software written completely in Java, buffer overflow errors are impossible. The language
simply does not provide any way to store data into memory that has not been properly allocated.
To do that, you would need a pointer that points to unallocated memory or you would have
to refer to an array location that lies outside the range allocated for the array. As explained
above, neither of these is possible in Java. (However, there could conceivably still be errors in
Java’s standard classes, since some of the methods in these classes are actually written in the
C programming language rather than in Java.)
It’s clear that language design can help prevent errors or detect them when they occur.
Doing so involves restricting what a programmer is allowed to do. Or it requires tests, such as
checking whether a pointer is null, that take some extra processing time. Some programmers
feel that the sacrifice of power and efficiency is too high a price to pay for the extra security. In
some applications, this is true. However, there are many situations where safety and security
are primary considerations. Java is designed for such situations.
The quadratic formula (from high-school mathematics) assures us that the value assigned to x
is a solution of the equation A*x2 + B*x + C = 0, provided that the value of disc is greater
than or equal to zero and the value of A is not zero. If we can guarantee that B*B-4*A*C >= 0
and that A != 0, then the fact that x is a solution of the equation becomes a postcondition
of the program segment. We say that the condition, B*B-4*A*C >= 0 is a precondition of
the program segment. The condition that A != 0 is another precondition. A precondition is
defined to be a condition that must be true at a given point in the execution of a program in
order for the program to continue correctly. A precondition is something that you want to be
true. It’s something that you have to check or force to be true, if you want your program to be
correct.
We’ve encountered preconditions and postconditions once before, in Subsection 4.7.1. That
section introduced preconditions and postconditions as a way of specifying the contract of
a subroutine. As the terms are being used here, a precondition of a subroutine is just a
precondition of the code that makes up the definition of the subroutine, and the postcondition
of a subroutine is a postcondition of the same code. In this section, we have generalized these
terms to make them more useful in talking about program correctness in general.
Let’s see how this works by considering a longer program segment:
do {
System.out.println("Enter A, B, and C.");
System.out.println("A must be non-zero and B*B-4*A*C must be >= 0.");
System.out.print("A = ");
A = TextIO.getlnDouble();
System.out.print("B = ");
B = TextIO.getlnDouble();
System.out.print("C = ");
C = TextIO.getlnDouble();
if (A == 0 || B*B - 4*A*C < 0)
System.out.println("Your input is illegal. Try again.");
} while (A == 0 || B*B - 4*A*C < 0);
disc = B*B - 4*A*C;
x = (-B + Math.sqrt(disc)) / (2*A);
After the loop ends, we can be sure that B*B-4*A*C >= 0 and that A != 0. The preconditions
for the last two lines are fulfilled, so the postcondition that x is a solution of the equation
A*x2 + B*x + C = 0 is also valid. This program segment correctly and provably computes a
solution to the equation. (Actually, because of problems with representing real numbers on
computers, this is not 100% true. The algorithm is correct, but the program is not a perfect
implementation of the algorithm. See the discussion in Subsection 8.1.3.)
Here is another variation, in which the precondition is checked by an if statement. In the
first part of the if statement, where a solution is computed and printed, we know that the
preconditions are fulfilled. In the other parts, we know that one of the preconditions fails to
hold. In any case, the program is correct.
System.out.println("Enter your values for A, B, and C.");
System.out.print("A = ");
A = TextIO.getlnDouble();
System.out.print("B = ");
B = TextIO.getlnDouble();
System.out.print("C = ");
C = TextIO.getlnDouble();
CHAPTER 8. CORRECTNESS, ROBUSTNESS, EFFICIENCY 393
8.2.3 Invariants
Let’s look at how loops work in more detail. Consider a subroutine for finding the sum of the
elements in an array of int:
static int arraySum( int[] A ) {
int total = 0;
int i = 0;
while ( i < A.length ) {
total = total + A[i];
i = i + 1;
}
return total;
}
(Note, by the way, that the requirements that A is not null is a precondition of the subroutine.
If it is violated, the code in the subroutine will throw a NullPointerException.)
How can we be sure that this subroutine works? We need to prove that when the return
statement is executed, the value of total is the sum of all the elements in A. One way to think
about this problem is in terms of loop invariants.
A loop invariant is, roughly, a statement that remains true as the loop is executed. More
precisely, we can show that a statement is an invariant for a loop if the following holds: As
long as the statement is true before the code inside the loop is executed, then it will also
be true after the code inside the loop has been executed. That is, a loop invariant is both a
precondition and a postcondition of the body of the loop.
A loop invariant for the loop in the above subroutine is, “total is equal to the sum of the
first i elements of A.” Suppose this is true at the beginning of the while loop. That is, before
the statement “total = total + A[i]” is executed, total is the sum of the first i elements of
the array (namely A[0] through A[i-1]). After A[i] is added to total, total is now the sum
of the first i+1 elements of the array. At this point, the loop invariant is not true. However, as
soon as the next statement, “i = i + 1” is executed, replacing i with i+1, the loop invariant
becomes true again. We have checked that if the loop invariant is true at the start of the body
of the loop, then is also true at the end.
Note that a loop invariant is not necessarily true at every point during the execution of
a loop. Executing one of the statements in the loop can make it false temporarily, as long as
later statements in the loop make it true again.
So, have we proved that the subroutine arraySum() is correct? Not quite. There are still a
few things to check. First of all, we need to make sure that the loop invariant is true before the
very first time the loop is executed. At that point, i is zero, and total is also equal to zero,
which is the correct sum of zero elements. So the loop invariant is true before the loop. Once
we know that, we know that it remains true after each execution of the loop (because it’s an
invariant), and in particular, we know that it will still be true after the loop ends.
But for that to do us any good, we need to check that the loop actually does end! In each
execution of the loop, the value of i goes up by one. That means that eventually it has to
reach A.length. At that point, the condition in the while loop is false, and the loop ends.
After the loop ends, we know that i equals A.length, and we know that the loop invariant
is true. At that point, since i is A.length, the loop invariant says, “total is the sum of the
first A.length” elements of A.” But that includes all of the elements of A. So, the loop invariant
gives us exactly what we wanted to show: When total is returned by the subroutine, it is equal
to the sum of all the elements of the array!
CHAPTER 8. CORRECTNESS, ROBUSTNESS, EFFICIENCY 395
This might seem like a lot of work to you to prove something that’s obvious. But if you
try to explain why it’s obvious that arraySum() works, you’ll probably find yourself using the
logic behind loop invariants, even if you don’t use the term.
Let’s look more quickly at a similar example. Consider a subroutine that finds the maximum
value in an array of int, where we assume that the array has length at least one:
static int maxInArray( int[] A ) {
int max = A[0];
int i = 1;
while ( i < A.length ) {
if ( A[i] > max )
max = A[i];
i = i + 1;
}
return max;
}
In this case, we have a loop invariant that says, “max is the largest value among the first i
elements of A.” Suppose this is true before the if statement. After the if statement, max is
greater than or equal to A[i], because that is a postcondition of the if statement, and it is
greater than or equal to A[0] through A[i-1], because of the truth of the loop invariant. Put
those two facts together, and you get that max is the largest value among the first i+1 elements
of A. When i is replaced by i+1 in the next statement, the loop invariant becomes true again.
After the loop ends, i is A.length, and the loop invariant tells us exactly what we need to
know: max is the largest value in the whole array.
Loop invariants are not just useful for proving that programs are correct. Thinking in terms
of loop invariants can be useful when you are trying to develop an algorithm. As an example,
let’s look at the insertion sort algorithm that was discussed in Subsection 7.4.3. Suppose that
we want to sort an array A. That is, at the end of the algorithm, we want it to be true that
A[0] <= A[1] <= ... <= A[A.length-1]
The question is, what step-by-step procedure can we use to make this statement true? Well,
can we come up with a loop invariant that, at the end, will become the statement that we want
to be true? If we want all of the elements to be sorted at the end, how about a loop invariant
that says that some of the elements are sorted—say, that the first i elements are sorted. This
leads to an outline for the algorithm:
i = 0;
while (i < A.length ) {
// Loop invariant: A[0] <= A[1] <= ... <= A[i-1]
.
. // Code that adds A[i] to the sorted portion of the array
.
i = i + 1;
}
// At this point, i = A.length, and A[0] <= A[1] <= ... <= A[A.length-1]
The loop invariant is true before the while loop, and when the loop ends, the loop invariant
becomes precisely the statement that we want to be true at the end of the algorithm. We know
what we have to do to complete the algorithm: Develop code for the inside of the loop that
will preserve the truth of the loop invariant. If we can do that, the loop invariant will assure
us that the algorithm that we have developed is correct. The algorithm for adding A[i] to the
CHAPTER 8. CORRECTNESS, ROBUSTNESS, EFFICIENCY 396
sorted portion of the array will require its own loop, with its own loop invariant. I’ll leave you
to think about that.
∗ ∗ ∗
There is another kind of invariant that is useful for thinking about programs: class in-
variants. A class invariant is, roughly, a statement that is always true about the state of a
class, or about objects created from that class. For example, suppose we have a PairOfDice
class in which the values shown on the dice are stored in instance variables die1 and die2.
(See Section 5.2 for a variety of such classes.) We might like to have a class invariant that says,
“the values of die1 and die2 are in the range 1 through 6.” (This would be a statement about
any object created from the PairOfDice class, not about the class as such.) After all, this is a
statement that should always be true about any pair of dice.
But in order to be a class invariant, the statement has to be guaranteed true at all times.
If die1 and die2 are public instance variables, then no such guarantee is possible, since there
is no way to control what values a program that uses the class might assign to them. So,
we are led to make die1 and die2 private. Then we just have to make sure that all of the
code in the class definition respects the class invariant. That is, first of all, when a PairOfDice
object is constructed, the variables die1 and die2 must be initialized to be in the range 1 to 6.
Furthermore, every method in the class must preserve the truth of the class invariant. In this
case, that means that any method that assigns a value to die1 or die2 must ensure that the
value is in the range 1 to 6. For example, a setter method would have to check that a legal
value is being assigned.
In general, we can say that a class invariant is a postcondition of every constructor and is
both a precondition and a postcondition of every method in the class. When you are writing
a class, a class invariant is something that you want to be true at all times. When you write a
method, you need to make sure that the code in that method respects the invariant: Assuming
that the class invariant is true when the method in called, you need to ensure that it will still
be true after the code in the method is executed. This kind of thinking can be a very useful
tool for class design.
As another example, consider a dynamic array class, like the one in Subsection 7.2.4. That
class uses an ordinary array to store values and a counter to keep track of how many items have
been added to the dyanmic array:
private int[] items = new int[8];
private int itemCount = 0;
Class invariants include the facts that “itemCount is the number of items,” that
“0 <= itemCount < items.length,” and that “the items are in the array elements items[0]
through items[itemCount-1].” Keeping these invariants in mind can be helpful when writing
the class. When writing a method for adding an item, the first invariant reminds you to incre-
ment itemCount in order to ensure that the invariant remains true. The second invariant tells
you where the new item has to be stored. And the third invariant tells you that if increment-
ing itemCount makes it equal to items.length, then you will need to do something to avoid
violating the invariant. (Since itemCount has to be incremented, the invariant means that you
will have to make the array bigger.)
In future chapters, I will occasionally point out how it can be useful to think in terms of
preconditions, postconditions, and invariants.
I should note that reasoning about invariants becomes much more complicated in parallel
programs, when several threads that are running at the same time and are accessing the same
data. This will be an issue when we encounter threads in Chapter 12.
CHAPTER 8. CORRECTNESS, ROBUSTNESS, EFFICIENCY 397
More specifically, the user will input lines containing one or more measurements such as “1
foot” or “3 miles 20 yards 2 feet”. The legal units of measure are inch, foot, yard, and mile.
The program will also recognize plurals (inches, feet, yards, miles) and abbreviations (in, ft,
yd, mi). Let’s write a subroutine that will read one line of input of this form and compute
the equivalent number of inches. The main program uses the number of inches to compute the
equivalent number of feet, yards, and miles. If there is any error in the input, the subroutine
will print an error message and return the value -1. The subroutine assumes that the input
line is not empty. The main program tests for this before calling the subroutine and uses an
empty line as a signal for ending the program.
Ignoring the possibility of illegal inputs, a pseudocode algorithm for the subroutine is
inches = 0 // This will be the total number of inches
while there is more input on the line:
read the numerical measurement
read the units of measure
add the measurement to inches
return inches
We can test whether there is more input on the line by checking whether the next non-blank
character is the end-of-line character. But this test has a precondition: We have to make sure
that the next character in the input is in fact either an end-of-line or is a non-blank. To ensure
that, we need to skip over any blank characters. So, the algorithm becomes
inches = 0
skipBlanks()
while TextIO.peek() is not ’\n’:
read the numerical measurement
read the unit of measure
add the measurement to inches
skipBlanks()
return inches
Note the call to skipBlanks() at the end of the while loop. The call to skipBlanks() ensures
that the precondition for the test is again true. More generally, if the test in a while loop has a
precondition, then you have to make sure that this precondition holds at the end of the while
loop, before the computer jumps back to re-evaluate the test, as well as before the start of the
loop.
What about error checking? Before reading the numerical measurement, we have to make
sure that there is really a number there to read. Before reading the unit of measure, we have
to test that there is something there to read. (The number might have been the last thing on
the line. An input such as “3”, without a unit of measure, is not acceptable.) Also, we have to
check that the unit of measure is one of the valid units: inches, feet, yards, or miles. Here is
an algorithm that includes error-checking:
inches = 0
skipBlanks()
while TextIO.peek() is not ’\n’:
if the next character is not a digit:
report an error and return -1
Let measurement = TextIO.getDouble();
skipBlanks() // Precondition for the next test!!
CHAPTER 8. CORRECTNESS, ROBUSTNESS, EFFICIENCY 399
by returning -1. */
while (ch != ’\n’) {
/* Get the next measurement and the units. Before reading
anything, make sure that a legal value is there to read. */
if ( ! Character.isDigit(ch) ) {
System.out.println(
"Error: Expected to find a number, but found " + ch);
return -1;
}
measurement = TextIO.getDouble();
skipBlanks();
if (TextIO.peek() == ’\n’) {
System.out.println(
"Error: Missing unit of measure at end of line.");
return -1;
}
units = TextIO.getWord();
units = units.toLowerCase();
/* Convert the measurement to inches and add it to the total. */
if (units.equals("inch")
|| units.equals("inches") || units.equals("in")) {
inches += measurement;
}
else if (units.equals("foot")
|| units.equals("feet") || units.equals("ft")) {
inches += measurement * 12;
}
else if (units.equals("yard")
|| units.equals("yards") || units.equals("yd")) {
inches += measurement * 36;
}
else if (units.equals("mile")
|| units.equals("miles") || units.equals("mi")) {
inches += measurement * 12 * 5280;
}
else {
System.out.println("Error: \"" + units
+ "\" is not a legal unit of measure.");
return -1;
}
/* Look ahead to see whether the next thing on the line is
the end-of-line. */
skipBlanks();
ch = TextIO.peek();
} // end while
return inches;
} // end readMeasurement()
CHAPTER 8. CORRECTNESS, ROBUSTNESS, EFFICIENCY 401
The source code for the complete program can be found in the file LengthConverter2.java.
caught by the program. The interpreter responds by terminating the program. In many other
programming languages, a crashed program will sometimes crash the entire system and freeze
the computer until it is restarted. With Java, such system crashes should be impossible—which
means that when they happen, you have the satisfaction of blaming the system rather than
your own program.
Exceptions were introduced in Section 3.7, along with the try..catch statement, which is
used to catch and handle exceptions. However, that section did not cover the complete syntax
of try..catch or the full complexity of exceptions. In this section, we cover these topics in full
detail.
∗ ∗ ∗
When an exception occurs, the thing that is actually “thrown” is an object. This object
can carry information (in its instance variables) from the point where the exception occurs to
the point where it is caught and handled. This information always includes the subroutine
call stack , which is a list of the subroutines that were being executed when the exception was
thrown. (Since one subroutine can call another, several subroutines can be active at the same
time.) Typically, an exception object also includes an error message describing what happened
to cause the exception, and it can contain other data as well. All exception objects must belong
to a subclass of the standard class java.lang.Throwable. In general, each different type of
exception is represented by its own subclass of Throwable, and these subclasses are arranged in
a fairly complex class hierarchy that shows the relationship among various types of exception.
Throwable has two direct subclasses, Error and Exception. These two subclasses in turn have
many other predefined subclasses. In addition, a programmer can create new exception classes
to represent new types of exception.
Most of the subclasses of the class Error represent serious errors within the Java virtual
machine that should ordinarily cause program termination because there is no reasonable way
to handle them. In general, you should not try to catch and handle such errors. An example is
a ClassFormatError, which occurs when the Java virtual machine finds some kind of illegal data
in a file that is supposed to contain a compiled Java class. If that class was being loaded as
part of the program, then there is really no way for the program to proceed.
On the other hand, subclasses of the class Exception represent exceptions that are meant
to be caught. In many cases, these are exceptions that might naturally be called “errors,” but
they are errors in the program or in input data that a programmer can anticipate and possibly
respond to in some reasonable way. (However, you should avoid the temptation of saying, “Well,
I’ll just put a thing here to catch all the errors that might occur, so my program won’t crash.”
If you don’t have a reasonable way to respond to the error, it’s best just to let the program
crash, because trying to go on will probably only lead to worse things down the road—in the
worst case, a program that gives an incorrect answer without giving you any indication that
the answer might be wrong!)
The class Exception has its own subclass, RuntimeException. This class groups together
many common exceptions, including all those that have been covered in previous sections. For
example, IllegalArgumentException and NullPointerException are subclasses of RuntimeException.
A RuntimeException generally indicates a bug in the program, which the programmer should
fix. RuntimeExceptions and Errors share the property that a program can simply ignore the
possibility that they might occur. (“Ignoring” here means that you are content to let your
program crash if the exception occurs.) For example, a program does this every time it uses
an array reference like A[i] without making arrangements to catch a possible ArrayIndexOut-
OfBoundsException. For all other exception classes besides Error, RuntimeException, and their
CHAPTER 8. CORRECTNESS, ROBUSTNESS, EFFICIENCY 403
Throwable
Error Exception
EOFException SocketException
ArrayIndexOutOfBoundsException
IllegalArgumentException
The class Throwable and
some of its subclasses.
NumberFormatException
The class Throwable includes several instance methods that can be used with any exception
object. If e is of type Throwable (or one of its subclasses), then e.getMessage() is a function
that returns a String that describes the exception. The function e.toString(), which is used
by the system whenever it needs a string representation of the object, returns a String that
contains the name of the class to which the exception belongs as well as the same string that
would be returned by e.getMessage(). And the method e.printStackTrace() writes a stack
trace to standard output that tells which subroutines were active when the exception occurred.
A stack trace can be very useful when you are trying to determine the cause of the problem.
(Note that if an exception is not caught by the program, then the default response to the
exception prints the stack trace to standard output.)
try {
double determinant = M[0][0]*M[1][1] - M[0][1]*M[1][0];
System.out.println("The determinant of M is " + determinant);
}
catch ( NullPointerException | ArrayIndexOutOfBoundsException err ) {
System.out.println("Sorry, an error has occurred.");
System.out.println("The error was: " + err);
}
Here, the two exception types are combined with a ”|”, the vertical line character that is also
used in the boolean or operator. This example will catch errors of type NullPointerException or
ArrayIndexOutOfBoundsException, and no other types.
The example I’ve been using here is not realistic, because you are not very likely to use
exception-handling to guard against null pointers and bad array indices. This is a case where
careful programming is better than exception handling: Just be sure that your program assigns
a reasonable, non-null value to the array M. You would certainly resent it if the designers of
Java forced you to set up a try..catch statement every time you wanted to use an array!
This is why handling of potential RuntimeExceptions is not mandatory. There are just too
many things that might go wrong! (This also shows that exception-handling does not solve
the problem of program robustness. It just gives you a tool that will in many cases let you
approach the problem in a more organized way.)
∗ ∗ ∗
I have still not completely specified the syntax of the try statement. The next variation is
the possibility of a finally clause at the end of a try statement. With this addition, syntax
of the try statement can be described as:
try {
hstatements i
}
hoptional-catch-clauses i
hoptional-finally-clause i
Note that the catch clauses are also listed as optional. The try statement can include zero or
more catch clauses and, optionally, a finally clause. The try statement must include one
or the other. That is, a try statement can have either a finally clause, or one or more catch
clauses, or both. The syntax for a catch clause is
catch ( hexception-class-names i hvariable-name i ) {
hstatements i
}
where hexception-class-namesi can be a single exception class or several classes separated by ”|”.
The syntax for a finally clause is
finally {
hstatements i
}
The semantics of the finally clause is that the block of statements in the finally clause is
guaranteed to be executed as the last step in the execution of the try statement, whether or not
any exception occurs and whether or not any exception that does occur is caught and handled.
The finally clause is meant for doing essential cleanup that under no circumstances should
be omitted. One example of this type of cleanup is closing a network connection. Although
CHAPTER 8. CORRECTNESS, ROBUSTNESS, EFFICIENCY 406
you don’t yet know enough about networking to look at the actual programming in this case,
we can consider some pseudocode:
try {
open a network connection
communicate over the connection
}
catch ( IOException e ) {
report the error
}
finally {
if the connection was successfully opened
close the connection
}
The finally clause ensures that the network connection will definitely be closed, whether or not
an error occurs during the communication. The pseudocode in this example follows a general
pattern that can be used to robustly obtain a resource, use the resource, and then release the
resource.
∗ ∗ ∗
The pattern of obtaining a resource, then using the resource, and then releasing the resource
is very common. Note that the resource can only be released if no error occurred while obtaining
it. And, if it was successfully obtained, then it should be closed whether or not an error occurs
while using it. This pattern is so common that it leads to one last option in the try statement
syntax. With this option, you only need code to obtain the resource, and you don’t need to
worry about releasing it. That will happen automatically at the end of the try statement.
In order for this to work, the resource must be represented by an object that implements an
interface named AutoCloseable, which defines a single method named close(), with no param-
eters. Standard Java classes that represent things like files and network connections already
implement AutoClosable. So does the Scanner class, which was introduced in Subsection 2.4.6.
In that section, I showed how to use a Scanner to read from System.in. Although I didn’t do it
in that section, it’s considered good form to close a Scanner after using it. Here is an example
that uses the pattern in a try statement to make sure that the Scanner is closed automatically:
try( Scanner in = new Scanner(System.in) ) {
// Use the Scanner to read from standard input
}
catch (Exception e) {
// ... some error occurred while using the Scanner
}
The statement that allocates the Scanner goes in parentheses after the word “try”. The state-
ment must have the form of a variable declaration that includes an initialization of the variable.
The variable is local to the try statement. (You can actually declare several variables in the
parentheses, separated by semicolons.) In this example, we can be sure that in.close() will
definitely be called by the system at the end of the try statement, as long as the Scanner was
successfully initialized.
This is all getting quite complicated, and I won’t continue the discussion here. The sample
program TryStatementDemo.java demonstrates a try statement with all its options, and it
includes a lot of comments to help you understand what can happen when you run the program.
CHAPTER 8. CORRECTNESS, ROBUSTNESS, EFFICIENCY 407
*/
static public double root( double A, double B, double C )
throws IllegalArgumentException {
if (A == 0) {
throw new IllegalArgumentException("A can’t be zero.");
}
else {
double disc = B*B - 4*A*C;
if (disc < 0)
throw new IllegalArgumentException("Discriminant < zero.");
return (-B + Math.sqrt(disc)) / (2*A);
}
}
As discussed in the previous section, the computation in this subroutine has the precon-
ditions that A != 0 and B*B-4*A*C >= 0. The subroutine throws an exception of type Ille-
galArgumentException when either of these preconditions is violated. When an illegal condition
is found in a subroutine, throwing an exception is often a reasonable response. If the program
that called the subroutine knows some good way to handle the error, it can catch the exception.
If not, the program will crash—and the programmer will know that the program needs to be
fixed.
A throws clause in a subroutine heading can declare several different types of exception,
separated by commas. For example:
void processArray(int[] A) throws NullPointerException,
ArrayIndexOutOfBoundsException { ...
executed. The caller will, in turn, be forced either to handle the exception in a try statement
or to declare the exception in a throws clause in its own header.
Exception-handling is mandatory for any exception class that is not a subclass of either
Error or RuntimeException. These checked exceptions generally represent conditions that are
outside the control of the programmer. For example, they might represent bad input or an
illegal action taken by the user. There is no way to avoid such errors, so a robust program
has to be prepared to handle them. The design of Java makes it impossible for programmers
to ignore the possibility of such errors.
Among the checked exceptions are several that can occur when using Java’s input/output
routines. This means that you can’t even use these routines unless you understand something
about exception-handling. Chapter 11 deals with input/output and uses checked exceptions
extensively.
A throw statement can be used in a program to throw an error of type ParseError. The
constructor for the ParseError object must specify an error message. For example:
throw new ParseError("Encountered an illegal negative number.");
or
throw new ParseError("The word ’" + word
+ "’ is not a valid file name.");
Since ParseError is defined as a subclass of Exception, it is a checked exception. If the throw
statement does not occur in a try statement that catches the error, then the subroutine that
contains the throw must declare that it can throw a ParseError by adding the clause “throws
ParseError” to the subroutine heading. For example,
void getUserData() throws ParseError {
. . .
}
This would not be required if ParseError were defined as a subclass of RuntimeException instead
of Exception, since in that case ParseErrors would not be checked exceptions.
A routine that wants to handle ParseErrors can use a try statement with a catch clause
that catches ParseErrors. For example:
try {
getUserData();
processUserData();
}
catch (ParseError pe) {
. . . // Handle the error
}
Note that since ParseError is a subclass of Exception, a catch clause of the form “catch
(Exception e)” would also catch ParseErrors, along with any other object of type Exception.
Sometimes, it’s useful to store extra data in an exception object. For example,
class ShipDestroyed extends RuntimeException {
Ship ship; // Which ship was destroyed.
int where x, where y; // Location where ship was destroyed.
ShipDestroyed(String message, Ship s, int x, int y) {
// Constructor creates a ShipDestroyed object
// carrying an error message plus the information
// that the ship s was destroyed at location (x,y)
// on the screen.
super(message);
ship = s;
where x = x;
where y = y;
}
}
Here, a ShipDestroyed object contains an error message and some information about a ship that
was destroyed. This could be used, for example, in a statement:
if ( userShip.isHit() )
throw new ShipDestroyed("You’ve been hit!", userShip, xPos, yPos);
CHAPTER 8. CORRECTNESS, ROBUSTNESS, EFFICIENCY 411
Note that the condition represented by a ShipDestroyed object might not even be considered
an error. It could be just an expected interruption to the normal flow of a game. Exceptions
can sometimes be used to handle such interruptions neatly.
∗ ∗ ∗
The ability to throw exceptions is particularly useful in writing general-purpose methods
and classes that are meant to be used in more than one program. In this case, the person writing
the method or class often has no reasonable way of handling the error, since that person has no
way of knowing exactly how the method or class will be used. In such circumstances, a novice
programmer is often tempted to print an error message and forge ahead, but this is almost
never satisfactory since it can lead to unpredictable results down the line. Printing an error
message and terminating the program is almost as bad, since it gives the program no chance
to handle the error.
The program that calls the method or uses the class needs to know that the error has
occurred. In languages that do not support exceptions, the only alternative is to return some
special value or to set the value of some global variable to indicate that an error has occurred.
For example, the readMeasurement() function in Subsection 8.2.2 returns the value -1 if the
user’s input is illegal. However, this only does any good if the main program bothers to test
the return value. It is very easy to be lazy about checking for special return values every time
a subroutine is called. And in this case, using -1 as a signal that an error has occurred makes
it impossible to allow negative measurements. Exceptions are a cleaner way for a subroutine to
react when it encounters an error.
It is easy to modify the readMeasurement() function to use exceptions instead of a special
return value to signal an error. My modified subroutine throws a ParseError when the user’s
input is illegal, where ParseError is the subclass of Exception that was defined above. (Arguably,
it might be reasonable to avoid defining a new class by using the standard exception class
IllegalArgumentException instead.) The changes from the original version are shown in italic:
/**
* Reads the user’s input measurement from one line of input.
* Precondition: The input line is not empty.
* Postcondition: If the user’s input is legal, the measurement
* is converted to inches and returned.
* @throws ParseError if the user’s input is not legal.
*/
static double readMeasurement() throws ParseError {
double inches; // Total number of inches in user’s measurement.
double measurement; // One measurement,
// such as the 12 in "12 miles."
String units; // The units specified for the measurement,
// such as "miles."
char ch; // Used to peek at next character in the user’s input.
inches = 0; // No inches have yet been read.
skipBlanks();
ch = TextIO.peek();
/* As long as there is more input on the line, read a measurement and
add the equivalent number of inches to the variable, inches. If an
error is detected during the loop, end the subroutine immediately
CHAPTER 8. CORRECTNESS, ROBUSTNESS, EFFICIENCY 412
by throwing a ParseError. */
while (ch != ’\n’) {
/* Get the next measurement and the units. Before reading
anything, make sure that a legal value is there to read. */
if ( ! Character.isDigit(ch) ) {
throw new ParseError("Expected to find a number, but found " + ch);
}
measurement = TextIO.getDouble();
skipBlanks();
if (TextIO.peek() == ’\n’) {
throw new ParseError("Missing unit of measure at end of line.");
}
units = TextIO.getWord();
units = units.toLowerCase();
/* Convert the measurement to inches and add it to the total. */
if (units.equals("inch")
|| units.equals("inches") || units.equals("in")) {
inches += measurement;
}
else if (units.equals("foot")
|| units.equals("feet") || units.equals("ft")) {
inches += measurement * 12;
}
else if (units.equals("yard")
|| units.equals("yards") || units.equals("yd")) {
inches += measurement * 36;
}
else if (units.equals("mile")
|| units.equals("miles") || units.equals("mi")) {
inches += measurement * 12 * 5280;
}
else {
throw new ParseError("\"" + units
+ "\" is not a legal unit of measure.");
}
/* Look ahead to see whether the next thing on the line is
the end-of-line. */
skipBlanks();
ch = TextIO.peek();
} // end while
return inches;
} // end readMeasurement()
In the main program, this subroutine is called in a try statement of the form
CHAPTER 8. CORRECTNESS, ROBUSTNESS, EFFICIENCY 413
try {
inches = readMeasurement();
}
catch (ParseError e) {
. . . // Handle the error.
}
The complete program can be found in the file LengthConverter3.java. From the user’s
point of view, this program has exactly the same behavior as the program LengthConverter2
from the previous section. Internally, however, the programs are significantly different, since
LengthConverter3 uses exception handling.
8.4.1 Assertions
Recall that a precondition is a condition that must be true at a certain point in a program, for
the execution of the program to continue correctly from that point. In the case where there is
a chance that the precondition might not be satisfied—for example, if it depends on input from
the user—then it’s a good idea to insert an if statement to test it. But then the question arises,
What should be done if the precondition does not hold? One option is to throw an exception.
This will terminate the program, unless the exception is caught and handled elsewhere in the
program.
In many cases, of course, instead of using an if statement to test whether a precondition
holds, a programmer tries to write the program in a way that will guarantee that the precondi-
tion holds. In that case, the test should not be necessary, and the if statement can be avoided.
The problem is that programmers are not perfect. In spite of the programmer’s intention, the
program might contain a bug that screws up the precondition. So maybe it’s a good idea to
check the precondition after all—at least during the debugging phase of program development.
Similarly, a postcondition is a condition that is true at a certain point in the program as
a consequence of the code that has been executed before that point. Assuming that the code
is correctly written, a postcondition is guaranteed to be true, but here again testing whether a
desired postcondition is actually true is a way of checking for a bug that might have screwed
up the postcondition. This is something that might be desirable during debugging.
And the same thing applies to loop invariants and class invariants. These are things that
should be true at certain points in a program. If they are not true at those points, it means
that the program contains a bug.
The programming languages C and C++ have always had a facility for adding what are
called assertions to a program. These assertions take the form “assert(hconditioni)”, where
hconditioni is a boolean-valued expression. This condition expresses a precondition or post-
condition that should hold at that point in the program. When the computer encounters an
assertion during the execution of the program, it evaluates the condition. If the condition is
false, the program is terminated. Otherwise, the program continues normally. This allows the
CHAPTER 8. CORRECTNESS, ROBUSTNESS, EFFICIENCY 414
programmer’s belief that the condition is true to be tested; if it is not true, that indicates that
the part of the program that preceded the assertion contained a bug. One nice thing about
assertions in C and C++ is that they can be “turned off” at compile time. That is, if the
program is compiled in one way, then the assertions are included in the compiled code. If the
program is compiled in another way, the assertions are not included. During debugging, the
first type of compilation is used, with assertions turned on. The release version of the program
is compiled with assertions turned off. The release version will be more efficient, because the
computer won’t have to evaluate all the assertions.
Although early versions of Java did not have assertions, an assertion facility similar to the
one in C/C++ has been available in Java since version 1.4. As with the C/C++ version, Java
assertions can be turned on during debugging and turned off during normal execution. In Java,
however, assertions are turned on and off at run time rather than at compile time. An assertion
in the Java source code is always included in the compiled class file. When the program is
run in the normal way, these assertions are ignored; since the condition in the assertion is not
evaluated in this case, there is little or no performance penalty for having the assertions in
the program. When the program is being debugged, it can be run with assertions enabled, as
discussed below, and then the assertions can be a great help in locating and identifying bugs.
∗ ∗ ∗
An assertion statement in Java takes one of the following two forms:
assert hcondition i ;
or
assert hcondition i : herror-message i ;
where hconditioni is a boolean-valued expression and herror-messagei is a string or an expres-
sion of type String. The word “assert” is a reserved word in Java, which cannot be used as an
identifier. An assertion statement can be used anyplace in Java where a statement is legal.
If a program is run with assertions disabled, an assertion statement is equivalent to an
empty statement and has no effect. When assertions are enabled and an assertion statement is
encountered in the program, the hconditioni in the assertion is evaluated. If the value is true,
the program proceeds normally. If the value of the condition is false, then an exception of
type java.lang.AssertionError is thrown, and the program will crash (unless the error is
caught by a try statement). If the assert statement includes an herror-messagei, then the
error message string becomes the message in the AssertionError.
So, the statement “assert hcondition i : herror-message i;" is similar to
if ( hcondition i == false )
throw new AssertionError( herror-message i );
except that the if statement is executed whenever the program is run, and the assert state-
ment is executed only when the program is run with assertions enabled.
The question is, when to use assertions instead of exceptions? The general rule is to use
assertions to test conditions that should definitely be true, if the program is written correctly.
Assertions are useful for testing a program to see whether or not it is correct and for finding
the errors in an incorrect program. After testing and debugging, when the program is used in
the normal way, the assertions in the program will be ignored. However, if a problem turns
up later, the assertions are still there in the program to be used to help locate the error. If
someone writes to you to say that your program doesn’t work when he does such-and-such, you
CHAPTER 8. CORRECTNESS, ROBUSTNESS, EFFICIENCY 415
can run the program with assertions enabled, do such-and-such, and hope that the assertions
in the program will help you locate the point in the program where it goes wrong.
Consider, for example, the root() method from Subsection 8.3.3 that calculates a root of
a quadratic equation. If you believe that your program will always call this method with legal
arguments, then it would make sense to write the method using assertions instead of exceptions:
/**
* Returns the larger of the two roots of the quadratic equation
* A*x*x + B*x + C = 0.
* Precondition: A != 0 and B*B - 4*A*C >= 0.
*/
static public double root( double A, double B, double C ) {
assert A != 0 : "Leading coefficient of quadratic equation cannot be zero.";
double disc = B*B - 4*A*C;
assert disc >= 0 : "Discriminant of quadratic equation cannot be negative.";
return (-B + Math.sqrt(disc)) / (2*A);
}
The assertions are not checked when the program is run in the normal way. If you are correct in
your belief that the method is never called with illegal arguments, then checking the conditions
in the assertions would be unnecessary. If your belief is not correct, the problem should turn
up during testing or debugging, when the program is run with the assertions enabled.
If the root() method is part of a software library that you expect other people to use, then
the situation is less clear. Oracle’s Java documentation advises that assertions should not be
used for checking the contract of public methods: If the caller of a method violates the contract
by passing illegal parameters, then an exception should be thrown. This will enforce the contract
whether or not assertions are enabled. (However, while it’s true that Java programmers expect
the contract of a method to be enforced with exceptions, there are reasonable arguments for
using assertions instead, in some cases.) One might say that assertions are for you, to help you
in debugging your code, while exceptions are for people who use your code, to alert them that
they are misusing it.
On the other hand, it never hurts to use an assertion to check a postcondition or an invariant.
These are conditions that are definitely expected to be true in any bug-free program, so an
assertion is the natural way to check the condition while debugging, without imposing an
efficiency penalty when the program is executed normally. If the postcondition or invariant
is false, there is a bug, and that is something that needs to be found during the testing and
debugging phase of programming.
∗ ∗ ∗
To have any effect, assertions must be enabled when the program is run. How to do this
depends on what programming environment you are using. (See Section 2.6 for a discussion of
programming environments.) In the usual command line environment, assertions are enabled by
adding the option -enableassertions to the java command that is used to run the program.
For example, if the class that contains the main program is RootFinder, then the command
java -enableassertions RootFinder
will run the program with assertions enabled. The -enableassertions option can be abbre-
viated to -ea, so the command can alternatively be written as
java -ea RootFinder
CHAPTER 8. CORRECTNESS, ROBUSTNESS, EFFICIENCY 416
In fact, it is possible to enable assertions in just part of a program. An option of the form
“-ea:hclass-name i” enables only the assertions in the specified class. Note that there are no
spaces between the -ea, the “:”, and the name of the class. To enable all the assertions in a
package and in its sub-packages, you can use an option of the form “-ea:hpackage-name i...”.
To enable assertions in the “default package” (that is, classes that are not specified to belong to
a package, like almost all the classes in this book), use “-ea:...”. For example, to run a Java
program named “MegaPaint” with assertions enabled for every class in the packages named
“paintutils” and “drawing”, you would use the command:
java -ea:paintutils... -ea:drawing... MegaPaint
If you are using the Eclipse integrated development environment, you can specify the -ea
option by creating a run configuration. Right-click the name of the main program class in the
Package Explorer pane, and select “Run As” from the pop-up menu and then “Run. . . ” from
the submenu. This will open a dialog box where you can manage run configurations. The name
of the project and of the main class will be already be filled in. Click the “Arguments” tab, and
enter -ea in the box under “VM Arguments”. The contents of this box are added to the java
command that is used to run the program. You can enter other options in this box, including
more complicated enableassertions options such as -ea:paintutils.... When you click the
“Run” button, the options will be applied. Furthermore, they will be applied whenever you run
the program, unless you change the run configuration or add a new configuration. Note that it
is possible to make two run configurations for the same class, one with assertions enabled and
one with assertions disabled.
8.4.2 Annotations
The term “annotation” commonly refers to notes added to or written alongside a main text, to
help you understand or appreciate the text. An annotation might be a note that you make to
yourself in the margin of a book. It might be a footnote added to an old novel by an editor to
explain the historical context of some event. The annotation is metadata or “metatext,” that
is, text written about the main text rather than as part of the main text itself.
Comments on a program are actually a kind of annotation. Since they are ignored by the
compiler, they have no effect on the meaning of the program. They are there to explain that
meaning to a human reader. It is possible, of course, for another computer program (not the
compiler) to process comments. That’s what is done in the case of Javadoc comments, which
are processed by a program that uses them to create API documentation. But comments are
only one type of metadata that might be added to programs.
In Java 5.0, a new feature called annotations was added to the Java language to make
it easier to create new kinds of metadata for Java programs. This has made it possible for
programmers to devise new ways of annotating programs, and to write programs that can read
and use their annotations.
Java annotations have no direct effect on the program that they annotate. But they do have
many potential uses. Some annotations are used to make the programmer’s intent more explicit.
Such annotations might be checked by a compiler to make sure that the code is consistent with
the programmer’s intention. For example, @Override is a standard annotation that can be
used to annotate method definitions. It means that the method is intended to override (that
is replace) a method with the same signature that was defined in some superclass. A compiler
can check that the superclass method actually exists; if not, it can inform the programmer. An
CHAPTER 8. CORRECTNESS, ROBUSTNESS, EFFICIENCY 417
annotation used in this way is an aid to writing correct programs, since the programmer can
be warned about a potential error in advance, instead of having to hunt it down later as a bug.
To annotate a method definition with the @Override annotation, simply place it in front of
the definition. Syntactically, annotations are modifiers that are used in much the same way as
built-in modifiers like “public” and “final.” For example,
@Override public void WindowClosed(WindowEvent evt) { ... }
If there is no "WindowClosed(WindowEvent)" method in any superclass, then the compiler
can issue an error. In fact, this example is based on a hard-to-find bug that I once intro-
duced when trying to override a method named “windowClosed” with a method that I called
“WindowClosed” (with an upper case “W”). If the @Override annotation had existed at that
time—and if I had used it—the compiler could have rejected my code and saved me the trouble
of tracking down the bug.
(Annotations are a fairly advanced feature, and I might not have mentioned them in this
textbook, except that the @Override annotation can show up in code generated by Eclipse and
other integrated development environments.)
There are two other standard annotations. One is @Deprecated, which can be used to mark
deprecated classes, methods, and variables. (A deprecated item is one that is considered to be
obsolete, but is still part of the Java language for backwards compatibility for old code.) Use of
this annotation would allow a compiler to generate warnings when the deprecated item is used.
The other standard annotation is @SurpressWarnings, which can be used by a compiler
to turn off warning messages that would ordinarily be generated when a class or method is
compiled. @SuppressWarnings is an example of an annotation that has a parameter. The
parameter tells what class of warnings are to be suppressed. For example, when a class or
method is annotated with
@SuppressWarnings("deprecation")
then no warnings about the use of deprecated items will be emitted when the class or method
is compiled. There are other types of warning that can be suppressed; unfortunately the list of
warnings and their names is not standardized and will vary from one compiler to another.
Note, by the way, that the syntax for annotation parameters—especially for an annotation
that accepts multiple parameters—is not the same as the syntax for method parameters. I
won’t cover the annotation syntax here.
Programmers can define new annotations for use in their code. Such annotations are ignored
by standard compilers and programming tools, but it’s possible to write programs that can
understand the annotations and check for their presence in source code. It is even possible to
create annotations that will be retained at run-time and become part of the running program.
In that case, a program can check for annotations in the actual compiled code that is being
executed, and take actions that depend on the presence of the annotation or the values of its
parameters.
Annotations can help programmers to write correct programs. To use an example from the
Java documentation, they can help with the creation of “boilerplate” code—that is, code that
has a very standardized format and that can be generated mechanically. Often, boilerplate code
is generated based on other code. Doing that by hand is a tedious and error-prone process.
A simple example might be code to save certain aspects of a program’s state to a file and to
restore it later. The code for reading and writing the values of all the relevant state variables is
highly repetitious. Instead of writing that code by hand, a programmer could use an annotation
to mark the variables that are part of the state that is to be saved. A program could then be
CHAPTER 8. CORRECTNESS, ROBUSTNESS, EFFICIENCY 418
used to check for the annotations and generate the save-and-restore code. In fact, it would even
be possible to do without that code altogether, if the program checks for the presence of the
annotation at run time to decide which variables to save and restore.
If this explanation is too mathematical for you, we can just note that for large values of
n, the c in the formula (a+b)*n+c is insignificant compared to the other term, (a+b)*n. We
say that c is a “lower order term.” When doing asymptotic analysis, lower order terms can
be discarded. A rough, but correct, asymptotic analysis of the algorithm would go something
like this: Each iteration of the for loop takes a certain constant amount of time. There are
n iterations of the loop, so the total run time is a constant times n, plus lower order terms
(to account for the initialization). Disregarding lower order terms, we see that the run time is
O(n).
∗ ∗ ∗
Note that to say that an algorithm has run time O(f(n)) is to say that its run time is no
bigger than some constant times f(n) (for large values of n). O(f(n)) puts an upper limit on
the run time. However, the run time could be smaller, even much smaller. For example, if the
run time is O(n), it would also be correct to say that the run time is O(n2 ) or even O(n10 ). If
the run time is less than a constant times n, then it is certainly less than the same constant
times n2 or n10 .
Of course, sometimes it’s useful to have a lower limit on the run time. That is, we want
to be able to say that the run time is greater than or equal to some constant times f(n) (for
large values of n). The notation for this is Ω(f(n)), read “Omega of f of n” or “Big Omega of f
of n.” “Omega” is the name of a letter in the Greek alphabet, and Ω is the upper case version
of that letter. (To be technical, saying that the run time of an algorithm is Ω(f(n)) means that
there is a positive number C and a positive integer M such that whenever n is greater than
M, the run time is greater than or equal to C*f(n).) O(f(n)) tells you something about the
maximum amount of time that you might have to wait for an algorithm to finish; Ω(f(n)) tells
you something about the minimum time.
The algorithm for adding up the numbers in an array has a run time that is Ω(n) as well as
O(n). When an algorithm has a run time that is both Ω(f(n)) and O(f(n)), its run time is said
to be Θ(f(n)), read “Theta of f of n” or “Big Theta of f of n.” (Theta is another letter from
the Greek alphabet.) To say that the run time of an algorithm is Θ(f(n)) means that for large
values of n, the run time is between a*f(n) and b*f(n), where a and b are constants (with b
greater than a, and both greater than 0).
Let’s look at another example. Consider the algorithm that can be expressed in Java in the
following method:
/**
* Sorts the n array elements A[0], A[1], ..., A[n-1] into increasing order.
*/
public static void simpleBubbleSort( int[] A, int n ) {
for (int i = 0; i < n; i++) {
// Do n passes through the array...
for (int j = 0; j < n-1; j++) {
if ( A[j] > A[j+1] ) {
// A[j] and A[j+1] are out of order, so swap them
int temp = A[j];
A[j] = A[j+1];
A[j+1] = temp;
}
}
}
}
CHAPTER 8. CORRECTNESS, ROBUSTNESS, EFFICIENCY 421
Here, the parameter n represents the problem size. The outer for loop in the method is executed
n times. Each time the outer for loop is executed, the inner for loop is executed n-1 times, so
the if statement is executed n*(n-1) times. This is n2 -n, but since lower order terms are not
significant in an asymptotic analysis, it’s good enough to say that the if statement is executed
about n2 times. In particular, the test A[j] > A[j+1] is executed about n2 times, and this fact
by itself is enough to say that the run time of the algorithm is Ω(n2 ), that is, the run time is
at least some constant times n2 . Furthermore, if we look at other operations—the assignment
statements, incrementing i and j, etc.—none of them are executed more than n2 times, so the
run time is also O(n2 ), that is, the run time is no more than some constant times n2 . Since it
is both Ω(n2 ) and O(n2 ), the run time of the simpleBubbleSort algorithm is Θ(n2 ).
You should be aware that some people use the notation O(f(n)) as if it meant Θ(f(n)). That
is, when they say that the run time of an algorithm is O(f(n)), they mean to say that the run
time is about equal to a constant times f(n). For that, they should use Θ(f(n)). Properly
speaking, O(f(n)) means that the run time is less than a constant times f(n), possibly much
less.
∗ ∗ ∗
So far, my analysis has ignored an important detail. We have looked at how run time
depends on the problem size, but in fact the run time usually depends not just on the size of
the problem but on the specific data that has to be processed. For example, the run time of a
sorting algorithm can depend on the initial order of the items that are to be sorted, and not
just on the number of items.
To account for this dependency, we can consider either the worst case run time analysis
or the average case run time analysis of an algorithm. For a worst case run time analysis, we
consider all possible problems of size n and look at the longest possible run time for all such
problems. For an average case analysis, we consider all possible problems of size n and look at
the average of the run times for all such problems. Usually, the average case analysis assumes
that all problems of size n are equally likely to be encountered, although this is not always
realistic—or even possible in the case where there is an infinite number of different problems of
a given size.
In many cases, the average and the worst case run times are the same to within a constant
multiple. This means that as far as asymptotic analysis is concerned, they are the same. That
is, if the average case run time is O(f(n)) or Θ(f(n)), then so is the worst case. However, later in
the book, we will encounter a few cases where the average and worst case asymptotic analyses
differ.
It is also possible to talk about best case run time analysis, which looks at the shortest
possible run time for all inputs of a given size. However, a best case analysis is only occasionally
useful.
∗ ∗ ∗
So, what do you really have to know about analysis of algorithms to read the rest of this
book? We will not do any rigorous mathematical analysis, but you should be able to follow
informal discussion of simple cases such as the examples that we have looked at in this section.
Most important, though, you should have a feeling for exactly what it means to say that the
running time of an algorithm is O(f(n)) or Θ(f(n)) for some common functions f(n). The main
point is that these notations do not tell you anything about the actual numerical value of the
running time of the algorithm for any particular case. They do not tell you anything at all
about the running time for small values of n. What they do tell you is something about the
rate of growth of the running time as the size of the problem increases.
CHAPTER 8. CORRECTNESS, ROBUSTNESS, EFFICIENCY 422
Suppose you compare two algorithms that solve the same problem. The run time of one
algorithm is Θ(n2 ), while the run time of the second algorithm is Θ(n3 ). What does this tell
you? If you want to know which algorithm will be faster for some particular problem of size,
say, 100, nothing is certain. As far as you can tell just from the asymptotic analysis, either
algorithm could be faster for that particular case—or in any particular case. But what you can
say for sure is that if you look at larger and larger problems, you will come to a point where the
Θ(n2 ) algorithm is faster than the Θ(n3 ) algorithm. Furthermore, as you continue to increase
the problem size, the relative advantage of the Θ(n2 ) algorithm will continue to grow. There
will be values of n for which the Θ(n2 ) algorithm is a thousand times faster, a million times
faster, a billion times faster, and so on. This is because for any positive constants a and b, the
function a*n3 grows faster than the function b*n2 as n gets larger. (Mathematically, the limit
of the ratio of a*n3 to b*n2 is infinite as n approaches infinity.)
This means that for “large” problems, a Θ(n2 ) algorithm will definitely be faster than a
Θ(n3 ) algorithm. You just don’t know—based on the asymptotic analysis alone—exactly how
large “large” has to be. In practice, in fact, it is likely that the Θ(n2 ) algorithm will be faster
even for fairly small values of n, and absent other information you would generally prefer a
Θ(n2 ) algorithm to a Θ(n3 ) algorithm.
So, to understand and apply asymptotic analysis, it is essential to have some idea of the
rates of growth of some common functions. For the power functions n, n2 , n3 , n4 , . . . , the larger
the exponent, the greater the rate of growth of the function. Exponential functions such as 2n
and 10n , where the n is in the exponent, have a growth rate that is faster than that of any power
function. In fact, exponential functions grow so quickly that an algorithm whose run time grows
exponentially is almost certainly impractical even for relatively modest values of n, because the
running time is just too long. Another function that often turns up in asymptotic analysis is
the logarithm function, log(n). There are actually many different logarithm functions, but the
one that is usually used in computer science is the so-called logarithm to the base two, which is
defined by the fact that log(2x ) = x for any number x. (Usually, this function is written log2 (n),
but I will leave out the subscript 2, since I will only use the base-two logarithm in this book.)
The logarithm function grows very slowly. The growth rate of log(n) is much smaller than the
growth rate of n. The growth rate of n*log(n) is a little larger than the growth rate of n, but
much smaller than the growth rate of n2 . The following table should help you understand the
differences among the rates of growth of various functions:
The reason that log(n) shows up so often is because of its association with multiplying and
dividing by two: Suppose you start with the number n and divide it by 2, then divide by 2
again, and so on, until you get a number that is less than or equal to 1. Then the number of
divisions is equal (to the nearest integer) to log(n).
As an example, consider the binary search algorithm from Subsection 7.4.1. This algorithm
searches for an item in a sorted array. The problem size, n, can be taken to be the length of
the array. Each step in the binary search algorithm divides the number of items still under
CHAPTER 8. CORRECTNESS, ROBUSTNESS, EFFICIENCY 423
consideration by 2, and the algorithm stops when the number of items under consideration is
less than or equal to 1 (or sooner). It follows that the number of steps for an array of length n
is at most log(n). This means that the worst-case run time for binary search is Θ(log(n)). (The
average case run time is also Θ(log(n)).) By comparison, the linear search algorithm, which
was also presented in Subsection 7.4.1 has a run time that is Θ(n). The Θ notation gives us
a quantitative way to express and to understand the fact that binary search is “much faster”
than linear search.
In binary search, each step of the algorithm divides the problem size by 2. It often happens
that some operation in an algorithm (not necessarily a single step) divides the problem size
by 2. Whenever that happens, the logarithm function is likely to show up in an asymptotic
analysis of the run time of the algorithm.
Analysis of Algorithms is a large, fascinating field. We will only use a few of the most basic
ideas from this field, but even those can be very helpful for understanding the differences among
algorithms.
Exercises 424
1. Write a program that uses the following subroutine, from Subsection 8.3.3, to solve equa- (solution)
tions specified by the user.
/**
* Returns the larger of the two roots of the quadratic equation
* A*x*x + B*x + C = 0, provided it has any roots. If A == 0 or
* if the discriminant, B*B - 4*A*C, is negative, then an exception
* of type IllegalArgumentException is thrown.
*/
static public double root( double A, double B, double C )
throws IllegalArgumentException {
if (A == 0) {
throw new IllegalArgumentException("A can’t be zero.");
}
else {
double disc = B*B - 4*A*C;
if (disc < 0)
throw new IllegalArgumentException("Discriminant < zero.");
return (-B + Math.sqrt(disc)) / (2*A);
}
}
Your program should allow the user to specify values for A, B, and C. It should call the
subroutine to compute a solution of the equation. If no error occurs, it should print the
root. However, if an error occurs, your program should catch that error and print an error
message. After processing one equation, the program should ask whether the user wants
to enter another equation. The program should continue until the user answers no.
2. As discussed in Section 8.1, values of type int are limited to 32 bits. Integers that are too (solution)
large to be represented in 32 bits cannot be stored in an int variable. Java has a standard
class, java.math.BigInteger, that addresses this problem. An object of type BigInteger
is an integer that can be arbitrarily large. (The maximum size is limited only by the
amount of memory available to the Java Virtual Machine.) Since BigIntegers are objects,
they must be manipulated using instance methods from the BigInteger class. For example,
you can’t add two BigIntegers with the + operator. Instead, if N and M are variables that
refer to BigIntegers, you can compute the sum of N and M with the function call N.add(M).
The value returned by this function is a new BigInteger object that is equal to the sum of
N and M.
The BigInteger class has a constructor new BigInteger(str), where str is a string.
The string must represent an integer, such as “3” or “39849823783783283733”. If the string
does not represent a legal integer, then the constructor throws a NumberFormatException.
There are many instance methods in the BigInteger class. Here are a few that you will
find useful for this exercise. Assume that N and M are variables of type BigInteger.
• N.add(M) — a function that returns a BigInteger representing the sum of N and M.
• N.multiply(M) — a function that returns a BigInteger representing the result of
multiplying N times M.
Exercises 425
3. A Roman numeral represents an integer using letters. Examples are XVII to represent 17, (solution)
MCMLIII for 1953, and MMMCCCIII for 3303. By contrast, ordinary numbers such as
17 or 1953 are called Arabic numerals. The following table shows the Arabic equivalent
of all the single-letter Roman numerals:
M 1000 X 10
D 500 V 5
C 100 I 1
L 50
When letters are strung together, the values of the letters are just added up, with the
following exception. When a letter of smaller value is followed by a letter of larger value,
the smaller value is subtracted from the larger value. For example, IV represents 5 - 1, or
4. And MCMXCV is interpreted as M + CM + XC + V, or 1000 + (1000 - 100) + (100 -
10) + 5, which is 1995. In standard Roman numerals, no more than three consecutive
copies of the same letter are used. Following these rules, every number between 1 and
3999 can be represented as a Roman numeral made up of the following one- and two-letter
combinations:
M 1000 X 10
CM 900 IX 9
D 500 V 5
CD 400 IV 4
Exercises 426
C 100 I 1
XC 90
L 50
XL 40
Write a class to represent Roman numerals. The class should have two constructors.
One constructs a Roman numeral from a string such as “XVII” or “MCMXCV”. It should
throw a NumberFormatException if the string is not a legal Roman numeral. The other
constructor constructs a Roman numeral from an int. It should throw a NumberForma-
tException if the int is outside the range 1 to 3999.
In addition, the class should have two instance methods. The method toString()
returns the string that represents the Roman numeral. The method toInt() returns the
value of the Roman numeral as an int.
At some point in your class, you will have to convert an int into the string that
represents the corresponding Roman numeral. One way to approach this is to gradually
“move” value from the Arabic numeral to the Roman numeral. Here is the beginning of a
routine that will do this, where number is the int that is to be converted:
String roman = "";
int N = number;
while (N >= 1000) {
// Move 1000 from N to roman.
roman += "M";
N -= 1000;
}
while (N >= 900) {
// Move 900 from N to roman.
roman += "CM";
N -= 900;
}
.
. // Continue with other values from the above table.
.
(You can save yourself a lot of typing in this routine if you use arrays in a clever way to
represent the data in the above table.)
Once you’ve written your class, use it in a main program that will read both Arabic
numerals and Roman numerals entered by the user. If the user enters an Arabic numeral,
print the corresponding Roman numeral. If the user enters a Roman numeral, print the
corresponding Arabic numeral. (You can tell the difference by using TextIO.peek() to
peek at the first character in the user’s input (see Subsection 8.2.2). If the first character
is a digit, then the user’s input is an Arabic numeral. Otherwise, it’s a Roman numeral.)
The program should end when the user inputs an empty line.
4. The source code file Expr.java defines a class, Expr, that can be used to represent math- (solution)
ematical expressions involving the variable x. The expression can use the operators +, -,
*, /, and ^ (where ^ represents the operation of raising a number to a power). It can use
mathematical functions such as sin, cos, abs, and ln. See the source code file for full
details. The Expr class uses some advanced techniques which have not yet been covered in
this textbook. However, the interface is easy to understand. It contains only a constructor
and two public methods.
Exercises 427
The constructor new Expr(def) creates an Expr object defined by a given expres-
sion. The parameter, def, is a string that contains the definition. For example,
new Expr("x^2") or new Expr("sin(x)+3*x"). If the parameter in the constructor call
does not represent a legal expression, then the constructor throws an IllegalArgumentEx-
ception. The message in the exception describes the error.
If func is a variable of type Expr and num is of type double, then func.value(num)
is a function that returns the value of the expression when the number num is substituted
for the variable x in the expression. For example, if Expr represents the expression 3*x+1,
then func.value(5) is 3*5+1, or 16. If the expression is undefined for the specified value
of x, then the special value Double.NaN is returned; no exception is thrown.
Finally, func.toString() returns the definition of the expression. This is just the
string that was used in the constructor that created the expression object.
For this exercise, you should write a program that lets the user enter an expression. If
the expression contains an error, print an error message. Otherwise, let the user enter some
numerical values for the variable x. Print the value of the expression for each number that
the user enters. However, if the expression is undefined for the specified value of x, print
a message to that effect. You can use the boolean-valued function Double.isNaN(val)
to check whether a number, val, is Double.NaN.
The user should be able to enter as many values of x as desired. After that, the user
should be able to enter a new expression.
5. This exercise uses the class Expr, which was described in Exercise 8.4 and which is defined (solution)
in the source code file Expr.java. For this exercise, you should write a GUI program that
can graph a function, f(x), whose definition is entered by the user. The program should
have a text-input box where the user can enter an expression involving the variable x, such
as x^2 or sin(x-3)/x. This expression is the definition of the function. When the user
clicks an “Enter” button or presses return, the program should use the contents of the
text input box to construct an object of type Expr. If an error is found in the definition,
then the program should display an error message. Otherwise, it should display a graph of
the function. (Recall: If you make a button into the default button for the window, then
pressing return will be equivalent to clicking the button (see the end of Subsection 6.4.2).)
The program will need a Canvas for displaying the graph. To keep things simple,
the canvas should represent a fixed region in the xy-plane, defined by -5 <= x <= 5
and -5 <= y <= 5. To draw the graph, compute a large number of points and connect
them with line segments. (This method does not handle discontinuous functions properly;
doing so is very hard, so you shouldn’t try to do it for this exercise.) My program
divides the interval -5 <= x <= 5 into 300 subintervals and uses the 301 endpoints of
these subintervals for drawing the graph. Note that the function might be undefined at
one of these x-values. In that case, you have to skip that point.
A point on the graph has the form (x,y) where y is obtained by evaluating the user’s
expression at the given value of x. You will have to convert x and y values in the range
from -5 to 5 to the pixel coordinates that you need for drawing on the canvas. The formulas
for the conversion are:
double a = ( (x + 5)/10 * width );
double b = ( (5 - y)/10 * height );
where a and b are the horizontal and vertical coordinates that you can use for drawing on
the canvas. The values of width and height give the size of the canvas.
Quiz 428
Quiz on Chapter 8
(answers)
1. Why do programming languages require that variables be declared before they are used?
What does this have to do with correctness and robustness?
4. Find a useful loop invariant for the while loop in the binary search algorithm (Subsec-
tion 7.4.1).
5. Java has a predefined class called Throwable. What does this class represent? Why does
it exist?
6. Write a method that prints out a 3N+1 sequence starting from a given integer, N. The
starting value should be a parameter to the method. If the parameter is less than or equal
to zero, throw an IllegalArgumentException. If the number in the sequence becomes too
large to be represented as a value of type int, throw an ArithmeticException.
7. Rewrite the method from the previous question, using assert statements instead of ex-
ceptions to check for errors. What is the difference between the two versions of the method
when the program is run?
8. Some classes of exceptions are checked exceptions that require mandatory exception han-
dling. Explain what this means.
10. Why should a subroutine throw an exception when it encounters an error? Why not just
terminate the program?
11. Suppose that you have a choice of two algorithms that perform the same task. One has
average-case run time that is Θ(n2 ) while the run time of the second algorithm has an
average-case run time that is Θ(n*log(n)). Suppose that you need to process an input
of size n = 100. Which algorithm would you choose? Can you be certain that you are
choosing the fastest algorithm for the input that you intend to process?
12. Analyze the run time of the following algorithm. That is, find a function f(n) such that
the run time of the algorithm is O(f(n)) or, better, Θ(f(n)). Assume that A is an array of
integers, and use the length of the array as the input size, n.
int total = 0;
for (int i = 0; i < A.length; i++) {
if (A[i] > 0)
total = total + A[i];
}
Chapter 9
In this chapter, we look at two advanced programming techniques, recursion and linked data
structures, and some of their applications. Both of these techniques are related to the seemingly
paradoxical idea of defining something in terms of itself. This turns out to be a remarkably
powerful idea.
A subroutine is said to be recursive if it calls itself, either directly or indirectly. What this
means is that the subroutine is used in its own definition. Recursion can often be used to solve
complex problems by reducing them to simpler problems of the same type.
A reference to one object can be stored in an instance variable of another object. The
objects are then said to be “linked.” Complex data structures can be built by linking objects
together. An especially interesting case occurs when an object contains a link to another object
that belongs to the same class. In that case, the class is used in its own definition. Several
important types of data structures are built using classes of this kind.
9.1 Recursion
At one time or another, you’ve probably been told that you can’t define something in
terms of itself. Nevertheless, if it’s done right, defining something at least partially in terms of
itself can be a very powerful technique. A recursive definition is one that uses the concept
or thing that is being defined as part of the definition. For example: An “ancestor” is either
a parent or an ancestor of a parent. A “sentence” can be, among other things, two sentences
joined by a conjunction such as “and.” A “directory” is a part of a disk drive that can hold files
and directories. In mathematics, a “set” is a collection of elements, which can themselves be
sets. A “statement” in Java can be a while statement, which is made up of the word “while”,
a boolean-valued condition, and a statement.
Recursive definitions can describe very complex situations with just a few words. A def-
inition of the term “ancestor” without using recursion might go something like “a parent, or
a grandparent, or a great-grandparent, or a great-great-grandparent, and so on.” But saying
“and so on” is not very rigorous. (I’ve often thought that recursion is really just a rigorous
way of saying “and so on.”) You run into the same problem if you try to define a “directory”
as “a file that is a list of files, where some of the files can be lists of files, where some of those
files can be lists of files, and so on.” Trying to describe what a Java statement can look like,
without using recursion in the definition, would be difficult and probably pretty comical.
429
CHAPTER 9. LINKED DATA STRUCTURES AND RECURSION 430
result can be an infinite recursion, where the subroutine keeps calling itself over and over,
without ever reaching a base case. Infinite recursion is similar to an infinite loop. However,
since each recursive call to the subroutine uses up some of the computer’s memory, a program
that is stuck in an infinite recursion will run out of memory and crash before long. In Java, the
program will crash with an exception of type StackOverflowError.
The problem is to move ten disks from Stack 0 to Stack 1, subject to the rules given above.
Stack 2 can be used as a spare location. Can we reduce this to smaller problems of the same
type, possibly generalizing the problem a bit to make this possible? It seems natural to consider
the size of the problem to be the number of disks to be moved. If there are N disks in Stack 0,
we know that we will eventually have to move the bottom disk from Stack 0 to Stack 1. But
before we can do that, according to the rules, the first N-1 disks must be on Stack 2. Once
we’ve moved the N-th disk to Stack 1, we must move the other N-1 disks from Stack 2 to Stack 1
to complete the solution. But moving N-1 disks is the same type of problem as moving N disks,
except that it’s a smaller version of the problem. This is exactly what we need to do recursion!
The problem has to be generalized a bit, because the smaller problems involve moving disks
CHAPTER 9. LINKED DATA STRUCTURES AND RECURSION 433
from Stack 0 to Stack 2 or from Stack 2 to Stack 1, instead of from Stack 0 to Stack 1. In the
recursive subroutine that solves the problem, the stacks that serve as the source and destination
of the disks have to be specified. It’s also convenient to specify the stack that is to be used as
a spare, even though we could figure that out from the other two parameters. The base case is
when there is only one disk to be moved. The solution in this case is trivial: Just move the disk
in one step. Here is a version of the subroutine that will print out step-by-step instructions for
solving the problem:
/**
* Solve the problem of moving the number of disks specified
* by the first parameter from the stack specified by the
* second parameter to the stack specified by the third
* parameter. The stack specified by the fourth parameter
* is available for use as a spare. Stacks are specified by
* number: 0, 1, or 2.
*/
static void towersOfHanoi(int disks, int from, int to, int spare) {
if (disks == 1) {
// There is only one disk to be moved. Just move it.
System.out.printf("Move disk 1 from stack %d to stack %d%n",
from, to);
}
else {
// Move all but one disk to the spare stack, then
// move the bottom disk, then put all the other
// disks on top of it.
towersOfHanoi(disks-1, from, spare, to);
System.out.printf("Move disk %d from stack %d to stack %d%n",
disks, from, to);
towersOfHanoi(disks-1, spare, to, from);
}
}
This subroutine just expresses the natural recursive solution. The recursion works be-
cause each recursive call involves a smaller number of disks, and the problem is trivial to
solve in the base case, when there is only one disk. To solve the “top level” problem of
moving N disks from Stack 0 to Stack 1, the subroutine should be called with the command
TowersOfHanoi(N,0,1,2). The subroutine is used in the sample program TowersOfHanoi.java.
Here, for example, is the output from the program when it is run with the number of disks set
equal to 4:
Move disk 1 from stack 0 to stack 2
Move disk 2 from stack 0 to stack 1
Move disk 1 from stack 2 to stack 1
Move disk 3 from stack 0 to stack 2
Move disk 1 from stack 1 to stack 0
Move disk 2 from stack 1 to stack 2
Move disk 1 from stack 0 to stack 2
Move disk 4 from stack 0 to stack 1
Move disk 1 from stack 2 to stack 1
Move disk 2 from stack 2 to stack 0
Move disk 1 from stack 1 to stack 0
Move disk 3 from stack 2 to stack 1
CHAPTER 9. LINKED DATA STRUCTURES AND RECURSION 434
}
else {
// Apply quicksortStep and get the new pivot position.
// Then apply quicksort to sort the items that
// precede the pivot and the items that follow it.
int pivotPosition = quicksortStep(A, lo, hi);
quicksort(A, lo, pivotPosition - 1);
quicksort(A, pivotPosition + 1, hi);
}
}
As usual, we had to generalize the problem. The original problem was to sort an array, but
the recursive algorithm is set up to sort a specified part of an array. To sort an entire array, A,
using the quickSort() subroutine, you would call quicksort(A, 0, A.length - 1).
Quicksort is an interesting example from the point of view of the analysis of algorithms
(Section 8.5), because its average case run time differs greatly from its worst case run time.
Here is a very informal analysis, starting with the average case: Note that an application of
quicksortStep divides a problem into two sub-problems. On the average, the subproblems will
be of approximately the same size. A problem of size n is divided into two problems that are
roughly of size n/2; these are then divided into four problems that are roughly of size n/4;
and so on. Since the problem size is divided by 2 on each level, there will be approximately
log(n) levels of subdivision. The amount of processing on each level is proportional to n. (On
the top level, each element in the array is looked at and possibly moved. On the second level,
where there are two subproblems, every element but one in the array is part of one of those
two subproblems and must be looked at and possibly moved, so there is a total of about n
steps in both subproblems combined. Similarly, on the third level, there are four subproblems
and a total of about n steps in the four subproblems on that level. . . .) With a total of n steps
on each level and approximately log(n) levels in the average case, the average case run time
for Quicksort is Θ(n*log(n)). This analysis assumes that quicksortStep divides a problem into
two approximately equal parts. However, in the worst case, each application of quicksortStep
divides a problem of size n into a problem of size 0 and a problem of size n-1. This happens
when the pivot element ends up at the beginning or end of the array. In this worst case, there
are n levels of subproblems, and the worst-case run time is Θ(n2 ). The worst case is very
rare—it depends on the items in the array being arranged in a very special way, so the average
performance of Quicksort can be very good even though it is not so good in certain rare cases.
(One of these “rare” cases is when the original array is already sorted or almost sorted, which
is really not all that rare in practice. Applying the Quicksort algorithm as given above to a
large sorted array will take a long time. One way to avoid that—with high probablility—is to
pick the pivot for QuickSort step at random, rather than always using the first item.)
There are sorting algorithms that have both an average case and a worst case run time of
Θ(n*log(n)). One example that is fairly easy to understand is MergeSort, which you can look
up if you are interested.
The gray or red squares are considered to be “filled” and the white squares are “empty.”
For the purposes of this example, we define a “blob” to consist of a filled square and all the
filled squares that can be reached from it by moving up, down, left, and right through other
filled squares. If the user clicks on any filled square in the program, the computer will count
the squares in the blob that contains the clicked square, and it will change the color of those
squares to red. In the picture, one of the blobs is shown in red. The program has several
controls. There is a “New Blobs” button; clicking this button will create a new random pattern
in the grid. A pop-up menu specifies the approximate percentage of squares that will be filled
in the new pattern. The more filled squares, the larger the blobs. And a button labeled “Count
the Blobs” will tell you how many different blobs there are in the pattern.
Recursion is used in this program to count the number of squares in a blob. Without
recursion, this would be a very difficult thing to implement. Recursion makes it relatively easy,
but it still requires a new technique, which is also useful in a number of other applications.
The data for the grid of squares is stored in a two dimensional array of boolean values,
boolean[][] filled;
The value of filled[r][c] is true if the square in row r and in column c of the grid is
filled. The number of rows in the grid is stored in an instance variable named rows, and the
number of columns is stored in columns. The program uses a recursive instance method named
getBlobSize(r,c) to count the number of squares in a blob. The parameters r and c tell
which blob to count, namely the blob that includes the square in a row r and column c. If
there is no filled square at position (r,c), then the answer is zero. Otherwise, getBlobSize()
has to count all the filled squares that can be reached from the square at position (r,c). The
idea is to use getBlobSize() recursively to get the number of filled squares that can be reached
from each of the neighboring positions: (r+1,c), (r-1,c), (r,c+1), and (r,c-1). Add up
these numbers, and add one to count the square at (r,c) itself, and you get the total number
of filled squares that can be reached from (r,c). Here is an implementation of this algorithm,
as stated. Unfortunately, it has a serious flaw: It leads to an infinite recursion!
int getBlobSize(int r, int c) { // BUGGY, INCORRECT VERSION!!
// This INCORRECT method tries to count all the filled
CHAPTER 9. LINKED DATA STRUCTURES AND RECURSION 439
}
if (filled[r][c] == false || visited[r][c] == true) {
// This square is not part of a blob, or else it has
// already been counted, so return zero.
return 0;
}
visited[r][c] = true; // Mark the square as visited so that
// we won’t count it again during the
// following recursive calls.
int size = 1; // Count the square at this position, then count the
// the blobs that are connected to this square
// horizontally or vertically.
size += getBlobSize(r-1,c);
size += getBlobSize(r+1,c);
size += getBlobSize(r,c-1);
size += getBlobSize(r,c+1);
return size;
} // end getBlobSize()
In the program, this method is used to determine the size of a blob when the user clicks
on a square. After getBlobSize() has performed its task, all the squares in the blob are still
marked as visited. The method that draws the grid of squares shows visited squares in red,
which makes the blob visible. The getBlobSize() method is also used for counting blobs. This
is done by the following method, which includes comments to explain how it works:
/**
* When the user clicks the "Count the Blobs" button, find the
* number of blobs in the grid and report the number in the
* message label.
*/
void countBlobs() {
int count = 0; // Number of blobs.
/* First clear out the visited array. The getBlobSize() method
will mark every filled square that it finds by setting the
corresponding element of the array to true. Once a square
has been marked as visited, it will stay marked until all the
blobs have been counted. This will prevent the same blob from
being counted more than once. */
for (int r = 0; r < rows; r++)
for (int c = 0; c < columns; c++)
visited[r][c] = false;
/* For each position in the grid, call getBlobSize() to get the
size of the blob at that position. If the size is not zero,
count a blob. Note that if we come to a position that was part
of a previously counted blob, getBlobSize() will return 0 and
the blob will not be counted again. */
for (int r = 0; r < rows; r++)
for (int c = 0; c < columns; c++) {
if (getBlobSize(r,c) > 0)
count++;
}
CHAPTER 9. LINKED DATA STRUCTURES AND RECURSION 441
Now, suppose that we want to know how many levels of supervisors there are between a
given employee and the boss. We just have to follow the chain of command through a series of
supervisor links, and count how many steps it takes to get to the boss:
if ( emp.supervisor == null ) {
System.out.println( emp.name + " is the boss!" );
}
else {
Employee runner; // For "running" up the chain of command.
runner = emp.supervisor;
if ( runner.supervisor == null) {
System.out.println( emp.name + " reports directly to the boss." );
}
else {
int count = 0;
while ( runner.supervisor != null ) {
count++; // Count the supervisor on this level.
runner = runner.supervisor; // Move up to the next level.
}
System.out.println( "There are " + count
+ " supervisors between " + emp.name
+ " and the boss." );
}
}
As the while loop is executed, runner points in turn to the original employee (emp), then
to emp’s supervisor, then to the supervisor of emp’s supervisor, and so on. The count
variable is incremented each time runner “visits” a new employee. The loop ends when
runner.supervisor is null, which indicates that runner has reached the boss. At that point,
count has counted the number of steps between emp and the boss.
In this example, the supervisor variable is quite natural and useful. In fact, data structures
that are built by linking objects together are so useful that they are a major topic of study
in computer science. We’ll be looking at a few typical examples. In this section and the
next, we’ll be looking at linked lists. A linked list consists of a chain of objects of the same
type, linked together by pointers from one object to the next. This is much like the chain of
supervisors between emp and the boss in the above example. It’s also possible to have more
complex situations, in which one object can contain links to several other objects. We’ll look
at an example of this in Section 9.4.
CHAPTER 9. LINKED DATA STRUCTURES AND RECURSION 443
null
null
null
If head is a variable of type IntNode that points to a linked list of integers, we can find the sum
of the integers in the list using:
int sum = 0;
IntNode runner = head;
while ( runner != null ) {
sum = sum + runner.item; // Add current item to the sum.
runner = runner.next;
}
System.out.println("The sum of the list of items is " + sum);
It is also possible to use recursion to process a linked list. Recursion is rarely the natural
way to process a list, since it’s so easy to use a loop to traverse the list. However, understanding
how to apply recursion to lists can help with understanding the recursive processing of more
complex data structures. A non-empty linked list can be thought of as consisting of two parts:
the head of the list, which is just the first node in the list, and the tail of the list, which
consists of the remainder of the list after the head. Note that the tail is itself a linked list and
that it is shorter than the original list (by one node). This is a natural setup for recursion,
where the problem of processing a list can be divided into processing the head and recursively
processing the tail. The base case occurs in the case of an empty list (or sometimes in the case
of a list of length one). For example, here is a recursive algorithm for adding up the numbers
in a linked list of integers:
if the list is empty then
return 0 (since there are no numbers to be added up)
otherwise
let listsum = the number in the head node
let tailsum be the sum of the numbers in the tail list (recursively)
add tailsum to listsum
return listsum
One remaining question is, how do we get the tail of a non-empty linked list? If head is a
variable that points to the head node of the list, then head.next is a variable that points to
the second node of the list—and that node is in fact the first node of the tail. So, we can view
head.next as a pointer to the tail of the list. One special case is when the original list consists
of a single node. In that case, the tail of the list is empty, and head.next is null. Since an
empty list is represented by a null pointer, head.next represents the tail of the list even in this
special case. This allows us to write a recursive list-summing function in Java as
/**
* Compute the sum of all the integers in a linked list of integers.
* @param head a pointer to the first node in the linked list
*/
public static int addItemsInList( IntNode head ) {
if ( head == null ) {
// Base case: The list is empty, so the sum is zero.
return 0;
}
else {
// Recursive case: The list is non-empty. Find the sum of
// the tail list, and add that to the item in the head node.
// (Note that this case could be written simply as
// return head.item + addItemsInList( head.next );)
int listsum = head.item;
CHAPTER 9. LINKED DATA STRUCTURES AND RECURSION 446
∗ ∗ ∗
In the rest of this section, we’ll look at a few more advanced operations on a linked list of
strings. The subroutines that we consider are instance methods in a class that I wrote named
StringList. An object of type StringList represents a linked list of strings. The class has a private
instance variable named head of type Node that points to the first node in the list, or is null
if the list is empty. Instance methods in class StringList access head as a global variable. The
source code for StringList is in the file StringList.java, and it is used in a sample program named
ListDemo.java, so you can take a look at the code in context if you want.
One of the methods in the StringList class searches the list, looking for a specified string.
If the string that we are looking for is searchItem, then we have to compare searchItem to
each item in the list. This is an example of basic list traversal and processing. However, in this
case, we can stop processing if we find the item that we are looking for.
/**
* Searches the list for a specified item.
* @param searchItem the item that is to be searched for
* @return true if searchItem is one of the items in the list or false if
* searchItem does not occur in the list.
*/
public boolean find(String searchItem) {
Node runner; // A pointer for traversing the list.
runner = head; // Start by looking at the head of the list.
// (head is an instance variable! )
while ( runner != null ) {
// Go through the list looking at the string in each
// node. If the string is the one we are looking for,
CHAPTER 9. LINKED DATA STRUCTURES AND RECURSION 447
// return true, since the string has been found in the list.
if ( runner.item.equals(searchItem) )
return true;
runner = runner.next; // Move on to the next node.
}
// At this point, we have looked at all the items in the list
// without finding searchItem. Return false to indicate that
// the item does not exist in the list.
return false;
} // end find()
It is possible that the list is empty, that is, that the value of head is null. We should be
careful that this case is handled properly. In the above code, if head is null, then the body
of the while loop is never executed at all, so no nodes are processed and the return value is
false. This is exactly what we want when the list is empty, since the searchItem can’t occur
in an empty list.
runner:
previous:
newNode: "larry"
Once we have previous and runner pointing to the right nodes, the command
“previous.next = newNode;” can be used to make previous.next point to the new node.
CHAPTER 9. LINKED DATA STRUCTURES AND RECURSION 448
And the command “newNode.next = runner” will set newNode.next to point to the correct
place. However, before we can use these commands, we need to set up runner and previous
as shown in the illustration. The idea is to start at the first node of the list, and then move
along the list past all the items that are less than the new item. While doing this, we have to
be aware of the danger of “falling off the end of the list.” That is, we can’t continue if runner
reaches the end of the list and becomes null. If insertItem is the item that is to be inserted,
and if we assume that it does, in fact, belong somewhere in the middle of the list, then the
following code would correctly position previous and runner:
Node runner, previous;
previous = head; // Start at the beginning of the list.
runner = head.next;
while ( runner != null && runner.item.compareTo(insertItem) < 0 ) {
previous = runner; // "previous = previous.next" would also work
runner = runner.next;
}
(This uses the compareTo() instance method from the String class to test whether the item in
the node is less than the item that is being inserted. See Subsection 2.3.3.)
This is fine, except that the assumption that the new node is inserted into the middle of
the list is not always valid. It might be that insertItem is less than the first item of the list.
In that case, the new node must be inserted at the head of the list. This can be done with the
instructions
newNode.next = head; // Make newNode.next point to the old head.
head = newNode; // Make newNode the new head of the list.
It is also possible that the list is empty. In that case, newNode will become the first and only
node in the list. This can be accomplished simply by setting head = newNode. The following
insert() method from the StringList class covers all of these possibilities:
/**
* Insert a specified item into the list, keeping the list in order.
* @param insertItem the item that is to be inserted.
*/
public void insert(String insertItem) {
Node newNode; // A Node to contain the new item.
newNode = new Node();
newNode.item = insertItem; // (N.B. newNode.next is null.)
if ( head == null ) {
// The new item is the first (and only) one in the list.
// Set head to point to it.
head = newNode;
}
else if ( head.item.compareTo(insertItem) >= 0 ) {
// The new item is less than the first item in the list,
// so it has to be inserted at the head of the list.
newNode.next = head;
head = newNode;
}
else {
// The new item belongs somewhere after the first item
// in the list. Search for its proper position and insert it.
CHAPTER 9. LINKED DATA STRUCTURES AND RECURSION 449
abstract data type. Any sequence of Strings that is arranged in increasing order is a possible
value of this data type. The operations on the data type include inserting a new string, deleting
a string, and finding a string in the list. There are often several different ways to implement the
same abstract data type. For example, the “ordered list of strings” ADT can be implemented
as a linked list or as an array. A program that only depends on the abstract definition of the
ADT can use either implementation, interchangeably. In particular, the implementation of the
ADT can be changed without affecting the program as a whole. This can make the program
easier to debug and maintain, so ADTs are an important tool in software engineering.
In this section, we’ll look at two common abstract data types, stacks and queues. Both
stacks and queues are often implemented as linked lists, but that is not the only possible
implementation. You should think of the rest of this section partly as a discussion of stacks
and queues and partly as a case study in ADTs.
9.3.1 Stacks
A stack consists of a sequence of items, which should be thought of as piled one on top of
the other like a physical stack of boxes or cafeteria trays. Only the top item on the stack is
accessible at any given time. It can be removed from the stack with an operation called pop.
An item lower down on the stack can only be removed after all the items on top of it have been
popped off the stack. A new item can be added to the top of the stack with an operation called
push. We can make a stack of any type of items. If, for example, the items are values of type
int, then the push and pop operations can be implemented as instance methods
• void push(int newItem) — Add newItem to top of stack.
• int pop() — Remove the top int from the stack and return it.
It is an error to try to pop an item from an empty stack, so it is important to be able to tell
whether a stack is empty. We need another stack operation to do the test, implemented as an
instance method
• boolean isEmpty() — Returns true if the stack is empty.
This defines “stack of ints” as an abstract data type. This ADT can be implemented in several
ways, but however it is implemented, its behavior must correspond to the abstract mental image
of a stack.
In a stack, all operations take place at the
"top" of the stack. The "push" operation
adds an item to the top of the stack.
The "pop" operation removes the item on
the top of the stack and returns it.
12 83
25 25 25
123 123 123
5 5 5
17 17 17
In the linked list implementation of a stack, the top of the stack is actually the node at the
head of the list. It is easy to add and remove nodes at the front of a linked list—much easier
than inserting and deleting nodes in the middle of the list. Here is a class that implements the
“stack of ints” ADT using a linked list. (It uses a static nested class to represent the nodes of
the linked list, but that is part of the private implementation of the ADT.)
public class StackOfInts {
/**
* An object of type Node holds one of the items in the linked list
* that represents the stack.
*/
private static class Node {
int item;
Node next;
}
private Node top; // Pointer to the Node that is at the top of
// of the stack. If top == null, then the
// stack is empty.
/**
* Add N to the top of the stack.
*/
public void push( int N ) {
Node newTop; // A Node to hold the new item.
newTop = new Node();
newTop.item = N; // Store N in the new Node.
newTop.next = top; // The new Node points to the old top.
top = newTop; // The new item is now on top.
}
/**
* Remove the top item from the stack, and return it.
* Throws an IllegalStateException if the stack is empty when
* this method is called.
*/
public int pop() {
if ( top == null )
throw new IllegalStateException("Can’t pop from an empty stack.");
int topItem = top.item; // The item that is being popped.
top = top.next; // The previous second item is now on top.
return topItem;
}
/**
* Returns true if the stack is empty. Returns false
* if there are one or more items on the stack.
*/
public boolean isEmpty() {
return (top == null);
}
} // end class StackOfInts
CHAPTER 9. LINKED DATA STRUCTURES AND RECURSION 453
You should make sure that you understand how the push and pop operations operate on the
linked list. Drawing some pictures might help. Note that the linked list is part of the private
implementation of the StackOfInts class. A program that uses this class doesn’t even need to
know that a linked list is being used.
Now, it’s pretty easy to implement a stack as an array instead of as a linked list. Since the
number of items on the stack varies with time, a counter is needed to keep track of how many
spaces in the array are actually in use. If this counter is called top, then the items on the stack
are stored in positions 0, 1, . . . , top-1 in the array. The item in position 0 is on the bottom
of the stack, and the item in position top-1 is on the top of the stack. Pushing an item onto
the stack is easy: Put the item in position top and add 1 to the value of top. If we don’t want
to put a limit on the number of items that the stack can hold, we can use the dynamic array
techniques from Subsection 7.2.4. Note that the typical picture of the array would show the
stack “upside down,” with the bottom of the stack at the top of the array. This doesn’t matter.
The array is just an implementation of the abstract idea of a stack, and as long as the stack
operations work the way they are supposed to, we are OK. Here is a second implementation of
the StackOfInts class, using a dynamic array:
import java.util.Arrays; // For the Arrays.copyOf() method.
public class StackOfInts { // (alternate version, using an array)
private int[] items = new int[10]; // Holds the items on the stack.
private int top = 0; // The number of items currently on the stack.
/**
* Add N to the top of the stack.
*/
public void push( int N ) {
if (top == items.length) {
// The array is full, so make a new, larger array and
// copy the current stack items into it.
items = Arrays.copyOf( items, 2*items.length );
}
items[top] = N; // Put N in next available spot.
top++; // Number of items goes up by one.
}
/**
* Remove the top item from the stack, and return it.
* Throws an IllegalStateException if the stack is empty when
* this method is called.
*/
public int pop() {
if ( top == 0 )
throw new IllegalStateException("Can’t pop from an empty stack.");
int topItem = items[top - 1]; // Top item in the stack.
top--; // Number of items on the stack goes down by one.
return topItem;
}
/**
* Returns true if the stack is empty. Returns false
* if there are one or more items on the stack.
*/
CHAPTER 9. LINKED DATA STRUCTURES AND RECURSION 454
9.3.2 Queues
Queues are similar to stacks in that a queue consists of a sequence of items, and there are
restrictions about how items can be added to and removed from the list. However, a queue has
two ends, called the front and the back of the queue. Items are always added to the queue at the
back and removed from the queue at the front. The operations of adding and removing items
are called enqueue and dequeue in this book. (These terms are not completely standardized,
in the way that “push” and “pop” are.) An item that is added to the back of the queue will
remain on the queue until all the items in front of it have been removed. This should sound
familiar. A queue is like a “line” or “queue” of customers waiting for service. Customers are
serviced in the order in which they arrive on the queue.
CHAPTER 9. LINKED DATA STRUCTURES AND RECURSION 455
Front Back
46 125 8 22 17
125 8 22 17
After d=>?=?=@A
125 8 22 17 83
Cfter en>?=?=@qD)
A queue can hold items of any type. For a queue of ints, the enqueue and dequeue operations
can be implemented as instance methods in a “QueueOfInts” class. We also need an instance
method for checking whether the queue is empty:
• void enqueue(int N) — Add N to the back of the queue.
• int dequeue() — Remove the item at the front and return it.
• boolean isEmpty() — Return true if the queue is empty.
A queue can be implemented as a linked list or as an array. An efficient array implementation
is trickier than the array implementation of a stack, so I won’t give it here. In the linked list
implementation, the first item of the list is at the front of the queue. Dequeueing an item from
the front of the queue is just like popping an item off a stack. The back of the queue is at the
end of the list. Enqueueing an item involves setting a pointer in the last node of the current
list to point to a new node that contains the item. To do this, we’ll need a command like
“tail.next = newNode;”, where tail is a pointer to the last node in the list. If head is a
pointer to the first node of the list, it would always be possible to get a pointer to the last node
of the list by saying:
Node tail; // This will point to the last node in the list.
tail = head; // Start at the first node.
while (tail.next != null) {
tail = tail.next; // Move to next node.
}
// At this point, tail.next is null, so tail points to
// the last node in the list.
However, it would be very inefficient to do this over and over every time an item is enqueued.
For the sake of efficiency, we’ll keep a pointer to the last node in an instance variable. This
complicates the class somewhat; we have to be careful to update the value of this variable
whenever a new node is added to the end of the list. Given all this, writing the QueueOfInts
class is not all that difficult:
public class QueueOfInts {
/**
* An object of type Node holds one of the items
* in the linked list that represents the queue.
CHAPTER 9. LINKED DATA STRUCTURES AND RECURSION 456
*/
private static class Node {
int item;
Node next;
}
private Node head = null; // Points to first Node in the queue.
// The queue is empty when head is null.
private Node tail = null; // Points to last Node in the queue.
/**
* Add N to the back of the queue.
*/
public void enqueue( int N ) {
Node newTail = new Node(); // A Node to hold the new item.
newTail.item = N;
if (head == null) {
// The queue was empty. The new Node becomes
// the only node in the list. Since it is both
// the first and last node, both head and tail
// point to it.
head = newTail;
tail = newTail;
}
else {
// The new node becomes the new tail of the list.
// (The head of the list is unaffected.)
tail.next = newTail;
tail = newTail;
}
}
/**
* Remove and return the front item in the queue.
* Throws an IllegalStateException if the queue is empty.
*/
public int dequeue() {
if ( head == null)
throw new IllegalStateException("Can’t dequeue from an empty queue.");
int firstItem = head.item;
head = head.next; // The previous second item is now first.
// If we have just removed the last item,
// then head is null.
if (head == null) {
// The queue has become empty. The Node that was
// deleted was the tail as well as the head of the
// list, so now there is no tail. (Actually, the
// class would work fine without this step.)
tail = null;
}
return firstItem;
}
/**
* Return true if the queue is empty.
CHAPTER 9. LINKED DATA STRUCTURES AND RECURSION 457
*/
boolean isEmpty() {
return (head == null);
}
} // end class QueueOfInts
To help you follow what is being done here with the tail pointer, it might help to think in
terms of a class invariant (Subsection 8.2.3): “If the queue is non-empty, then tail points to
the last node in the queue.” This invariant must be true at the beginning and at the end of each
method call. For example, applying this to the enqueue() method, in the case of a non-empty
list, the invariant tells us that a new node can be added to the back of the list simply by saying
“tail.next = newNode. It also tells us how the value of tail must be set before returning
from the method: it must be set to point to the node that was just added to the queue.
Queues are typically used in a computer (as in real life) when only one item can be processed
at a time, but several items can be waiting for processing. For example:
• In a Java program that has multiple threads, the threads that want processing time on
the CPU are kept in a queue. When a new thread is started, it is added to the back of the
queue. A thread is removed from the front of the queue, it is given some processing time,
and then—if it has not terminated—is sent to the back of the queue to wait for another
turn.
• Events such as keystrokes and mouse clicks are stored in a queue called the “event queue.”
A program removes events from the event queue and processes them. It’s possible for
several more events to occur while one event is being processed, but since the events are
stored in a queue, they will always be processed in the order in which they occurred.
• A web server is a program that receives requests from web browsers for “pages.” It is easy
for new requests to arrive while the web server is still fulfilling a previous request. Requests
that arrive while the web server is busy are placed into a queue to await processing. Using
a queue ensures that requests will be processed in the order in which they were received.
Queues are said to implement a FIFO policy: First In, First Out. Or, as it is more
commonly expressed, first come, first served. Stacks, on the other hand implement a LIFO
policy: Last In, First Out. The item that comes out of the stack is the last one that was put
in. Just like queues, stacks can be used to hold items that are waiting for processing (although
in applications where queues are typically used, a stack would be considered “unfair”).
∗ ∗ ∗
To get a better handle on the difference between stacks and queues, consider the sample
program DepthBreadth.java. I suggest that you try out the program. The program shows
a grid of squares. Initially, all the squares are white. When you click on a white square,
that square is “marked” by turning it red. The program than starts marking squares that are
connected, horizontally or vertically, to squares that have already been marked. This process
will eventually process every square in the grid. To understand how the program works, think
of yourself in the place of the program. When the user clicks a square, you are handed an index
card. The location of the square—its row and column—is written on the card. You put the
card in a pile, which then contains just that one card. Then, you repeat the following: If the
pile is empty, you are done. Otherwise, remove an index card from the pile. The index card
specifies a square. Look at each horizontal and vertical neighbor of that square. If the neighbor
has not already been encountered, write its location on a new index card and put the card in
the pile. You are done when there are no more index cards waiting in the pile to be processed.
CHAPTER 9. LINKED DATA STRUCTURES AND RECURSION 458
In the program, while a square is in the pile, waiting to be processed, it is colored red;
that is, red squares have been encountered but not yet processed. When a square is taken from
the pile and processed, its color changes to gray. Once a square has been colored gray, the
program will never consider it again, since all of its neighbors have already been accounted for.
Eventually, all the squares have been processed, all the squares are gray, and the procedure
ends. In the index card analogy, the pile of cards has been emptied.
The program can use your choice of three methods: Stack, Queue, and Random. In each
case, the same general procedure is used. The only difference is how the “pile of index cards” is
managed. For a stack, cards are added and removed at the top of the pile. For a queue, cards
are added to the bottom of the pile and removed from the top. In the random case, the card to
be processed is picked at random from among all the cards in the pile. The order of processing
is very different in these three cases. Here are three pictures from the program, using the three
different processing methods. In each case, the process was started by selecting a square near
the middle of the grid. A stack is used for the picture on the left, a queue for the picture in
the middle, and random selection for the picture on the right:
The patterns that are produced are very different. When using a stack, the program explores
out as far as possible before it starts backtracking to look at previously encountered squares.
With a queue, squares are processed roughly in the order of their distance from the starting
point. When random selection is used, the result is an irregular blob, but it is a connected blob
since a square can only be encountered if it is next to a previously encountered square.
You should experiment with the program to see how it all works. Try to understand how
stacks and queues are being used. Try starting from one of the corner squares. While the
process is going on, you can click on other white squares, and they will be added to the list of
encountered squares. When you do this with a stack, you should notice that the square you
click is processed immediately, and all the red squares that were already waiting for processing
have to wait. On the other hand, if you do this with a queue, the square that you click will
wait its turn until all the squares that were already in the pile have been processed. Again, the
source code for the program is DepthBreadth.java.
∗ ∗ ∗
Queues seem very natural because they occur so often in real life, but there are times when
stacks are appropriate and even essential. For example, consider what happens when a routine
calls a subroutine. The first routine is suspended while the subroutine is executed, and it will
continue only when the subroutine returns. Now, suppose that the subroutine calls a second
subroutine, and the second subroutine calls a third, and so on. Each subroutine is suspended
CHAPTER 9. LINKED DATA STRUCTURES AND RECURSION 459
while the subsequent subroutines are executed. The computer has to keep track of all the
subroutines that are suspended. It does this with a stack.
When a subroutine is called, an activation record is created for that subroutine. The
activation record contains information relevant to the execution of the subroutine, such as its
local variables and parameters. The activation record for the subroutine is placed on a stack.
It will be removed from the stack and destroyed when the subroutine returns. If the subroutine
calls another subroutine, the activation record of the second subroutine is pushed onto the
stack, on top of the activation record of the first subroutine. The stack can continue to grow
as more subroutines are called, and it shrinks as those subroutines return.
In the case of a recursive subroutine, which calls itself, there can be several activation records
on the stack for the same subroutine. This is how the computer keeps track of many recursive
calls at the same time: It has a different activation record for each call.
}
else {
// Since the next item is not a number, the only thing
// it can legally be is an operator. Get the operator
// and perform the operation.
char op; // The operator, which must be +, -, *, /, or ^.
double x,y; // The operands, from the stack, for the operation.
double answer; // The result, to be pushed onto the stack.
op = TextIO.getChar();
if (op != ’+’ && op != ’-’ && op != ’*’ && op != ’/’ && op != ’^’) {
// The character is not one of the acceptable operations.
System.out.println("\nIllegal operator found in input: " + op);
return;
}
if (stack.isEmpty()) {
System.out.println(" Stack is empty while trying to evaluate " + op);
System.out.println("\nNot enough numbers in expression!");
return;
}
y = stack.pop();
if (stack.isEmpty()) {
System.out.println(" Stack is empty while trying to evaluate " + op);
System.out.println("\nNot enough numbers in expression!");
return;
}
x = stack.pop();
switch (op) {
case ’+’:
answer = x + y;
break;
case ’-’:
answer = x - y;
break;
case ’*’:
answer = x * y;
break;
case ’/’:
answer = x / y;
break;
default:
answer = Math.pow(x,y); // (op must be ’^’.)
}
stack.push(answer);
System.out.println(" Evaluated " + op + " and pushed " + answer);
}
TextIO.skipBlanks();
} // end while
// If we get to this point, the input has been read successfully.
// If the expression was legal, then the value of the expression is
// on the stack, and it is the only thing on the stack.
if (stack.isEmpty()) { // Impossible if the input is really non-empty.
CHAPTER 9. LINKED DATA STRUCTURES AND RECURSION 462
} // end readAndEvaluate()
Postfix expressions are often used internally by computers. In fact, the Java virtual machine
is a “stack machine” which uses the stack-based approach to expression evaluation that we have
been discussing. The algorithm can easily be extended to handle variables, as well as constants.
When a variable is encountered in the expression, the value of the variable is pushed onto the
stack. It also works for operators with more or fewer than two operands. As many operands as
are needed are popped from the stack and the result is pushed back onto the stack. For example,
the unary minus operator, which is used in the expression “-x”, has a single operand. We
will continue to look at expressions and expression evaluation in the next two sections.
gbbi `bc]
2 3
null
4 5 6
null null null
null null null
\]^_ `bc]f
A node that has no children is called a leaf . A leaf node can be recognized by the fact that
both the left and right pointers in the node are null. In the standard picture of a binary tree,
the root node is shown at the top and the leaf nodes at the bottom—which doesn’t show much
respect for the analogy to real trees. But at least you can see the branching, tree-like structure
that gives a binary tree its name.
/**
* Print all the items in the tree to which root points.
* The items in the left subtree are printed first, followed
* by the item in the root node and then the items
* in the right subtree.
*/
CHAPTER 9. LINKED DATA STRUCTURES AND RECURSION 465
roops
jkmn
bill mary
muvw juxw
Binary sort trees have this useful property: An inorder traversal of the tree will process the
items in increasing order. In fact, this is really just another way of expressing the definition.
For example, if an inorder traversal is used to print the items in the tree shown above, then the
items will be in alphabetical order. The definition of an inorder traversal guarantees that all
the items in the left subtree of “judy” are printed before “judy”, and all the items in the right
subtree of “judy” are printed after “judy”. But the binary sort tree property guarantees that
the items in the left subtree of “judy” are precisely those that precede “judy” in alphabetical
order, and all the items in the right subtree follow “judy” in alphabetical order. So, we know
that “judy” is output in its proper alphabetical position. But the same argument applies to
the subtrees. “Bill” will be output after “alice” and before “fred” and its descendents. “Fred”
will be output after “dave” and before “jane” and “joe”. And so on.
Suppose that we want to search for a given item in a binary search tree. Compare that item
to the root item of the tree. If they are equal, we’re done. If the item we are looking for is
less than the root item, then we need to search the left subtree of the root—the right subtree
can be eliminated because it only contains items that are greater than or equal to the root.
Similarly, if the item we are looking for is greater than the item in the root, then we only need
to look in the right subtree. In either case, the same procedure can then be applied to search
the subtree. Inserting a new item is similar: Start by searching the tree for the position where
the new item belongs. When that position is found, create a new node and attach it to the tree
at that position.
Searching and inserting are efficient operations on a binary search tree, provided that the
tree is close to being balanced . A binary tree is balanced if for each node, the left subtree of
that node contains approximately the same number of nodes as the right subtree. In a perfectly
balanced tree, the two numbers differ by at most one. Not all binary trees are balanced, but if
the tree is created by inserting items in a random order, there is a high probability that the tree
is approximately balanced. (If the order of insertion is not random, however, it’s quite possible
for the tree to be very unbalanced.) During a search of any binary sort tree, every comparison
eliminates one of two subtrees from further consideration. If the tree is balanced, that means
cutting the number of items still under consideration in half. This is exactly the same as the
binary search algorithm, and the result is a similarly efficient algorithm.
In terms of asymptotic analysis (Section 8.5), searching, inserting, and deleting in a binary
search tree have average case run time Θ(log(n)). The problem size, n, is the number of items
in the tree, and the average is taken over all the different orders in which the items could have
been inserted into the tree. As long as the actual insertion order is random, the actual run
time can be expected to be close to the average. However, the worst case run time for binary
search tree operations is Θ(n), which is much worse than Θ(log(n)). The worst case occurs
for particular insertion orders. For example, if the items are inserted into the tree in order of
increasing size, then every item that is inserted moves always to the right as it moves down
the tree. The result is a “tree” that looks more like a linked list, since it consists of a linear
string of nodes strung together by their right child pointers. Operations on such a tree have
the same performance as operations on a linked list. Now, there are data structures that are
similar to simple binary sort trees, except that insertion and deletion of nodes are implemented
in a way that will always keep the tree balanced, or almost balanced. For these data structures,
searching, inserting, and deleting have both average case and worst case run times that are
Θ(log(n)). Here, however, we will look at only the simple versions of inserting and searching.
The sample program SortTreeDemo.java is a demonstration of binary sort trees. The pro-
gram includes subroutines that implement inorder traversal, searching, and insertion. We’ll
CHAPTER 9. LINKED DATA STRUCTURES AND RECURSION 467
look at the latter two subroutines below. The main() routine tests the subroutines by letting
you type in strings to be inserted into the tree.
In SortTreeDemo, nodes in the binary tree are represented using the following static nested
class, including a simple constructor that makes creating nodes easier:
/**
* An object of type TreeNode represents one node in a binary tree of strings.
*/
private static class TreeNode {
String item; // The data in this node.
TreeNode left; // Pointer to left subtree.
TreeNode right; // Pointer to right subtree.
TreeNode(String str) {
// Constructor. Make a node containing str.
// Note that left and right pointers are null.
item = str;
}
} // end class TreeNode
A static member variable of type TreeNode points to the binary sort tree that is used by the
program:
private static TreeNode root; // Pointer to the root node in the tree.
// When the tree is empty, root is null.
A recursive subroutine named treeContains is used to search for a given item in the tree. This
routine implements the search algorithm for binary trees that was outlined above:
/**
* Return true if item is one of the items in the binary
* sort tree to which root points. Return false if not.
*/
static boolean treeContains( TreeNode root, String item ) {
if ( root == null ) {
// Tree is empty, so it certainly doesn’t contain item.
return false;
}
else if ( item.equals(root.item) ) {
// Yes, the item has been found in the root node.
return true;
}
else if ( item.compareTo(root.item) < 0 ) {
// If the item occurs, it must be in the left subtree.
return treeContains( root.left, item );
}
else {
// If the item occurs, it must be in the right subtree.
return treeContains( root.right, item );
}
} // end treeContains()
When this routine is called in the main() routine, the first parameter is the static member
variable root, which points to the root of the entire binary sort tree.
It’s worth noting that recursion is not really essential in this case. A simple, non-recursive
algorithm for searching a binary sort tree follows the rule: Start at the root and move down
CHAPTER 9. LINKED DATA STRUCTURES AND RECURSION 468
the tree until you find the item or reach a null pointer. Since the search follows a single path
down the tree, it can be implemented as a while loop. Here is a non-recursive version of the
search routine:
private static boolean treeContainsNR( TreeNode root, String item ) {
TreeNode runner; // For "running" down the tree.
runner = root; // Start at the root node.
while (true) {
if (runner == null) {
// We’ve fallen off the tree without finding item.
return false;
}
else if ( item.equals(runner.item) ) {
// We’ve found the item.
return true;
}
else if ( item.compareTo(runner.item) < 0 ) {
// If the item occurs, it must be in the left subtree.
// So, advance the runner down one level to the left.
runner = runner.left;
}
else {
// If the item occurs, it must be in the right subtree.
// So, advance the runner down one level to the right.
runner = runner.right;
}
} // end while
} // end treeContainsNR();
The subroutine for inserting a new item into the tree turns out to be more similar to the
non-recursive search routine than to the recursive. The insertion routine has to handle the case
where the tree is empty. In that case, the value of root must be changed to point to a node
that contains the new item:
root = new TreeNode( newItem );
But this means, effectively, that the root can’t be passed as a parameter to the subroutine,
because it is impossible for a subroutine to change the value stored in an actual parameter.
(I should note that this is something that is possible in other languages.) Recursion uses
parameters in an essential way. There are ugly ways to work around the problem, but the
easiest thing is just to use a non-recursive insertion routine that accesses the static member
variable root directly. One difference between inserting an item and searching for an item is
that we have to be careful not to fall off the tree. That is, we have to stop searching just before
runner becomes null. When we get to an empty spot in the tree, that’s where we have to
insert the new node:
/**
* Add the item to the binary sort tree to which the global variable
* "root" refers. (Note that root can’t be passed as a parameter to
* this routine because the value of root might change, and a change
* in the value of a formal parameter does not change the actual parameter.)
*/
private static void treeInsert(String newItem) {
if ( root == null ) {
CHAPTER 9. LINKED DATA STRUCTURES AND RECURSION 469
values. The process is shown by the upward-directed arrows in the illustration. The value
computed for the root node is the value of the expression as a whole. There are other uses for
expression trees. For example, a postorder traversal of the tree will output the postfix form of
the expression.
ents 18
ession answer
- 5)
g
arrows show how the
value of the on
can be com
6 12
+
3 * 2 17 5
3 8 / 4 17 5
7 + 1 4
7 1
An expression tree contains two types of nodes: nodes that contain numbers and nodes that
contain operators. Furthermore, we might want to add other types of nodes to make the trees
more useful, such as nodes that contain variables. If we want to work with expression trees in
Java, how can we deal with this variety of nodes? One way—which will be frowned upon by
object-oriented purists—is to include an instance variable in each node object to record which
type of node it is:
enum NodeType { NUMBER, OPERATOR } // Possible kinds of node.
class ExpNode { // A node in an expression tree.
NodeType kind; // Which type of node is this?
double number; // The value in a node of type NUMBER.
char op; // The operator in a node of type OPERATOR.
ExpNode left; // Pointers to subtrees,
ExpNode right; // in a node of type OPERATOR.
ExpNode( double val ) {
// Constructor for making a node of type NUMBER.
kind = NodeType.NUMBER;
number = val;
}
ExpNode( char op, ExpNode left, ExpNode right ) {
// Constructor for making a node of type OPERATOR.
kind = NodeType.OPERATOR;
this.op = op;
this.left = left;
this.right = right;
}
} // end class ExpNode
CHAPTER 9. LINKED DATA STRUCTURES AND RECURSION 471
Given this definition, the following recursive subroutine will find the value of an expression tree:
static double getValue( ExpNode node ) {
// Return the value of the expression represented by
// the tree to which node refers. Node must be non-null.
if ( node.kind == NodeType.NUMBER ) {
// The value of a NUMBER node is the number it holds.
return node.number;
}
else { // The kind must be OPERATOR.
// Get the values of the operands and combine them
// using the operator.
double leftVal = getValue( node.left );
double rightVal = getValue( node.right );
switch ( node.op ) {
case ’+’: return leftVal + rightVal;
case ’-’: return leftVal - rightVal;
case ’*’: return leftVal * rightVal;
case ’/’: return leftVal / rightVal;
default: return Double.NaN; // Bad operator.
}
}
} // end getValue()
Although this approach works, a more object-oriented approach is to note that since there
are two types of nodes, there should be two classes to represent them, perhaps named ConstNode
and BinOpNode. To represent the general idea of a node in an expression tree, we need another
class, ExpNode. Both ConstNode and BinOpNode will be subclasses of ExpNode. Since any
actual node will be either a ConstNode or a BinOpNode, ExpNode should be an abstract class.
(See Subsection 5.5.5.) Since one of the things we want to do with nodes is find their values,
each class should have an instance method for finding the value:
abstract class ExpNode {
// Represents a node of any type in an expression tree.
abstract double value(); // Return the value of this node.
} // end class ExpNode
which are pretty well understood, can give some insight into the much more complex and difficult
natural languages. And programming languages raise more than enough interesting issues to
make them worth studying in their own right. How can it be, after all, that computers can be
made to “understand” even the relatively simple languages that are used to write programs?
Computers can only directly use instructions expressed in very simple machine language. Higher
level languages must be translated into machine language. But the translation is done by a
compiler, which is just a program. How could such a translation program be written?
∗ ∗ ∗
When we try to parse a phrase that contains a syntax error, we need some way to respond
to the error. A convenient way of doing this is to throw an exception. I’ll use an exception
class called ParseError, defined as follows:
/**
* An object of type ParseError represents a syntax error found in
* the user’s input.
*/
private static class ParseError extends Exception {
ParseError(String message) {
super(message);
}
} // end nested class ParseError
Another general point is that our BNF rules don’t say anything about spaces between items,
but in reality we want to be able to insert spaces between items at will. To allow for this, I’ll
always call the routine TextIO.skipBlanks() before trying to look ahead to see what’s coming
up next in input. TextIO.skipBlanks() skips past any whitespace, such as spaces and tabs, in
the input, and stops when the next character in the input is either a non-blank character or the
end-of-line character. (For a discussion of robust handling of TextIO input, see Subsection 8.2.4.)
Let’s start with a very simple example. A “fully parenthesized expression” can be specified
in BNF by the rules
<expression> ::= <number> |
"(" <expression> <operator> <expression> ")"
<operator> ::= "+" | "-" | "*" | "/"
where <number> refers to any non-negative real number. An example of a fully parenthesized
expression is “(((34-17)*8)+(2*7))”. Since every operator corresponds to a pair of parenthe-
ses, there is no ambiguity about the order in which the operators are to be applied. Suppose
we want a program that will read and evaluate such expressions. We’ll read the expressions
from standard input, using TextIO. To apply recursive descent parsing, we need a subroutine
for each rule in the grammar. Corresponding to the rule for <operator>, we get a subroutine
that reads an operator. The operator can be a choice of any of four things. Any other input
will be an error.
/**
* If the next character in input is one of the legal operators,
* read it and return it. Otherwise, throw a ParseError.
*/
static char getOperator() throws ParseError {
TextIO.skipBlanks();
char op = TextIO.peek(); // look ahead at the next char, without reading it
if ( op == ’+’ || op == ’-’ || op == ’*’ || op == ’/’ ) {
TextIO.getAnyChar(); // read the operator, to remove it from the input
return op;
}
else if (op == ’\n’)
throw new ParseError("Missing operator at end of line.");
else
throw new ParseError("Missing operator. Found \"" +
op + "\" instead of +, -, *, or /.");
} // end getOperator()
CHAPTER 9. LINKED DATA STRUCTURES AND RECURSION 476
I’ve tried to give a reasonable error message, depending on whether the next character is an
end-of-line or something else. I use TextIO.peek() to look ahead at the next character before I
read it, and I call TextIO.skipBlanks() before testing TextIO.peek() in order to ignore any
blanks that separate items. I will follow this same pattern in every case.
When we come to the subroutine for <expression>, things are a little more interesting. The
rule says that an expression can be either a number or an expression enclosed in parentheses.
We can tell which it is by looking ahead at the next character. If the character is a digit,
we have to read a number. If the character is a “(“, we have to read the “(“, followed by an
expression, followed by an operator, followed by another expression, followed by a “)”. If the
next character is anything else, there is an error. Note that we need recursion to read the
nested expressions. The routine doesn’t just read the expression. It also computes and returns
its value. This requires semantical information that is not specified in the BNF rule.
/**
* Read an expression from the current line of input and return its value.
* @throws ParseError if the input contains a syntax error
*/
private static double expressionValue() throws ParseError {
TextIO.skipBlanks();
if ( Character.isDigit(TextIO.peek()) ) {
// The next item in input is a number, so the expression
// must consist of just that number. Read and return
// the number.
return TextIO.getDouble();
}
else if ( TextIO.peek() == ’(’ ) {
// The expression must be of the form
// "(" <expression> <operator> <expression> ")"
// Read all these items, perform the operation, and
// return the result.
TextIO.getAnyChar(); // Read the "("
double leftVal = expressionValue(); // Read and evaluate first operand.
char op = getOperator(); // Read the operator.
double rightVal = expressionValue(); // Read and evaluate second operand.
TextIO.skipBlanks();
if ( TextIO.peek() != ’)’ ) {
// According to the rule, there must be a ")" here.
// Since it’s missing, throw a ParseError.
throw new ParseError("Missing right parenthesis.");
}
TextIO.getAnyChar(); // Read the ")"
switch (op) { // Apply the operator and return the result.
case ’+’: return leftVal + rightVal;
case ’-’: return leftVal - rightVal;
case ’*’: return leftVal * rightVal;
case ’/’: return leftVal / rightVal;
default: return 0; // Can’t occur since op is one of the above.
// (But Java syntax requires a return value.)
}
}
else { // No other character can legally start an expression.
throw new ParseError("Encountered unexpected character, \"" +
TextIO.peek() + "\" in input.");
CHAPTER 9. LINKED DATA STRUCTURES AND RECURSION 477
}
} // end expressionValue()
I hope that you can see how this routine corresponds to the BNF rule. Where the rule
uses “|” to give a choice between alternatives, there is an if statement in the routine to
determine which choice to take. Where the rule contains a sequence of items, “(“ <expression>
<operator> <expression> “)”, there is a sequence of statements in the subroutine to read each
item in turn.
When expressionValue() is called to evaluate the expression (((34-17)*8)+(2*7)), it
sees the “(“ at the beginning of the input, so the else part of the if statement is executed.
The “(“ is read. Then the first recursive call to expressionValue() reads and evaluates the
subexpression ((34-17)*8), the call to getOperator() reads the “+” operator, and the sec-
ond recursive call to expressionValue() reads and evaluates the second subexpression (2*7).
Finally, the “)” at the end of the expression is read. Of course, reading the first subexpression,
((34-17)*8), involves further recursive calls to the expressionValue() routine, but it’s better
not to think too deeply about that! Rely on the recursion to handle the details.
You’ll find a complete program that uses these routines in the file SimpleParser1.java.
∗ ∗ ∗
Fully parenthesized expressions aren’t very natural for people to use. But with ordinary
expressions, we have to worry about the question of operator precedence, which tells us, for
example, that the “*” in the expression “5+3*7” is applied before the “+”. The complex
expression “3*6+8*(7+1)/4-24” should be seen as made up of three “terms”, 3*6, 8*(7+1)/4,
and 24, combined with “+” and “-” operators. A term, on the other hand, can be made up
of several factors combined with “*” and “/” operators. For example, 8*(7+1)/4 contains the
factors 8, (7+1) and 4. This example also shows that a factor can be either a number or an
expression in parentheses. To complicate things a bit more, we allow for leading minus signs in
expressions, as in “-(3+4)” or “-7”. (Since a <number> is a positive number, this is the only
way we can get negative numbers. It’s done this way to avoid “3 * -7”, for example.) This
structure can be expressed by the BNF rules
<expression> ::= [ "-" ] <term> [ ( "+" | "-" ) <term> ]...
<term> ::= <factor> [ ( "*" | "/" ) <factor> ]...
<factor> ::= <number> | "(" <expression> ")"
The first rule uses the “[ ]...” notation, which says that the items that it encloses can occur
zero, one, two, or more times. The rule means that an <expression> can begin, optionally,
with a “-”. Then there must be a <term> which can optionally be followed by one of the
operators “+” or “-” and another <term>, optionally followed by another operator and <term>,
and so on. In a subroutine that reads and evaluates expressions, this repetition is handled by
a while loop. An if statement is used at the beginning of the loop to test whether a leading
minus sign is present:
/**
* Read an expression from the current line of input and return its value.
* @throws ParseError if the input contains a syntax error
*/
private static double expressionValue() throws ParseError {
TextIO.skipBlanks();
boolean negative; // True if there is a leading minus sign.
negative = false;
if (TextIO.peek() == ’-’) {
CHAPTER 9. LINKED DATA STRUCTURES AND RECURSION 478
In some real compilers, the parser creates a tree to represent the program that is being
parsed. This tree is called a parse tree or abstract syntax tree. Parse trees are somewhat
different in form from expression trees, but the purpose is the same. Once you have the tree,
there are a number of things you can do with it. For one thing, it can be used to generate
machine language code. But there are also techniques for examining the tree and detecting
certain types of programming errors, such as an attempt to reference a local variable before
it has been assigned a value. (The Java compiler, of course, will reject the program if it
contains such an error.) It’s also possible to manipulate the tree to optimize the program.
In optimization, the tree is transformed to make the program more efficient before the code is
generated.
And so we are back where we started in Chapter 1, looking at programming languages,
compilers, and machine language. But looking at them, I hope, with a lot more understanding
and a much wider perspective.
Exercises 481
1. In many textbooks, the first examples of recursion are the mathematical functions factorial (solution)
and fibonacci. These functions are defined for non-negative integers using the following
recursive formulas:
factorial(0) = 1
factorial(N) = N*factorial(N-1) for N > 0
fibonacci(0) = 1
fibonacci(1) = 1
fibonacci(N) = fibonacci(N-1) + fibonacci(N-2) for N > 1
Write recursive functions to compute factorial(N) and fibonacci(N) for a given non-
negative integer N, and write a main() routine to test your functions. Consider using the
BigInteger class (see Exercise 8.2)
(In fact, factorial and fibonacci are really not very good examples of recursion, since
the most natural way to compute them is to use simple for loops. Furthermore, fibonacci
is a particularly bad example, since the natural recursive approach to computing this
function is extremely inefficient.)
2. Exercise 7.6 asked you to read a file, make an alphabetical list of all the words that occur (solution)
in the file, and write the list to another file. In that exercise, you were asked to use an
ArrayList<String> to store the words. Write a new version of the same program that stores
the words in a binary sort tree instead of in an arraylist. You can use the binary sort tree
routines from SortTreeDemo.java, which was discussed in Subsection 9.4.2.
3. Suppose that linked lists of integers are made from objects belonging to the class (solution)
class ListNode {
int item; // An item in the list.
ListNode next; // Pointer to the next node in the list.
}
Write a subroutine that will make a copy of a list, with the order of the items of the list
reversed. The subroutine should have a parameter of type ListNode, and it should return
a value of type ListNode. The original list should not be modified.
You should also write a main() routine to test your subroutine.
4. Subsection 9.4.1 explains how to use recursion to print out the items in a binary tree in (solution)
various orders. That section also notes that a non-recursive subroutine can be used to
print the items, provided that a stack or queue is used as an auxiliary data structure.
Assuming that a queue is used, here is an algorithm for such a subroutine:
Add the root node to an empty queue
while the queue is not empty:
Get a node from the queue
Print the item in the node
if node.left is not null:
add it to the queue
if node.right is not null:
add it to the queue
Exercises 482
Write a subroutine that implements this algorithm, and write a program to test the sub-
routine. Note that you will need a queue of TreeNodes, so you will need to write a class
to represent such queues.
(Note that the order in which items are printed by this algorithm is different from all
three of the orders considered in Subsection 9.4.1.)
5. In Subsection 9.4.2, I say that “if the [binary sort] tree is created by inserting items in a (solution)
random order, there is a high probability that the tree is approximately balanced.” For
this exercise, you will do an experiment to test whether that is true.
The depth of a node in a binary tree is the length of the path from the root of the tree
to that node. That is, the root has depth 0, its children have depth 1, its grandchildren
have depth 2, and so on. In a balanced tree, all the leaves in the tree are about the same
depth. For example, in a perfectly balanced tree with 1023 nodes, all the leaves are at
depth 9. In an approximately balanced tree with 1023 nodes, the average depth of all the
leaves should be not too much bigger than 9.
On the other hand, even if the tree is approximately balanced, there might be a few
leaves that have much larger depth than the average, so we might also want to look at the
maximum depth among all the leaves in a tree.
For this exercise, you should create a random binary sort tree with 1023 nodes. The
items in the tree can be real numbers, and you can create the tree by generating 1023
random real numbers and inserting them into the tree, using the usual treeInsert()
method for binary sort trees. Once you have the tree, you should compute and output the
average depth of all the leaves in the tree and the maximum depth of all the leaves. To
do this, you will need three recursive subroutines: one to count the leaves, one to find the
sum of the depths of all the leaves, and one to find the maximum depth. The latter two
subroutines should have an int-valued parameter, depth, that tells how deep in the tree
you’ve gone. When you call this routine from the main program, the depth parameter is
0; when you call the routine recursively, the parameter increases by 1.
6. The parsing programs in Section 9.5 work with expressions made up of numbers and (solution)
operators. We can make things a little more interesting by allowing the variable “x” to
occur. This would allow expression such as “3*(x-1)*(x+1)”, for example. Make a new
version of the sample program SimpleParser3.java that can work with such expressions.
In your program, the main() routine can’t simply print the value of the expression, since
the value of the expression now depends on the value of x. Instead, it should print the
value of the expression for x=0, x=1, x=2, and x=3.
The original program will have to be modified in several other ways. Currently, the
program uses classes ConstNode, BinOpNode, and UnaryMinusNode to represent nodes
in an expression tree. Since expressions can now include x, you will need a new class,
VariableNode, to represent an occurrence of x in the expression.
In the original program, each of the node classes has an instance method,
“double value()”, which returns the value of the node. But in your program, the
value can depend on x, so you should replace this method with one of the form
“double value(double xValue)”, where the parameter xValue is the value of x.
Finally, the parsing subroutines in your program will have to take into account the
fact that expressions can contain x. There is just one small change in the BNF rules for
the expressions: A <factor> is allowed to be the variable x:
<factor> ::= <number> | <x-variable> | "(" <expression> ")"
Exercises 483
where <x-variable> can be either a lower case or an upper case “X”. This change in the
BNF requires a change in the factorTree() subroutine.
7. This exercise builds on the previous exercise, Exercise 9.6. To understand it, you should (solution)
have some background in Calculus. The derivative of an expression that involves the
variable x can be defined by a few recursive rules:
• The derivative of a constant is 0.
• The derivative of x is 1.
• If A is an expression, let dA be the derivative of A. Then the derivative of -A is -dA.
• If A and B are expressions, let dA be the derivative of A and let dB be the derivative
of B. Then the derivative of A+B is dA+dB.
• The derivative of A-B is dA-dB.
• The derivative of A*B is A*dB + B*dA.
• The derivative of A/B is (B*dA - A*dB) / (B*B).
For this exercise, you should modify your program from the previous exercise so that
it can compute the derivative of an expression. You can do this by adding a derivative-
computing method to each of the node classes. First, add another abstract method to the
ExpNode class:
abstract ExpNode derivative();
Then implement this method in each of the four subclasses of ExpNode. All the information
that you need is in the rules given above. In your main program, instead of printing the
stack operations for the original expression, you should print out the stack operations
that define the derivative. Note that the formula that you get for the derivative can be
much more complicated than it needs to be. For example, the derivative of 3*x+1 will be
computed as (3*1+0*x)+0. This is correct, even though it’s kind of ugly, and it would be
nice for it to be simplified. However, simplifying expressions is not easy.
As an alternative to printing out stack operations, you might want to print the deriva-
tive as a fully parenthesized expression. You can do this by adding a printInfix() routine
to each node class. It would be nice to leave out unnecessary parentheses, but again, the
problem of deciding which parentheses can be left out without altering the meaning of the
expression is a fairly difficult one, which I don’t advise you to attempt.
(There is one curious thing that happens here: If you apply the rules, as given, to an
expression tree, the result is no longer a tree, since the same subexpression can occur at
multiple points in the derivative. For example, if you build a node to represent B*B by
saying “new BinOpNode(’*’,B,B)”, then the left and right children of the new node are
actually the same node! This is not allowed in a tree. However, the difference is harmless
in this case since, like a tree, the structure that you get has no loops in it. Loops, on the
other hand, would be a disaster in most of the recursive tree-processing subroutines that
we have written, since it would lead to infinite recursion. The type of structure that is
built by the derivative functions is technically referred to as a directed acyclic graph.)
Quiz 484
Quiz on Chapter 9
(answers)
3. Suppose that a linked list is formed from objects that belong to the class
class ListNode {
int item; // An item in the list.
ListNode next; // Pointer to next item in the list.
}
Write a subroutine that will count the number of zeros that occur in a given linked list
of ints. The subroutine should have a parameter of type ListNode and should return a
value of type int.
4. Let ListNode be defined as in the previous problem. Suppose that head is a variable of
type ListNode that points to the first node in a linked list. Write a code segment that
will add the number 42 in a new node at the end of the list. Assume that the list is not
empty. (There is no “tail pointer” for the list.)
7. What is an activation record ? What role does a stack of activation records play in a
computer?
8. Suppose that a binary tree of integers is formed from objects belonging to the class
class TreeNode {
int item; // One item in the tree.
TreeNode left; // Pointer to the left subtree.
TreeNode right; // Pointer to the right subtree.
}
Quiz 485
Write a recursive subroutine that will find the sum of all the nodes in the tree. Your
subroutine should have a parameter of type TreeNode, and it should return a value of
type int.
9. Let TreeNode be the same class as in the previous problem. Write a recursive subroutine
that makes a copy of a binary tree. The subroutine has a parameter that points to the
root of the tree that is to be copied. The return type is TreeNode, and the return value
should be a pointer to the root of the copy. The copy should consist of newly created
nodes, and it should have exactly the same structure as the original tree.
How to avoid reinventing the wheel? Many data structures and algorithms, such as
those from Chapter 9, have been studied, programmed, and re-programmed by generations of
computer science students. This is a valuable learning experience. Unfortunately, they have
also been programmed and re-programmed by generations of working computer professionals,
taking up time that could be devoted to new, more creative work. A programmer who needs
a list or a binary tree shouldn’t have to re-code these data structures from scratch. They are
well-understood and have been programmed thousands of times before. The problem is how to
make pre-written, robust data structures available to programmers. In this chapter, we’ll look
at Java’s attempt to address this problem.
You have already seen part of the solution in Section 7.3. That section introduced parame-
terized types and the ArrayList class. Parameterized types make it possible for the same class
to work with many different kinds of data. This idea—that the same code can be used for a
variety of data types—is called generic programming .
487
CHAPTER 10. GENERIC PROGRAMMING AND COLLECTION CLASSES 488
In Section 10.5, we will see that it is possible to define new generic classes, interfaces, and
methods. Until then, we will stick to using Java’s predefined generics. And in Section 10.6, we
will look at streams, a relatively new feature of Java that makes extensive use of generics.
It is no easy task to design a library for generic programming. Java’s solution has many
nice features but is certainly not the only possible approach. It is almost certainly not the
best, and has a few features that in my opinion can only be called bizarre, but in the context
of the overall design of Java, it might be close to optimal. To get some perspective on generic
programming in general, it might be useful to look very briefly at some other approaches to
generic programming.
thing as “ArrayList<String>”; only the non-parameterized type ArrayList exists at run time.
(However, although you can’t have an array of ArrayList<String>, you can have an ArrayList of
ArrayList<String>—with the type written as ArrayList<ArrayList<String>>—which is just as good
or better.)
Fortunately, most programmers don’t have to deal with such problems, since they turn
up only in fairly advanced programming. Most people who use parameterized types will not
encounter the problems, and they will get the benefits of type-safe generic programming with
little difficulty.
It’s worth noting that if the type parameter in a parameterized type can be deduced by the
compiler, then the name of the type parameter can be omitted. For example, the word “String”
is optional in the constructor in the following statement, because the ArrayList that is created
must be an ArrayList<String> to match the type of the variable:
ArrayList<String> words = new ArrayList<>();
∗ ∗ ∗
The interface Collection<T> specifies methods for performing some basic operations on any
collection of objects. Since “collection” is a very general concept, operations that can be applied
to all collections are also very general. They are generic operations in the sense that they can
be applied to various types of collections containing various types of objects. Suppose that
coll is an object that implements the interface Collection<T> (for some specific non-primitive
type T ). Then the following operations, which are specified in the interface Collection<T>, are
defined for coll:
• coll.size() — returns an int that gives the number of objects in the collection.
• coll.isEmpty() — returns a boolean value which is true if the size of the collection is 0.
• coll.clear() — removes all objects from the collection.
• coll.add(tobject) — adds tobject to the collection. The parameter must be of type T ;
if not, a syntax error occurs at compile time. (Remember that if T is a class, this includes
objects belonging to a subclass of T, and if T is an interface, it includes any object that
implements T.) The add() method returns a boolean value which tells you whether the
operation actually modified the collection. For example, adding an object to a Set has no
effect if that object was already in the set.
• coll.contains(object) — returns a boolean value that is true if object is in the
collection. Note that object is not required to be of type T, since it makes sense to
check whether object is in the collection, no matter what type object has. (For testing
equality, null is considered to be equal to itself. The criterion for testing non-null objects
for equality can differ from one kind of collection to another; see Subsection 10.1.6, below.)
• coll.remove(object) — removes object from the collection, if it occurs in the collection,
and returns a boolean value that tells you whether the object was found. Again, object
is not required to be of type T. The test for equality is the same test that is used by
contains().
• coll.containsAll(coll2) — returns a boolean value that is true if every object in
coll2 is also in coll. The parameter can be any collection.
• coll.addAll(coll2) — adds all the objects in coll2 to coll. The parameter, coll2,
can be any collection of type Collection<T>. However, it can also be more general. For
example, if T is a class and S is a sub-class of T, then coll2 can be of type Collection<S>.
This makes sense because any object of type S is automatically of type T and so can
legally be added to coll.
• coll.removeAll(coll2) — removes every object from coll that also occurs in the
collection coll2. coll2 can be any collection.
• coll.retainAll(coll2) — removes every object from coll that does not occur in
the collection coll2. It “retains” only the objects that do occur in coll2. coll2 can be
any collection.
• coll.toArray() — returns an array of type Object[ ] that contains all the items in the
collection. Note that the return type is Object[ ], not T[ ]! However, there is another
version of this method that takes an array of type T[ ] as a parameter: the method
coll.toArray(tarray) returns an array of type T[ ] containing all the items in the collec-
tion. If the array parameter tarray is large enough to hold the entire collection, then the
items are stored in tarray and tarray is also the return value of the collection. If tarray
is not large enough, then a new array is created to hold the items; in that case tarray
CHAPTER 10. GENERIC PROGRAMMING AND COLLECTION CLASSES 493
serves only to specify the type of the array. For example, coll.toArray(new String[0])
can be used if coll is a collection of Strings and will return a new array of type String[ ].
Since these methods are part of the Collection<T> interface, they must be defined for every
object that implements that interface. There is a problem with this, however. For example,
the size of some collections cannot be changed after they are created. Methods that add or
remove objects don’t make sense for these collections. While it is still legal to call the methods,
an exception will be thrown when the call is evaluated at run time. The type of the exception
is UnsupportedOperationException. Furthermore, since Collection<T> is only an interface, not a
concrete class, the actual implementation of the method is left to the classes that implement
the interface. This means that the semantics of the methods, as described above, are not
guaranteed to be valid for all collection objects; they are valid, however, for classes in the Java
Collection Framework.
There is also the question of efficiency. Even when an operation is defined for several types of
collections, it might not be equally efficient in all cases. Even a method as simple as size() can
vary greatly in efficiency. For some collections, computing the size() might involve counting
the items in the collection. The number of steps in this process is equal to the number of items.
Other collections might have instance variables to keep track of the size, so evaluating size()
just means returning the value of a variable. In this case, the computation takes only one step,
no matter how many items there are. When working with collections, it’s good to have some
idea of how efficient operations are and to choose a collection for which the operations that you
need can be implemented most efficiently. We’ll see specific examples of this in the next two
sections.
of type Iterator<T>, with the same type T as its type parameter. The interface Iterator<T>
defines just three methods. If iter refers to an object that implements Iterator<T>, then we
have:
• iter.next() — returns the next item, and advances the iterator. The return value is of
type T. This method lets you look at one of the items in the collection. Note that there is
no way to look at an item without advancing the iterator past that item. If this method
is called when no items remain, it will throw a NoSuchElementException.
• iter.hasNext() — returns a boolean value telling you whether there are more items to
be processed. In general, you should test this before calling iter.next().
• iter.remove() — if you call this after calling iter.next(), it will remove the item
that you just saw from the collection. Note that this method has no parameter. It
removes the item that was most recently returned by iter.next(). This might produce
an UnsupportedOperationException, if the collection does not support removal of items.
Using iterators, we can write code for printing all the items in any collection. Suppose,
for example, that coll is of type Collection<String>. In that case, the value returned by
coll.iterator() is of type Iterator<String>, and we can say:
Iterator<String> iter; // Declare the iterator variable.
iter = coll.iterator(); // Get an iterator for the collection.
while ( iter.hasNext() ) {
String item = iter.next(); // Get the next item.
System.out.println(item);
}
The same general form will work for other types of processing. For example, the following
code will remove all null values from any collection of type Collection<Color> (as long as that
collection supports removal of values):
Iterator<Color> iter = coll.iterator():
while ( iter.hasNext() ) {
Color item = iter.next();
if (item == null)
iter.remove();
}
(Note, by the way, that when Collection<T>, Iterator<T>, or any other parameterized type is
used in actual code, they are always used with actual types such as String or Color in place of
the “formal type parameter” T. An iterator of type Iterator<String> is used to iterate through
a collection of Strings; an iterator of type Iterator<Color> is used to iterate through a collection
of Color ; and so on.)
An iterator is often used to apply the same operation to all the elements in a collection. In
many cases, it’s possible to avoid the use of iterators for this purpose by using a for-each loop.
The for-each loop was discussed in Subsection 7.1.1 for use with arrays and in Subsection 7.3.3
for use with ArrayLists. But in fact, a for-each loop can be used to iterate through any
collection. For a collection coll of type Collection<T>, a for-each loop takes the form:
for ( T x : coll ) { // "for each object x, of type T, in coll"
// process x
}
CHAPTER 10. GENERIC PROGRAMMING AND COLLECTION CLASSES 495
Here, x is the loop control variable. Each object in coll will be assigned to x in turn, and the
body of the loop will be executed for each object. Since objects in coll are of type T, x is
declared to be of type T. For example, if namelist is of type Collection<String>, we can print
out all the names in the collection with:
for ( String name : namelist ) {
System.out.println( name );
}
This for-each loop could, of course, be written as a while loop using an iterator, but the for-each
loop is much easier to follow.
try {
FullName other = (FullName)obj; // Type-cast obj to type FullName
return firstName.equals(other.firstName)
&& lastName.equals(other.lastName);
}
catch (Exception e) {
return false; // if obj is null or is not of type FullName
}
}
public int compareTo( FullName other ) {
if ( lastName.compareTo(other.lastName) < 0 ) {
// If lastName comes before the last name of
// the other object, then this FullName comes
// before the other FullName. Return a negative
// value to indicate this.
return -1;
}
else if ( lastName.compareTo(other.lastName) > 0 ) {
// If lastName comes after the last name of
// the other object, then this FullName comes
// after the other FullName. Return a positive
// value to indicate this.
return 1;
}
else {
// Last names are the same, so base the comparison on
// the first names, using compareTo from class String.
return firstName.compareTo(other.firstName);
}
}
.
. // other methods
.
}
(I find it a little odd that the class here is declared as “class FullName implements
Comparable<FullName>”, with “FullName” repeated as a type parameter in the name of the
interface. However, it does make sense. It means that we are going to compare objects that
belong to the class FullName to other objects of the same type. Even though this is the only
reasonable thing to do, that fact is not obvious to the Java compiler—and the type parameter
in Comparable<FullName> is there for the compiler.)
There is another way to allow for comparison of objects in Java, and that is to provide
a separate object that is capable of making the comparison. The object must implement the
interface Comparator<T>, where T is the type of the objects that are to be compared. The
interface Comparator<T> defines the method:
public int compare( T obj1, T obj2 )
This method compares two objects of type T and returns a value that is negative, or positive,
or zero, depending on whether obj1 comes before obj2, or comes after obj2, or is considered to
be the same as obj2 for the purposes of this comparison. Comparators are useful for comparing
CHAPTER 10. GENERIC PROGRAMMING AND COLLECTION CLASSES 498
objects that do not implement the Comparable interface and for defining several different order-
ings on the same collection of objects. Since Comparator is a functional interface, comparators
are often defined by lambda expressions (see Section 4.5).
In the next two sections, we’ll see how Comparable and Comparator are used in the context
of collections and maps.
These classes are part of the Java Collection Framework. Each implements the interface List<T>,
and therefore the interface Collection<T>. An object of type ArrayList<T> represents an ordered
sequence of objects of type T, stored in an array that will grow in size whenever necessary as
new items are added. An object of type LinkedList<T> also represents an ordered sequence of
objects of type T, but the objects are stored in nodes that are linked together with pointers.
Both list classes support the basic list operations that are defined in the interface List<T>,
and an abstract data type is defined by its operations, not by its representation. So why two
classes? Why not a single List class with a single representation? The problem is that there is
no single representation of lists for which all list operations are efficient. For some operations,
linked lists are more efficient than arrays. For others, arrays are more efficient. In a particular
application of lists, it’s likely that only a few operations will be used frequently. You want
to choose the representation for which the frequently used operations will be as efficient as
possible.
Broadly speaking, the LinkedList class is more efficient in applications where items will often
be added or removed at the beginning of the list or in the middle of the list. In an array, these
operations require moving a large number of items up or down one position in the array, to
make a space for a new item or to fill in the hole left by the removal of an item. In terms of
asymptotic analysis (Section 8.5), adding an element at the beginning or in the middle of an
array has run time Θ(n), where n is the number of items in the array. In a linked list, nodes can
be added or removed at any position by changing a few pointer values, an operation that has
run time Θ(1). That is, the operation takes only some constant amount of time, independent
of how many items are in the list.
On the other hand, the ArrayList class is more efficient when random access to items is
required. Random access means accessing the k-th item in the list, for any integer k. Random
access is used when you get or change the value stored at a specified position in the list. This is
trivial for an array, with run time Θ(1). But for a linked list it means starting at the beginning
of the list and moving from node to node along the list for k steps, an operation that has run
time Θ(k).
Operations that can be done efficiently for both types of lists include sorting and adding an
item at the end of the list.
All lists implement the methods from interface Collection<T> that were discussed in Sub-
section 10.1.4. These methods include size(), isEmpty(), add(T), remove(Object), and
clear(). The add(T) method adds the object at the end of the list. The remove(Object)
method involves first finding the object, which uses linear search and is not very efficient for
any list since it involves going through the items in the list from beginning to end until the
object is found. The interface List<T> adds some methods for accessing list items according to
their numerical positions in the list. Suppose that list is an object of type List<T>. Then we
have the methods:
• list.get(index) — returns the object of type T that is at position index in the list,
where index is an integer. Items are numbered 0, 1, 2, . . . , list.size()-1. The param-
eter must be in this range, or an IndexOutOfBoundsException is thrown.
• list.set(index,obj) — stores the object obj at position number index in the list,
replacing the object that was there previously. The object obj must be of type T. This
does not change the number of elements in the list or move any of the other elements.
• list.add(index,obj) — inserts an object obj into the list at position number index,
where obj must be of type T. The number of items in the list increases by one, and items
that come after position index move down one position to make room for the new item.
CHAPTER 10. GENERIC PROGRAMMING AND COLLECTION CLASSES 500
The value of index must be in the range 0 to list.size(), inclusive. If index is equal
to list.size(), then obj is added at the end of the list.
• list.remove(index) — removes the object at position number index, and returns that
object as the return value of the method. Items after this position move up one space in
the list to fill the hole, and the size of the list decreases by one. The value of index must
be in the range 0 to list.size()-1.
• list.indexOf(obj) — returns an int that gives the position of obj in the list, if it occurs.
If it does not occur, the return value is -1. The object obj can be of any type, not just
of type T. If obj occurs more than once in the list, the index of the first occurrence is
returned.
These methods are defined both in class ArrayList<T> and in class LinkedList<T>, although
some of them—such as get and set—are only efficient for ArrayLists. The class LinkedList<T>
adds a few additional methods, which are not defined for an ArrayList. If linkedlist is an
object of type LinkedList<T>, then we have
• linkedlist.getFirst() — returns the object of type T that is the first item in the list.
The list is not modified. If the list is empty when the method is called, an exception of
type NoSuchElementException is thrown (the same is true for the next three methods as
well).
• linkedlist.getLast() — returns the object of type T that is the last item in the list.
The list is not modified.
• linkedlist.removeFirst() — removes the first item from the list, and returns that
object of type T as its return value. The functions linkedlist.remove() and
linkedlist.pop() are also defined, with the same meaning as removeFirst().
• linkedlist.removeLast() — removes the last item from the list, and returns that object
of type T as its return value.
• linkedlist.addFirst(obj) — adds the obj, which must be of type T, to the beginning
of the list. The function linkedlist.push(obj) has the same meaning.
• linkedlist.addLast(obj) — adds the object obj, which must be of type T, to the end
of the list. This is exactly the same as linkedlist.add(obj) but is defined to keep the
naming consistent.
There is some redundancy here, apparently to make it easy to use a LinkedList as if it were
a stack or a queue. (See Section 9.3.) For example, we can use a LinkedList as a stack by
using the methods named push() and pop(), or as a queue by using add() and remove() to
implement the enqueue and dequeue operations.
If list is an object of type List<T>, then the method list.iterator(), defined in the
interface Collection<T>, returns an Iterator that can be used to traverse the list from beginning
to end. However, for Lists, there is a special type of Iterator, called a ListIterator, which offers
additional capabilities. ListIterator<T> is an interface that extends the interface Iterator<T>.
The method list.listIterator() returns an object of type ListIterator<T>.
A ListIterator has the usual Iterator methods, hasNext(), next(), and remove(), but it
also has methods hasPrevious(), previous(), add(obj), and set(obj) that make it possible
to move backwards in the list, to add an item at the current position of the iterator, and to
replace one of the items in the list. To understand how these work, it’s best to think of an
iterator as pointing to a position between two list elements, or at the beginning or end of the
CHAPTER 10. GENERIC PROGRAMMING AND COLLECTION CLASSES 501
list. In this diagram, the items in a list are represented by squares, and arrows indicate the
possible positions of an iterator:
If iter is of type ListIterator<T>, then iter.next() moves the iterator one space to the
right along the list and returns the item that the iterator passes as it moves. The method
iter.previous() moves the iterator one space to the left along the list and returns the item
that it passes. The method iter.remove() removes an item from the list; the item that is
removed is the item that the iterator passed most recently in a call to either iter.next()
or iter.previous(). The method iter.set(obj) works similarly; it replaces the item that
would be removed by iter.remove(). There is also a method iter.add(obj) that adds the
specified object to the list at the current position of the iterator (where obj must be of type
T ). This can be between two existing items or at the beginning of the list or at the end of the
list.
(By the way, the lists that are used in class LinkedList<T> are doubly linked lists.
That is, each node in the list contains two pointers—one to the next node in the list and
one to the previous node. This makes it possible to efficiently implement both the next()
and previous() methods of a ListIterator. Also, to make the addLast() and getLast()
methods of a LinkedList efficient, the class LinkedList<T> includes a “tail pointer” that points
to the last node in the list.)
As an example of using a ListIterator, suppose that we want to maintain a list of items
that is always sorted into increasing order. When adding an item to the list, we can use a
ListIterator to find the position in the list where the item should be added. Once the position
has been found, we use the same list iterator to place the item in that position. The idea is to
start at the beginning of the list and to move the iterator forward past all the items that are
smaller than the item that is being inserted. At that point, the iterator’s add() method can
be used to insert the item. To be more definite, suppose that stringList is a variable of type
List<String>. Assume that the strings that are already in the list are stored in ascending order
and that newItem is a string that we would like to insert into the list. The following code will
place newItem in the list in its correct position, so that the modified list is still in ascending
order:
ListIterator<String> iter = stringList.listIterator();
// Move the iterator so that it points to the position where
// newItem should be inserted into the list. If newItem is
// bigger than all the items in the list, then the while loop
// will end when iter.hasNext() becomes false, that is, when
// the iterator has reached the end of the list.
while (iter.hasNext()) {
String item = iter.next();
if (newItem.compareTo(item) <= 0) {
// newItem should come BEFORE item in the list.
// Move the iterator back one space so that
// it points to the correct insertion point,
// and end the loop.
iter.previous();
CHAPTER 10. GENERIC PROGRAMMING AND COLLECTION CLASSES 502
break;
}
}
iter.add(newItem);
Here, stringList might be of type ArrayList<String> or of type LinkedList<String>. The
algorithm that is used to insert newItem into the list will be about equally efficient for both types
of lists, and it will even work for other classes that implement the interface List<String>. You
would probably find it easier to design an insertion algorithm that uses array-like indexing with
the methods get(index) and add(index,obj). However, that algorithm would be horribly
inefficient for LinkedLists because random access is so inefficient for linked lists. (By the way,
the insertion algorithm works when the list is empty. It might be useful for you to think about
why this is true.)
10.2.2 Sorting
Sorting a list is a fairly common operation, and there should really be a sorting method in
the List interface. There is not, presumably because it only makes sense to sort lists of certain
types of objects. However, methods for sorting lists are available as static methods in the class
java.util.Collections. This class contains a variety of static utility methods for working
with collections. The methods are generic; that is, they will work for collections of objects of
various types. (You have already seen similar methods for arrays in the Arrays class.) Suppose
that list is of type List<T>. The command
Collections.sort(list);
can be used to sort the list into ascending order. The items in the list should implement the
interface Comparable<T> (see Subsection 10.1.6). The method Collections.sort() will work,
for example, for lists of String and for lists of any of the wrapper classes such as Integer and
Double. There is also a sorting method that takes a Comparator as its second argument:
Collections.sort(list,comparator);
In this method, the comparator will be used to compare the items in the list. As mentioned
in the previous section, a Comparator is an object that defines a compare() method that can
be used to compare two objects. We’ll see an example of using a Comparator in Section 10.4.
The sorting method that is used by Collections.sort() is the so-called “merge sort”
algorithm, which has both worst-case and average-case run times that are Θ(n*log(n)) for
a list of size n. Although the average run time for MergeSort is a little slower than that of
QuickSort, its worst-case performance is much better than QuickSort’s. (QuickSort was covered
in Subsection 9.1.3.) MergeSort also has a nice property called “stability” that we will encounter
at the end of Subsection 10.4.3.
The Collections class has at least two other useful methods for modifying lists.
Collections.shuffle(list) will rearrange the elements of the list into a random order.
Collections.reverse(list) will reverse the order of the elements, so that the last element is
moved to the beginning of the list, the next-to-last element to the second position, and so on.
Since an efficient sorting method is provided for Lists, there is no need to write one yourself.
occurs twice in the set. For example, if set is an object of type Set<T>, then set.add(obj)
will have no effect on the set if obj is already an element of the set. Java has two classes that
implement the interface Set<T>: java.util.TreeSet and java.util.HashSet.
In addition to being a Set, a TreeSet has the property that the elements of the set are
arranged into ascending sorted order. An Iterator (or a for-each loop) for a TreeSet will always
visit the elements of the set in ascending order.
A TreeSet cannot hold arbitrary objects, since there must be a way to determine the sorted
order of the objects it contains. Ordinarily, this means that the objects in a set of type
TreeSet<T> should implement the interface Comparable<T> and that obj1.compareTo(obj2)
should be defined in a reasonable way for any two objects obj1 and obj2 in the set. Alter-
natively, an object of type Comparator<T> can be provided as a parameter to the constructor
when the TreeSet is created. In that case, the compare() method of the Comparator will be
used to compare objects that are added to the set.
A TreeSet does not use the equals() method to test whether two objects are the same.
Instead, it uses the compareTo() (or compare()) method. This can be a problem. Recall from
Subsection 10.1.6 that compareTo() can consider two objects to be the same for the purpose of
the comparison even though the objects are not equal. For a TreeSet, this means that only one
of those objects can be in the set. For example, if the TreeSet contains mailing addresses and if
the compareTo() method for addresses just compares their zip codes, then the set can contain
only one address in each zip code. Clearly, this is not right! But that only means that you
have to be aware of the semantics of TreeSets, and you need to make sure that compareTo()
is defined in a reasonable way for objects that you put into a TreeSet. This will be true, by
the way, for Strings, Integers, and many other built-in types, since the compareTo() method for
these types considers two objects to be the same only if they are actually equal.
In the implementation of a TreeSet, the elements are stored in something similar to a binary
sort tree. (See Subsection 9.4.2.) However, the data structure that is used is balanced in the
sense that all the leaves of the tree are at about the same distance from the root of the tree.
This ensures that all the basic operations—inserting, deleting, and searching—are efficient, with
worst-case run time Θ(log(n)), where n is the number of items in the set.
The fact that a TreeSet sorts its elements and removes duplicates makes it very useful in
some applications. Exercise 7.6 asked you to write a program that would read a file and output
an alphabetical list of all the words that occurred in the file, with duplicates removed. The
words were to be stored in an ArrayList, so it was up to you to make sure that the list was
sorted and contained no duplicates. The same task can be programmed much more easily
using a TreeSet instead of a list. A TreeSet automatically eliminates duplicates, and an iterator
for the set will automatically visit the items in the set in sorted order. An algorithm for the
program, using a TreeSet, would be:
TreeSet<String> words = new TreeSet<String>();
while there is more data in the input file:
Let word = the next word from the file
Convert word to lower case
words.add(word) // Adds the word only if not already present.
for ( String w : words ) // for each String w in words
Output w // words are output in sorted order
If you would like to see a complete, working program, you can find it in the file WordListWith-
TreeSet.java.
CHAPTER 10. GENERIC PROGRAMMING AND COLLECTION CLASSES 504
As another example, suppose that coll is any Collection of Strings. (This would also
work for any other type for which compareTo() is properly defined.) We can use a TreeSet to
sort the items of coll and remove the duplicates simply by saying:
TreeSet<String> set = new TreeSet<String>();
set.addAll(coll);
The second statement adds all the elements of the collection to the set. Since it’s a Set,
duplicates are ignored. Since it’s a TreeSet, the elements of the set are sorted. If you would like
to have the data in some other type of data structure, it’s easy to copy the data from the set.
For example, to place the answer in an ArrayList, you could say:
TreeSet<String> set = new TreeSet<String>();
set.addAll(coll);
ArrayList<String> list = new ArrayList<String>();
list.addAll(set);
Now, in fact, every one of Java’s collection classes has a constructor that takes a Collection
as an argument. All the items in that Collection are added to the new collection when it is
created. So, if coll is of type Collection<String>, then “new TreeSet<String>(coll)” creates
a TreeSet that contains the same elements as coll, but with duplicates removed and in sorted
order. This means that we can abbreviate the four lines in the above example to the single
command:
ArrayList<String> list = new ArrayList<>( new TreeSet<>(coll) );
This makes a sorted list of the elements of coll with no duplicates. This is a nice example of
the power of generic programming. (Note that the type parameter, String, is optional in the
two constructors in this example, since it can be deduced by the compiler.)
∗ ∗ ∗
A HashSet stores its elements in a hash table, a type of data structure that I will discuss
in the next section. The operations of finding, adding, and removing elements are implemented
very efficiently in hash tables, even more so than for TreeSets. The elements of a HashSet are
not stored in any particular order, and so do not need to implement the Comparable interface.
(They do, however, need to define a proper “hash code,” as we’ll see in the next section.)
The equals() method is used to determine whether two objects in a HashSet are to be
considered the same. An Iterator for a HashSet will visit its elements in what seems to be a
completely arbitrary order, and it’s even possible for the order to change when a new element
is added. Use a HashSet instead of a TreeSet when the elements it contains are not comparable,
or when the order is not important, or when the small advantage in efficiency is important.
∗ ∗ ∗
A note about the mathematics of sets: In mathematical set theory, the items in a set are
called members or elements of that set. Important operations include adding an element
to a set, removing an element from a set, and testing whether a given entity is an element of
a set. Operations that can be performed on two sets include union, intersection, and set
difference. All these operations are defined in Java for objects of type Set, but with different
names. Suppose that A and B are Sets. Then:
• A.add(x) adds the element x to the set A.
• A.remove(x) removes the element x from the set A.
• A.contains(x) tests whether x is an element of the set A.
CHAPTER 10. GENERIC PROGRAMMING AND COLLECTION CLASSES 505
can be used in a similar way to sort a collection without removing duplicates. For example, if
coll is of type Collection<String>, then the following code segment will print all the items from
coll in order, including duplicates:
PriorityQueue<String> pq = new PriorityQueue<>();
pq.addAll( coll );
while ( ! pq.isEmpty() ) {
System.out.println( pq.remove() );
}
(Note, by the way, that we can’t use an iterator or a for-each loop to print the items in this
example, since iterators and for-each loops do not traverse a priority queue in ascending order.)
The sample program WordListWithPriorityQueue.java makes a sorted list of words from
a file without removing duplicates, using a priority queue to hold the words. It is a minor
modification of WordListWithTreeSet.java.
Although priority queues can be used for sorting, they have other natural applications. For
example, consider the problem of scheduling “jobs” to be executed on a computer, where each
job is assigned a priority and jobs with lower priority should always be executed before jobs
with higher priority. Jobs can be placed into a priority queue as they are created. When the
computer removes jobs from the queue for execution, they will be removed in order of increasing
priority.
10.3 Maps
An array of N elements can be thought of as a way of associating some item with each of
the integers 0, 1, . . . , N-1. If i is one of these integers, it’s possible to get the item associated
with i, and it’s possible to put a new item in the i-th position. These “get” and “put”
operations define what it means to be an array.
A map is a kind of generalized array. Like an array, a map is defined by “get” and “put”
operations. But in a map, these operations are defined not for integers 0, 1, . . . , N-1, but for
arbitrary objects of some specified type T. Associated to these objects of type T are objects of
some possibly different type S.
In fact, some programming languages use the term associative array instead of “map”
and use the same notation for associative arrays as for regular arrays. In those languages,
for example, you might see the notation A["fred"] used to indicate the item associated to the
string “fred” in an associative array A. Java does not use array notation for maps, unfortunately,
but the idea is the same: A map is like an array, but the indices for a map are objects, not
integers. In a map, an object that serves as an “index” is called a key . The object that is
associated with a key is called a value. Note that a key can have at most one associated value,
but the same value can be associated to several different keys. A map can be considered to be
a set of “associations,” where each association is a key/value pair.
values of type Button to keys of type Date. For a map of type Map<String,String>, both the
keys and the values are of type String.
Suppose that map is a variable of type Map<K,V> for some specific types K and V. Then the
following are some of the methods that are defined for map:
• map.get(key) — returns the object of type V that is associated by the map to the key.
If the map does not associate any value with key, then the return value is null. Note that
it’s also possible for the return value to be null when the map explicitly associates the
value null with the key. Referring to “map.get(key)” is similar to referring to “A[key]”
for an array A. (But note that there is nothing like an IndexOutOfBoundsException for
maps.)
• map.put(key,value) — Associates the specified value with the specified key, where key
must be of type K and value must be of type V. If the map already associated some
other value with the key, then the new value replaces the old one. This is similar to the
command “A[key] = value” for an array.
• map.putAll(map2) — if map2 is another map of type Map<K,V>, this copies all the asso-
ciations from map2 into map.
• map.remove(key) — if map associates a value to the specified key, that association is
removed from the map.
• map.containsKey(key) — returns a boolean value that is true if the map associates some
value to the specified key.
• map.containsValue(value) — returns a boolean value that is true if the map associates
the specified value to some key.
• map.size() — returns an int that gives the number of key/value associations in the map.
• map.isEmpty() — returns a boolean value that is true if the map is empty, that is if it
contains no associations.
• map.clear() — removes all associations from the map, leaving it empty.
The put and get methods are certainly the most commonly used of the methods in the Map
interface. In many applications, these are the only methods that are needed, and in such cases
a map is really no more difficult to use than a standard array.
Java includes two classes that implement the interface Map<K,V>: TreeMap<K,V> and
HashMap<K,V>. In a TreeMap, the key/value associations are stored in a sorted tree, in which
they are sorted according to their keys. For this to work, it must be possible to compare the
keys to one another. This means either that the keys must implement the interface Compa-
rable<K>, or that a Comparator must be provided for comparing keys. (The Comparator can
be provided as a parameter to the TreeMap constructor.) Note that in a TreeMap, as in a
TreeSet, the compareTo() (or compare()) method is used to decide whether two keys are to
be considered the same. This can have undesirable consequences if the compareTo() method
does not agree with the usual notion of equality, and you should keep this in mind when using
TreeMaps.
A HashMap does not store associations in any particular order, so the keys that can be used
in a HashMap do not have to be comparable. However, the key class should have reasonable
definitions for the equals() method and for a hashCode() method that is discussed later in
this section; most of Java’s standard classes define these methods correctly. Most operations are
a little faster on HashMaps than they are on TreeMaps. In general, you should use a HashMap
CHAPTER 10. GENERIC PROGRAMMING AND COLLECTION CLASSES 508
unless you have some particular need for the ordering property of a TreeMap. In particular, if
you are only using the put and get operations, you can safely use a HashMap.
Let’s consider an example where maps would be useful. In Subsection 7.4.2, I presented
a simple PhoneDirectory class that associated phone numbers with names. That class defined
operations addEntry(name,number) and getNumber(name), where both name and number are
given as Strings. In fact, the phone directory is acting just like a map, with the addEntry
method playing the role of the put operation and getNumber playing the role of get. In a real
programming application, there would be no need to define a new class; we could simply use a
map of type Map<String,String>. A directory could be defined as
Map<String,String> directory = new TreeMap<>();
(using TreeMap so that the entries are kept in sorted order by name). Then
directory.put(name,number) would record a phone number in the directory and
directory.get(name) would retrieve the phone number associated with a given name.
This is certainly a place where it would be convenient to use var to declare the variables
(See Subsection 4.8.2). With var, the example using an iterator becomes:
var entries = map.entrySet();
var entryIter = entries.iterator();
System.out.println("The map contains the following associations:");
while (entryIter.hasNext()) { . . .
(But note that this syntax requires Java 10 or later.)
∗ ∗ ∗
Maps are not the only place in Java’s generic programming framework where views are
used. For example, the interface List<T> defines a sublist as a view of a part of a list. If list
implements the interface List<T>, then the method
list.subList( fromIndex, toIndex )
where fromIndex and toIndex are integers, returns a view of the part of the list consisting
of the list elements in positions between fromIndex and toIndex (including fromIndex but
excluding toIndex). This view lets you operate on the sublist using any of the operations
defined for lists, but the sublist is not an independent list. Changes made to the sublist are
actually made to the original list.
Similarly, it is possible to obtain views that represent certain subsets of a sorted set. If
set is of type TreeSet<T>, then set.subSet(fromElement,toElement) returns a Set<T> that
contains all the elements of set that are between fromElement and toElement (including
fromElement and excluding toElement). The parameters fromElement and toElement must
be objects of type T. For example, if words is a set of type TreeSet<String> in which all the
elements are strings of lower case letters, then words.subSet("m","n") contains all the ele-
ments of words that begin with the letter ’m’. This subset is a view of part of the original
set. That is, creating the subset does not involve copying elements. And changes made to
the subset, such as adding or removing elements, are actually made to the original set. The
view set.headSet(toElement) consists of all elements from the set which are strictly less than
toElement, and set.tailSet(fromElement) is a view that contains all elements from the set
that are greater than or equal to fromElement.
The class TreeMap<K,V> defines three submap views. A submap is similar to a subset.
A submap is a Map that contains a subset of the keys from the original Map, along with
their associated values. If map is a variable of type TreeMap<K,V>, and if fromKey and toKey
are of type K, then map.subMap(fromKey,toKey) returns a view that contains all key/value
pairs from map whose keys are between fromKey and toKey (including fromKey and excluding
toKey). There are also views map.headMap(toKey) and map.tailMap(fromKey) which are
defined analogously to headSet and tailSet. Suppose, for example, that phoneBook is a map
of type TreeMap<String,String> in which the keys are names and the values are phone numbers.
We can print out all the entries from phoneBook where the name begins with “M” as follows:
Map<String,String> ems = phoneBook.subMap("M","N");
// This submap contains entries for which the key is greater
// than or equal to "M" and strictly less than "N".
if (ems.isEmpty()) {
System.out.println("No entries beginning with M.");
}
else {
System.out.println("Entries beginning with M:");
CHAPTER 10. GENERIC PROGRAMMING AND COLLECTION CLASSES 511
A collision is not an error. We can’t reject a key just because another key happened to have
the same hash code. A hash table must be able to handle collisions in some reasonable way. In
the type of hash table that is used in Java, each array location actually holds a linked list of
key/value pairs (possibly an empty list). When two items have the same hash code, they are
in the same linked list. The structure of the hash table looks something like this:
0 item item
1
2 item
3
4
5 item item item
6
7
8 item
9 item
10
11
.. item item
.
..
.
In this diagram, there are two items with hash code 0, no items with hash code 1, one item
with hash code 2, and so on. In a properly designed hash table, most of the linked lists are of
length zero or one, and the average length of the lists is less than one. Although the hash code
of a key doesn’t necessarily take you directly to that key, there are probably no more than one
or two other items that you have to look through before finding the key you want. For this to
work properly, the number of items in the hash table should be somewhat less than the number
of locations in the array. In Java’s implementation, whenever the number of items exceeds 75%
of the array size, the array is replaced by a larger one and all the items in the old array are
inserted into the new one. (This is why adding one new item will sometimes cause the ordering
of all the items in the hash table to change completely.)
There is still the question of where hash codes come from. Every object in Java has a
hash code. The Object class defines the method hashCode(), which returns a value of type
int. When an object, obj, is stored in a hash table that has N locations, a hash code in the
range 0 to N-1 is needed. This hash code is computed as Math.abs(obj.hashCode()) % N,
the remainder when the absolute value of obj.hashCode() is divided by N. (The Math.abs
is necessary because obj.hashCode() can be a negative integer, and we need a non-negative
number to use as an array index.)
For hashing to work properly, two objects that are equal according to the equals() method
must have the same hash code. In the Object class, this condition is satisfied because both
equals() and hashCode() are based on the address of the memory location where the object
is stored. However, as noted in Subsection 10.1.6, many classes redefine the equals() method.
If a class redefines the equals() method, and if objects of that class will be used as keys in
hash tables, then the class must also redefine the hashCode() method. For example, in the
String class, the equals() method is redefined so that two objects of type String are considered
to be equal if they contain the same sequence of characters. The hashCode() method is also
redefined in the String class, so that the hash code of a string is computed from the characters
in that string rather than from its location in memory. For Java’s standard classes, you can
expect equals() and hashCode() to be correctly defined. However, you might need to define
these methods in classes that you write yourself.
CHAPTER 10. GENERIC PROGRAMMING AND COLLECTION CLASSES 513
Writing a good hash function is something of an art. In order to work well, the hash function
must spread the possible keys fairly evenly over the hash table. Otherwise, the items in a table
can be concentrated in a subset of the available locations, and the linked lists at those locations
can grow to a large size; that would destroy the efficiency that is the major reason for hash
tables to exist in the first place. However, I won’t cover techniques for creating good hash
functions in this book.
program. The Java program is executing a sort of program typed in by the user. I am talking
about variables in the user’s program. The user gets to make up variable names, so there is no
way for the Java program to know in advance what the variables will be.)
In Subsection 9.5.2, we saw how to write a program, SimpleParser2.java, that can evaluate
expressions that do not contain variables. Here, I will discuss another example program, Sim-
pleInterpreter.java, that is based on the older program. I will only talk about the parts that
are relevant to the symbol table.
The program uses a HashMap as the symbol table. A TreeMap could also be used, but since
the program does not need to access the variables in alphabetical order, we don’t need to have
the keys stored in sorted order. The symbol table in the program is represented by a variable
named symbolTable of type HashMap<String,Double>. At the beginning of the program, the
symbol table object is created with the command:
symbolTable = new HashMap<>();
This creates a map that initially contains no key/value associations. To execute a “let” com-
mand, the program uses the symbol table’s put() method to associate a value with the variable
name. Suppose that the name of the variable is given by a String, varName, and the value of
the variable is stored in a variable, val, of type double. The following command would then
set the value associated with the variable in the symbol table:
symbolTable.put( varName, val );
In the program SimpleInterpreter.java, you’ll find this in the method named doLetCommand().
The actual value that is stored in the symbol table is an object of type Double. We can use the
double value val in the call to put because Java does an automatic conversion of type double
to Double when necessary.
Just for fun, I decided to pre-define two variables named “pi” and “e” whose values are
the usual mathematical constants π and e. In Java, the values of these constants are given by
Math.PI and Math.E. To make these variables available to the user of the program, they are
added to the symbol table with the commands:
symbolTable.put( "pi", Math.PI );
symbolTable.put( "e", Math.E );
When the program encounters a variable while evaluating an expression, the symbol table’s
get() method is used to retrieve its value. The function symbolTable.get(varName) returns
a value of type Double. It is possible that the return value is null; this will happen if no value
has ever been assigned to varName in the symbol table. It’s important to check this possibility.
It indicates that the user is trying to use a variable that the user has not defined. The program
considers this to be an error, so the processing looks something like this:
Double val = symbolTable.get(varName);
if (val == null) {
... // Throw an exception: Undefined variable.
}
// The value associated to varName is val.doubleValue()
You will find this code, more or less, in a method named primaryValue() in SimpleInter-
preter.java.
As you can see from this example, Maps are very useful and are really quite easy to use.
CHAPTER 10. GENERIC PROGRAMMING AND COLLECTION CLASSES 515
have a set of page references, and we should just add the new page reference to the set. Here
is a subroutine that does this:
/**
* Add a page reference to the index.
*/
void addReference(String term, int pageNum) {
TreeSet<Integer> references; // The set of page references that we
// have so far for the term.
references = index.get(term);
if (references == null){
// This is the first reference that we have
// found for the term. Make a new set containing
// the page number and add it to the index, with
// the term as the key.
TreeSet<Integer> firstRef = new TreeSet<>();
firstRef.add( pageNum ); // pageNum is "autoboxed" to give an Integer!
index.put(term,firstRef);
}
else {
// references is the set of page references
// that we have found previously for the term.
// Add the new page number to that set. This
// set is already associated to term in the index.
references.add( pageNum );
}
}
The only other thing we need to do with the index is print it out. We want to iterate
through the index and print out each term, together with the set of page references for that
term. We could use an Iterator to iterate through the index, but it’s much easier to do it with
a for-each loop. The loop will iterate through the entry set of the map (see Subsection 10.3.2).
Each “entry” is a key/value pair from the map; the key is a term and the value is the associated
set of page references. Inside the for-each loop, we will have to print out a set of Integers, which
can also be done with a for-each loop. So, here we have an example of nested for-each loops.
(You might try to do the same thing entirely with iterators; doing so should give you some
appreciation for the for-each loop!) Here is a subroutine that will print the index:
/**
* Print each entry in the index.
*/
void printIndex() {
for ( Map.Entry<String,TreeSet<Integer>> entry : index.entrySet() ) {
String term = entry.getKey();
TreeSet<Integer> pageSet = entry.getValue();
System.out.print( term + ": " );
for ( int page : pageSet ) {
System.out.print( page + " " );
}
System.out.println();
}
CHAPTER 10. GENERIC PROGRAMMING AND COLLECTION CLASSES 517
}
The hardest thing here is the name of the type Map.Entry<String,TreeSet<Integer>>! Re-
member that the entries in a map of type Map<K,V> have type Map.Entry<K,V>, so the type
parameters in Map.Entry<String,TreeSet<Integer>> are simply copied from the declaration
of index. Another thing to note is that I used a loop control variable, page, of type int to
iterate through the elements of pageSet, which is of type TreeSet<Integer>. You might have
expected page to be of type Integer, not int, and in fact Integer would have worked just as well
here. However, int does work, because of automatic type conversion: It’s legal to assign a value
of type Integer to a variable of type int. (To be honest, I was sort of surprised that this worked
when I first tried it!)
This is not a lot of code, considering the complexity of the operations. I have not written
a complete indexing program, but Exercise 10.6 presents a problem that is almost identical to
the indexing problem.
∗ ∗ ∗
By the way, in this example, I would prefer to print each list of page references with the
integers separated by commas. In the printIndex() method given above, they are separated
by spaces. There is an extra space after the last page reference in the list, but it does no harm
since it’s invisible in the printout. An extra comma at the end of the list would be annoying.
The lists should be in a form such as “17,42,105” and not “17,42,105,”. The problem is, how
to leave that last comma out. Unfortunately, this is not so easy to do with a for-each loop. It
might be fun to look at a few ways to solve this problem. One alternative is to use an iterator:
Iterator<Integer> iter = pageSet.iterator();
int firstPage = iter.next(); // In this program, we know the set has
// at least one element.
System.out.print(firstPage);
while ( iter.hasNext() ) {
int nextPage = iter.next();
System.out.print("," + nextPage);
}
Another possibility is to use the fact that the TreeSet class defines a method first() that
returns the first item in the set, that is, the one that is smallest in terms of the ordering that
is used to compare items in the set. (It also defines the method last().) We can solve our
problem using this method and a for-each loop:
int firstPage = pageSet.first(); // Find out the first page number in the set.
for ( int page : pageSet ) {
if ( page != firstPage )
System.out.print(","); // Output comma only if this is not the first page.
System.out.print(page);
}
Finally, here is an elegant solution using a subset view of the tree. (See Subsection 10.3.2.)
Actually, this solution might be a bit extreme:
int firstPage = pageSet.first(); // Get first item, which we know exists.
System.out.print(firstPage); // Print first item, with no comma.
for ( int page : pageSet.tailSet( firstPage+1 ) ) // Process remaining items.
System.out.print( "," + page );
CHAPTER 10. GENERIC PROGRAMMING AND COLLECTION CLASSES 518
each word occurs. The file will be selected by the user. The output of the program will consist
of two lists. Each list contains all the words from the file, along with the number of times that
the word occurred. One list is sorted alphabetically, and the other is sorted according to the
number of occurrences, with the most common words at the top and the least common at the
bottom. The problem here is a generalization of Exercise 7.6, which asked you to make an
alphabetical list of all the words in a file, without counting the number of occurrences.
My word counting program can be found in the file WordCount.java. As the program reads
an input file, it must keep track of how many times it encounters each word. We could simply
throw all the words, with duplicates, into a list and count them later. But that would require
a lot of extra storage space and would not be very efficient. A better method is to keep a
counter for each word. The first time the word is encountered, the counter is initialized to 1.
On subsequent encounters, the counter is incremented. To keep track of the data for one word,
the program uses a simple class that holds a word and the counter for that word. The class is
a static nested class:
/**
* Represents the data we need about a word: the word and
* the number of times it has been encountered.
*/
private static class WordData {
String word;
int count;
WordData(String w) {
// Constructor for creating a WordData object when
// we encounter a new word.
word = w;
count = 1; // The initial value of count is 1.
}
} // end class WordData
The program has to store all the WordData objects in some sort of data structure. We want
to be able to add new words efficiently. Given a word, we need to check whether a WordData
object already exists for that word, and if it does, we need to find that object so that we can
increment its counter. A Map can be used to implement these operations. Given a word, we
want to look up a WordData object in the Map. This means that the word is the key, and
the WordData object is the value. (It might seem strange that the key is also one of the
instance variables in the value object, but in fact this is a very common situation: The value
object contains all the information about some entity, and the key is one of those pieces of
information; the partial information in the key is used to retrieve the full information in the
value object.) After reading the file, we want to output the words in alphabetical order, so we
should use a TreeMap rather than a HashMap. This program converts all words to lower case
so that the default ordering on Strings will put the words in alphabetical order. The data is
stored in a variable named words of type TreeMap<String,WordData>. The variable is declared
and the map object is created with the statement:
TreeMap<String,WordData> words = new TreeMap<>();
When the program reads a word from a file, it calls words.get(word) to find out if that
word is already in the map. If the return value is null, then this is the first time the word
has been encountered, so a new WordData object is created and inserted into the map with the
command words.put(word, new WordData(word)). If words.get(word) is not null, then
CHAPTER 10. GENERIC PROGRAMMING AND COLLECTION CLASSES 520
its value is the WordData object for this word, and the program only has to increment the
counter in that object. The program uses a method readNextWord(), which was given in
Exercise 7.6, to read one word from the file. This method returns null when the end of the file
is encountered. Here is the complete code segment that reads the file and collects the data:
String word = readNextWord();
while (word != null) {
word = word.toLowerCase(); // convert word to lower case
WordData data = words.get(word);
if (data == null)
words.put( word, new WordData(word) );
else
data.count++;
word = readNextWord();
}
After reading the words and printing them out in alphabetical order, the program has
to sort the words by frequency and print them again. To do the sorting using a generic
algorithm, we can copy the WordData objects into a list and then use the generic method
Collections.sort(list,comparator), which specifies a comparator as its second parameter.
The comparator can be given as a lambda expression.
The WordData objects that we need are the values in the map, words. Recall that
words.values() returns a Collection that contains all the values from the map. The con-
structor for the ArrayList class lets you specify a collection to be copied into the list when it is
created. So, we can use the following commands to create a list of type ArrayList<WordData>
containing the word data and then sort that list according to frequency:
ArrayList<WordData> wordsByFrequency = new ArrayList<>( words.values() );
Collections.sort( wordsByFrequency, (a,b) -> b.count - a.count );
You should notice that these two lines replace a lot of code! It requires some practice to think
in terms of generic data structures and algorithms, but the payoff is significant in terms of saved
time and effort.
The only remaining problem is to print the data. We have to print the data from all the
WordData objects twice, first in alphabetical order and then sorted according to frequency
count. The data is in alphabetical order in the TreeMap, or more precisely, in the values of
the TreeMap. We can use a for-each loop to print the data in the collection words.values(),
and the words will appear in alphabetical order. Another for-each loop can be used to print
the data in the list wordsByFrequency, and the words will be printed in order of decreasing
frequency. Here is the code that does it:
TextIO.putln("List of words in alphabetical order"
+ " (with counts in parentheses):\n");
for ( WordData data : words.values() )
TextIO.putln(" " + data.word + " (" + data.count + ")");
TextIO.putln("\n\nList of words by frequency of occurrence:\n");
for ( WordData data : wordsByFrequency )
TextIO.putln(" " + data.word + " (" + data.count + ")");
You can find the complete word-counting program in the file WordCount.java. Note that for
reading and writing files, it uses the file I/O capabilities of TextIO.java, which were discussed
in Subsection 2.4.4.
CHAPTER 10. GENERIC PROGRAMMING AND COLLECTION CLASSES 521
By the way, if you run the WordCount program on a reasonably large file and take a look
at the output, it will illustrate something about the Collections.sort() method. The second
list of words in the output is ordered by frequency, but if you look at a group of words that all
have the same frequency, you will see that the words in that group are in alphabetical order.
The method Collections.sort() was applied to sort the words by frequency, but before it was
applied, the words were already in alphabetical order. When Collections.sort() rearranged
the words, it did not change the ordering of words that have the same frequency, so they were
still in alphabetical order within the group of words with that frequency. This is because the
algorithm used by Collections.sort() is a stable sorting algorithm. A sorting algorithm is
said to be stable if it satisfies the following condition: When the algorithm is used to sort a list
according to some property of the items in the list, then the sort does not change the relative
order of items that have the same value of that property. That is, if item B comes after item A
in the list before the sort, and if both items have the same value for the property that is being
used as the basis for sorting, then item B will still come after item A after the sorting has
been done. Neither SelectionSort nor QuickSort are stable sorting algorithms. Insertion sort is
stable, but is not very fast. Merge sort, the sorting algorithm used by Collections.sort(),
is both stable and fast.
I hope that the programming examples in this section have convinced you of the usefulness
of the Java Collection Framework!
return items.removeFirst();
}
public boolean isEmpty() {
return (items.size() == 0);
}
}
This is a fine and useful class. But, if this is how we write queue classes, and if we want
queues of Integers or Doubles or Colors or any other type, then we will have to write a different
class for each type. The code for all of these classes will be almost identical, which seems like a
lot of redundant programming. To avoid the redundancy, we can write a generic Queue class
that can be used to define queues of any type of object.
The syntax for writing the generic class is straightforward: We replace the specific type
String with a type parameter such as T, and we add the type parameter to the name of the
class:
class Queue<T> {
private LinkedList<T> items = new LinkedList<>();
public void enqueue(T item) {
items.addLast(item);
}
public T dequeue() {
return items.removeFirst();
}
public boolean isEmpty() {
return (items.size() == 0);
}
}
Note that within the class, the type parameter T is used just like any regular type name.
It’s used to declare the return type for dequeue, as the type of the formal parameter item in
enqueue, and even as the actual type parameter in LinkedList<T>. Given this class definition,
we can use parameterized types such as Queue<String> and Queue<Integer> and Queue<Color>.
That is, the Queue class is used in exactly the same way as built-in generic classes like LinkedList
and HashSet.
Note that you don’t have to use “T” as the name of the type parameter in the definition of
the generic class. Type parameters are like formal parameters in subroutines. You can make up
any name you like in the definition of the class. The name in the definition will be replaced by
an actual type name when the class is used to declare variables or create objects. If you prefer
to use a more meaningful name for the type parameter, you might define the Queue class as:
class Queue<ItemType> {
private LinkedList<ItemType> items = new LinkedList<>();
public void enqueue(ItemType item) {
items.addLast(item);
}
public ItemType dequeue() {
return items.removeFirst();
}
public boolean isEmpty() {
return (items.size() == 0);
}
}
CHAPTER 10. GENERIC PROGRAMMING AND COLLECTION CLASSES 523
Changing the name from “T” to “ItemType” has absolutely no effect on the meaning of the
class definition or on the way that Queue is used.
Generic interfaces can be defined in a similar way. It’s also easy to define generic classes
and interfaces that have two or more type parameters, as is done with the standard interface
Map<K,V>. A typical example is the definition of a “Pair” that contains two objects, possibly
of different types. A simple version of such a class can be defined as:
class Pair<T,S> {
public T first;
public S second;
public Pair( T a, S b ) { // Constructor.
first = a;
second = b;
}
}
This class can be used to declare variables and create objects such as:
Pair<String,Color> colorName = new Pair<>("Red", Color.RED);
Pair<Double,Double> coordinates = new Pair<>(17.3,42.8);
Note that in the definition of the constructor in this class, the name “Pair” does not have type
parameters. You might have expected “Pair<T,S>”. However, the name of the class is “Pair”,
not “Pair<T,S>”, and within the definition of the class, “T” and “S” are used as if they are the
names of specific, actual types. Note in any case that type parameters are never added to the
names of methods or constructors, only to the names of classes and interfaces.
Once again, we have some code that works for type String, and we can imagine writing
almost identical code to work with other types of objects. By writing a generic method, we get
to write a single method definition that will work for objects of any type. We need to replace
the specific type String in the definition of the method with the name of a type parameter, such
as T. However, if that’s the only change we make, the compiler will think that “T” is the name
of an actual type, and it will mark it as an undeclared identifier. We need some way of telling
the compiler that “T” is a type parameter. That’s what the “<T>” does in the definition of the
generic class “class Queue<T> { ...”. For a generic method, the “<T>” goes just before the
name of the return type of the method:
public static <T> int countOccurrences(T[] list, T itemToCount) {
int count = 0;
if (itemToCount == null) {
for ( T listItem : list )
if (listItem == null)
count++;
}
else {
for ( T listItem : list )
if (itemToCount.equals(listItem))
count++;
}
return count;
}
The “<T>” marks the method as being generic and specifies the name of the type parameter
that will be used in the definition. Of course, the name of the type parameter doesn’t have to
be “T”; it can be anything. (The “<T>” looks a little strange in that position, I know, but it
had to go somewhere and that’s just where the designers of Java decided to put it.)
Given the generic method definition, we can apply it to objects of any type. If wordList is
a variable of type String[ ] and word is a variable of type String, then
int ct = countOccurrences( wordList, word );
will count the number of times that word occurs in wordList. If palette is a variable of type
Color[ ] and color is a variable of type Color, then
int ct = countOccurrences( palette, color );
will count the number of times that color occurs in palette. If numbers is a variable of type
Integer[ ], then
int ct = countOccurrences( numbers, 17 );
will count the number of times that 17 occurs in numbers. This last example uses autoboxing;
the 17 is automatically converted to a value of type Integer. Note that, since generic program-
ming in Java applies only to objects, we cannot use countOccurrences to count the number
of occurrences of 17 in an array of type int[ ].
A generic method can have one or more type parameters, such as the “T” in
countOccurrences. Note that when a generic method is used, as in the function call
“countOccurrences(wordlist, word)”, there is no explicit mention of the type that is sub-
stituted for the type parameter. The compiler deduces the type from the types of the actual
parameters in the method call. Since wordlist is of type String[ ], the compiler can tell that
CHAPTER 10. GENERIC PROGRAMMING AND COLLECTION CLASSES 525
class Queue<T> {
private LinkedList<T> items = new LinkedList<T>();
public void enqueue(T item) {
items.addLast(item);
}
public T dequeue() {
return items.removeFirst();
}
public boolean isEmpty() {
return (items.size() == 0);
}
public void addAll(Collection<? extends T> collection) {
// Add all the items from the collection to the end of the queue
for ( T item : collection )
enqueue(item);
}
}
Here, T is a type parameter in the generic class definition. We are combining wildcard types
with generic classes. Inside the generic class definition, “T” is used as if it is a specific, though
unknown, type. The wildcard type “? extends T” means some type that is equal to or extends
that specific type. When we create a queue of type Queue<Shape>, “T” refers to “Shape”, and
the wildcard type “? extends T” in the class definition means “? extends Shape”. This
ensures that the addAll method of the queue can be applied to collections of Rects and Ovals
as well as to collections of Shapes.
The for-each loop in the definition of addAll iterates through the collection using a
variable, item, of type T. Now, collection can be of type Collection<S>, where S is a subclass
of T. Since item is of type T, not S, do we have a problem here? No, no problem. As long as
S is a subclass of T, a value of type S can be assigned to a variable of type T. The restriction
on the wildcard type makes everything work nicely.
The addAll method adds all the items from a collection to the queue. Suppose that we
wanted to do the opposite: Add all the items that are currently in the queue to a given collection.
An instance method defined as
public void addAllTo(Collection<T> collection)
would only work for collections whose base type is exactly the same as T. This is too restrictive.
We need some sort of wildcard. However, “? extends T” won’t work. Suppose we try it:
public void addAllTo(Collection<? extends T> collection) {
// Remove all items currently on the queue and add them to collection
while ( ! isEmpty() ) {
T item = dequeue(); // Remove an item from the queue.
collection.add( item ); // Add it to the collection. ILLEGAL!!
}
}
The problem is that we can’t add an item of type T to a collection that might only be able
to hold items belonging to some subclass, S, of T. The containment is going in the wrong
direction: An item of type T is not necessarily of type S. For example, if we have a queue of
type Queue<Shape>, it doesn’t make sense to add items from the queue to a collection of type
Collection<Rect>, since not every Shape is a Rect. On the other hand, if we have a Queue<Rect>,
CHAPTER 10. GENERIC PROGRAMMING AND COLLECTION CLASSES 528
it would make sense to add items from that queue to a Collection<Shape> or indeed to any
collection Collection<S> where S is a superclass of Rect.
To express this type of relationship, we need a new kind of type wildcard: “? super T”.
This wildcard means, roughly, “either T itself or any class that is a superclass of T.” For
example, Collection<? super Rect> would match the types Collection<Shape>, ArrayList<Object>,
and Set<Rect>. This is what we need for our addAllTo method. With this change, our complete
generic queue class becomes:
class Queue<T> {
private LinkedList<T> items = new LinkedList<T>();
public void enqueue(T item) {
items.addLast(item);
}
public T dequeue() {
return items.removeFirst();
}
public boolean isEmpty() {
return (items.size() == 0);
}
public void addAll(Collection<? extends T> collection) {
// Add all the items from the collection to the end of the queue
for ( T item : collection )
enqueue(item);
}
public void addAllTo(Collection<? super T> collection) {
// Remove all items currently on the queue and add them to collection
while ( ! isEmpty() ) {
T item = dequeue(); // Remove an item from the queue.
collection.add( item ); // Add it to the collection.
}
}
}
In a wildcard type such as “? extends T”, T can be an interface instead of a class. Note
that the term “extends” (not “implements”) is used in the wildcard type, even if T is an
interface. For example, we have seen that Runnable is an interface that defines the method
public void run(). Here is a method that runs all the objects in a collection of Runnables by
executing the run() method from each runnable object:
public static runAll( Collection<? extends Runnable> runnables ) {
for ( Runnable runnable : runnables ) {
runnable.run();
}
}
∗ ∗ ∗
Wildcard types are used only as type parameters in parameterized types, such as Collec-
tion<? extends Runnable>. The place where a wildcard type is most likely to occur, by far, is
in a formal parameter list, where the wildcard type is used in the declaration of the type of a
formal parameter. However, they can also be used in a few other places. For example, they can
be used in the type specification in a variable declaration statement.
CHAPTER 10. GENERIC PROGRAMMING AND COLLECTION CLASSES 529
One final remark: The wildcard type “<?>” is equivalent to “<? extends Object>”. That
is, it matches any possible type. For example, the removeAll() method in the generic interface
Collection<T> is declared as
public boolean removeAll( Collection<?> c ) { ...
This just means that the removeAll method can be applied to any collection of any type of
object.
Let’s look at a generic method in which a bounded type parameter is essential. In Subsec-
tion 10.2.1, I presented a code segment for inserting a string into a sorted list of strings, in such
a way that the modified list is still in sorted order. Here is the same code, but this time in the
form of a method definition (and without the comments):
static void sortedInsert(List<String> sortedList, String newItem) {
ListIterator<String> iter = sortedList.listIterator();
while (iter.hasNext()) {
String item = iter.next();
if (newItem.compareTo(item) <= 0) {
iter.previous();
break;
}
}
iter.add(newItem);
}
This method works fine for lists of strings, but it would be nice to have a generic method
that can be applied to lists of other types of objects. The problem, of course, is that the code
assumes that the compareTo() method is defined for objects in the list, so the method can
only work for lists of objects that implement the Comparable interface. We can’t simply use
a wildcard type to enforce this restriction. Suppose we try to do it, by replacing List<String>
with List<? extends Comparable>:
static void sortedInsert(List<? extends Comparable> sortedList, ???? newItem) {
ListIterator<????> iter = sortedList.listIterator();
...
We immediately run into a problem, because we have no name for the unknown type represented
by the wildcard. We need a name for that type because the type of newItem and of iter should
be the same as the type of the items in the list. The problem is solved if we write a generic
method with a bounded type parameter, since then we have a name for the unknown type, and
we can write a valid generic method:
static <T extends Comparable> void sortedInsert(List<T> sortedList, T newItem) {
ListIterator<T> iter = sortedList.listIterator();
while (iter.hasNext()) {
T item = iter.next();
if (newItem.compareTo(item) <= 0) {
iter.previous();
break;
}
}
iter.add(newItem);
}
There is still one technicality to cover in this example. Comparable is itself a parameterized
type, but I have used it here without a type parameter. This is legal but the compiler might give
you a warning about using a “raw type.” In fact, the objects in the list should implement the
parameterized interface Comparable<T>, since they are being compared to items of type T. This
just means that instead of using Comparable as the type bound, we should use Comparable<T>:
static <T extends Comparable<T>> void sortedInsert(List<T> sortedList, ...
CHAPTER 10. GENERIC PROGRAMMING AND COLLECTION CLASSES 532
streams, so the list would have to be fairly large before you would see any speedup. In fact, for
small lists, the stream version will almost certainly take longer than the for loop.
The stream API is complex, and I can only give a basic introduction to it here—but hopefully
enough to convey some of its spirit.
DoublePredicate and LongPredicate. This is typical of how the stream API deals with primitive
types. For example, it defines IntStream to represent a stream of ints as a more efficient
alternative to Stream<Integer>.
The functional interface Supplier<T> defines a function, get() with no parameters and a
return type of T. It represents a source of values of type T. There is a companion interface
Consumer<T> that defines the void function accept(t) with a parameter of type T. There are
also specialized versions for primitive types, including IntSupplier, IntConsumer, DoubleSupplier
and DoubleConsumer. I will give examples of using suppliers and consumers below.
Function<T,R> represents functions from values of type T to values of type R. This functional
interface defines the function apply(t), where t is of type T and the return type is R. The
interface UnaryOperator<T> is essentially Function<T,T>; that is, it represents a function whose
input and output types are the same. Note that DoubleUnaryOperator is a specialized version
of UnaryOperator<Double>, and of course there is also IntUnaryOperator.
Finally, I will mention BinaryOperator<T> and its specializations such as IntBinaryOperator.
The interface BinaryOperator<T> defines the function apply(t1,t2) where t1 and t2 are both
of type T and the return type is also T. Binary operators include things like addition of numbers
or concatenation of strings.
The stream is sequential and is effectively infinite. That is, it will continue to produce values
forever or until trying to do so produces an error. Similarly, IntStream.generate(s) will
create the stream of int values from an IntSupplier, and DoubleStream.generate(s) creates a
stream of doubles from a DoubleSupplier. For example,
DoubleStream.generate( () -> Math.random() )
creates an infinite stream of random numbers. In fact, you can get a similar stream of random
values from a variable, rand, of type Random (see Subsection 5.3.1): rand.doubles() is an
infinite stream of random numbers in the range 0 to 1. If you only want a finite number of
random numbers, use rand.doubles(count). The Random class has other methods for creating
streams of random doubles and ints. You will find other methods that create streams in various
standard classes.
The IntStream interface defines a method for creating a stream containing a given range of
integer values. The stream
IntStream.range( start, end )
is a sequential stream containing the values start, start+1, . . . , end-1. Note that end is not
included.
Some additional methods for making streams have been introduced in newer versions of Java.
For example, for a Scanner, input, Java 9 introduced the method input.tokens(), which makes
a stream consisting of all the strings that would be returned by calling input.next() over and
over. And for a String, str, that contains multiple lines of text, Java 11 added str.lines()
that creates a stream consisting of the lines from the string.
Here are a few more intermediate operations on a stream, S, that can be useful: S.limit(n),
where n is an integer, creates a stream containing only the first n values from S. (If S has fewer
than n values, then S.limit(n) is the same as S.) S.distinct() creates a stream from the
values of S by omitting duplicate values, so that all the values in S.distinct() are different.
And S.sorted() creates a stream containing the same values as S, but in sorted order; to sort
items that do not have a natural ordering, you can provide a Comparator as a parameter to
sorted(). (Comparators are discussed in Subsection 10.1.6.) Note that S.limit(n) can be
especially useful for truncating what would otherwise be an infinite stream, such as a stream
generated from a Supplier.
∗ ∗ ∗
To actually get anything done with a stream, you need to apply a terminal operation at
the end. The operator forEach(c) applies a Consumer, c, to each element of the stream. The
result is not a stream, since consumers do not produce values. The effect of S.forEach(c) on
a stream S is simply to do something with each value in the stream. For example, we have a
whole new way to print all the strings in a list of strings:
stringList.stream().forEach( s -> System.out.println(s) );
For parallel streams, the consumer function is not guaranteed to be applied to the values from
the stream in the same order that they occur in the stream. If you want that guarantee, you
can use forEachOrdered(c) instead of forEach(c).
If we want to print out only some of the strings, say those that have length at least 5, and
if we want them in sorted order, we can apply some filters:
stringList.stream()
.filter( s -> (s.length() >= 5) )
.sorted()
.forEachOrdered( s -> System.out.println(s) )
Some terminal operations output a single value. For example, S.count() returns the num-
ber of values in the stream S. And IntStreams and DoubleStreams have the terminal operation
sum(), to compute the sum of all the values in the stream. Suppose, for example, that you
would like to test the random number generator by generating 10000 random numbers and
counting how many of them are less than 0.5:
long half = DoubleStream.generate( Math::random )
.limit(10000)
.filter( x -> (x < 0.5) )
.count();
Note that count() returns a long rather than an int. Also note that I’ve used the method refer-
ence Math::random here instead of the equivalent lambda expression “() -> Math.random()”
(see Subsection 4.5.4). If you are having trouble reading things like this, keep in mind that the
pattern is: Create a stream, apply some intermediate operations, apply a terminal operation.
Here, an infinite stream of random numbers is generated by calling Math.random() over and
over. The operation limit(10000) truncates that stream to 10000 values, so that in fact only
10000 values are generated. The filter() operation only lets through numbers x such that
x < 0.5 is true. And finally, count() returns the number of items in the resulting stream.
A Stream<T> also has terminal operations min(c) and max(c) to return the smallest and
largest values in the stream. The parameter, c, is of type Comparator<T>; it is used for com-
paring the values. However, the return type of min() and max() is a little peculiar: The return
type is Optional<T>, which represents a value of type T that might or might not exist. The
CHAPTER 10. GENERIC PROGRAMMING AND COLLECTION CLASSES 537
problem is that an empty stream does not have a largest or smallest value, so the minimum
and maximum of an empty stream do not exist. An Optional has a get() method that returns
the value of the Optional, if there is one; it will throw an exception if the Optional is empty. For
example, if words is a Collection<String>, you can get the longest string in the collection with
String longest = words.parallelStream()
.max( (s1,s2) -> s1.length() - s2.length() )
.get();
But this will throw an exception if the collection is empty. (The boolean-valued method
isPresent() in an Optional can be used to test whether the value exists.)
Similarly, IntStream and DoubleStream provide terminal operations min() and max() that
return values of type OptionalInt and OptionalDouble.
The terminal operators allMatch(p) and anyMatch(p) take a predicate as parameter and
compute a boolean value. The value of allMatch(p) is true if the predicate, p, is true for every
value in the stream to which it is applied. The value of anyMatch(p) is true if there is at least
one value in the stream for which p is true. Note that anyMatch() will stop processing, and
will return true as its output, if it finds a value that satisfies the predicate. And allMatch()
will stop processing if it finds a value that does not match the predicate.
Many terminal operations that compute a single value can be expressed in terms of a more
general operation, reduce. A reduce operation combines the values from a string using a
BinaryOperator. For example, a sum is computed by a reduce operation in which the binary
operation is addition. The binary operator should be associative, which means that the order in
which the operator is applied doesn’t matter. There is no built-in terminal operator to compute
the product of the values in a stream, but we can do that directly with reduce. Suppose, for
example, that A is an array of double, and we want the product of all the non-zero elements
in A:
double multiply = Arrays.stream(A).filter( x -> (x != 0) )
.reduce( 1, (x,y) -> x*y );
The binary operator here maps a pair of numbers (x,y) to their product x*y. The first
parameter to reduce() is an “identity” for the binary operation. That is, it is a value such
that 1*x = x for any x. The maximum of a stream of double could be computed with reduce()
by using reduce(Double.NEGATIVE INFINITY, Math::max).
The last major terminal operation is collect(c), a very general operation which collects
all of the values in the stream into a data structure or a single summary result of some type.
The parameter, c is something called a collector. The collector will ordinarily be given by one
of the static functions in the Collectors class. This can get very complicated, and I will only
give a couple of examples. The function Collectors.toList() returns a Collector that can be
used with collect() to put all of the values from the stream into a List. For example, suppose
that A is an array of non-null Strings, and we want a list of all the strings in A that begin with
the substring “Fred”:
List<String> freds = Arrays.stream(A)
.filter( s -> s.startsWith("Fred") )
.collect( Collectors.toList() );
That’s actually pretty easy! Even more useful are collectors that group the items from a stream
according to some criterion. The collector Collectors.groupingBy(f) takes a parameter, f,
whose type is specified by the functional interface Function<T,S>, representing a function from
values of type T to values of type S. When used with collect(), Collectors.groupingBy(f)
CHAPTER 10. GENERIC PROGRAMMING AND COLLECTION CLASSES 538
operates on a stream of type Stream<T>, and it separates the items in the stream into groups,
based on the value of the function f when applied to the items. That is, all the items, x, in a
given group have the same value for f(x). The result is a Map<S,List<T>>. In this map, a key
is one of the function values, f(x), and the associated value for that key is a list containing all
the items from the stream to which f assigns that function value.
An example will clarify things. Suppose we have an array of people, where each person has
a first name and a last name. And suppose that we want to put the people into groups, where
each group consists of all the people with a given last name. A person can be represented by
an object of type Person that contains instance variables named firstname and lastname. Let’s
say that population is a variable of type Person[ ]. Then Arrays.stream(population) is a
stream of Persons, and we can group the people in the stream by last name with the following
code:
Map<String, List<Person>> families;
families = Arrays.stream(population)
.collect(Collectors.groupingBy( person -> person.lastname ));
Here, the lambda expression, person -> person.lastname, defines the grouping function. The
function takes a Person as input and outputs a String. In the resulting Map, families, a key
is one of the last names from the Persons in the array, and the value associated with that last
name is a List containing all the Persons with that last name. We could print out the groups
as follows:
for ( String lastName : families.keySet() ) {
System.out.println("People with last name " + lastName + ":");
for ( Person name : families.get(lastName) ) {
System.out.println(" " + name.firstname + " " + name.lastname);
}
System.out.println();
}
Although the explanation is a bit long-winded, the result should be reasonably easy to under-
stand.
10.6.4 An Experiment
Most of the examples of using streams that I have given so far are not very practical. In most
cases, a simple for loop would have been just as easy to write and probably more efficient. That’s
especially true since I’ve mostly used sequential streams, and most of the examples cannot
be efficiently parallelized. (A notable exception is the reduce operation, which is important
precisely because it parallelizes well.) Let’s look at an example where the stream API is applied
to a long computation that might get some real speedup with parallelization. The problem is to
compute a Riemann sum. This is something from Calculus, but you don’t need to understand
anything at all about what it means. Here is a traditional method for computing the desired
sum:
/**
* Use a basic for loop to compute a Riemann sum.
* @param f The function that is to be summed.
* @param a The left endpoint of the interval over which f is summed.
* @param b The right endpoint.
* @param n The number of subdivisions of the interval.
* @return The value computed for the Riemann sum.
CHAPTER 10. GENERIC PROGRAMMING AND COLLECTION CLASSES 539
*/
private static double riemannSumWithForLoop(
DoubleUnaryOperator f, double a, double b, int n) {
double sum = 0;
double dx = (b - a) / n;
for (int i = 0; i < n; i++) {
sum = sum + f.applyAsDouble(a + i*dx);
}
return sum * dx;
}
The type for the first parameter is a functional interface, so we could call this method, for
example, with
reimannSumWithForLoop( x -> Math.sin(x), 0, Math.PI, 10000 )
How can we apply the stream API to this problem? To imitate the for loop, we can start by
generating the integers from 0 to n as a stream, using IntStream.range(0,n). This gives a
sequential stream. To enable parallelism, we have to convert it to a parallel stream by applying
the .parallel() operation. To compute the values that we want to sum up, we can apply a
map operation that maps the stream of ints to a stream of doubles by mapping the integer i
to f.applyAsDouble(a+i*dx). Finally, we can apply sum() as the terminal operation. Here is
a version of the Riemann sum method that uses a parallel stream:
private static double riemannSumWithParallelStream(
DoubleUnaryOperator f, double a, double b, int n) {
double dx = (b - a) / n;
double sum = IntStream.range(0,n)
.parallel()
.mapToDouble( i -> f.applyAsDouble(a + i*dx) )
.sum();
return sum * dx;
}
I also wrote a version riemannSumWithSequentialStream(), that leaves out the .parallel()
operator. All three versions can be found in the sample program RiemannSumStreamExper-
iment.java. The main routine in that program calls each of the three methods, using various
values for n. It times how long each method takes to compute the sum, and reports the result.
As might be expected, I found that the version that uses a sequential stream is uniformly
slower than the other versions. The sequential stream version does essentially the same thing
as the for loop version, but with the extra overhead involved with creating and manipulating
streams. The situation for parallel streams is more interesting, and the results depend on the
machine on which the program is executed. On one old machine with four processors, the for
loop version was faster for n = 100,000, but the parallel version was much faster for 1,000,000
items or more. On another machine, the parallel version was faster for 10,000 or more items.
Note that there is a limit to how much faster the parallel version can be. On a machine with K
processors, the parallel version cannot be more than K times faster than the sequential version,
and will probably in reality be somewhat slower than that. I encourage you to try out the
sample program on your own computer!
It is even conceivable (or at least this is a goal of the stream API) that you have a machine
on which Java can run parallel code on your graphics card, making use of the many processors
that it contains. If that happens, you might see a very large speedup.
Exercises 540
1. Rewrite the PhoneDirectory class from Subsection 7.4.2 so that it uses a TreeMap to store (solution)
directory entries, instead of an array. (Doing this was suggested in Subsection 10.3.1.)
You should also write a short program to test the class.
2. In mathematics, several operations are defined on sets. The union of two sets A and B is (solution)
a set that contains all the elements that are in A together with all the elements that are
in B. The intersection of A and B is the set that contains elements that are in both A
and B. The difference of A and B is the set that contains all the elements of A except
for those elements that are also in B.
Suppose that A and B are variables of type set in Java. The mathematical opera-
tions on A and B can be computed using methods from the Set interface. In particular:
A.addAll(B) computes the union of A and B; A.retainAll(B) computes the intersection
of A and B; and A.removeAll(B) computes the difference of A and B. (These operations
change the contents of the set A, while the mathematical operations create a new set
without changing A, but that difference is not relevant to this exercise.)
For this exercise, you should write a program that can be used as a “set calculator” for
simple operations on sets of non-negative integers. (Negative integers are not allowed.) For
input and output, a set of such integers will be written as a list of integers, separated by
commas and, optionally, spaces and enclosed in square brackets. For example: [1,2,3] or
[17, 42, 9, 53, 108]. The characters +, *, and - will be used for the union, intersection,
and difference operations. The user of the program will type in lines of input containing
two sets, separated by an operator. The program should perform the operation and print
the resulting set. Here are some examples:
Input Output
------------------------- -------------------
[1, 2, 3] + [3, 5, 7] [1, 2, 3, 5, 7]
[10,9,8,7] * [2,4,6,8] [8]
[ 5, 10, 15, 20 ] - [ 0, 10, 20 ] [5, 15]
To represent sets of non-negative integers, use sets of type TreeSet<Integer>. Read the
user’s input, create two TreeSets, and use the appropriate TreeSet method to perform the
requested operation on the two sets. Your program should be able to read and process
any number of lines of input. If a line contains a syntax error, your program should not
crash. It should report the error and move on to the next line of input. (Note: To print
out a Set, A, of Integers, you can just say System.out.print(A). I’ve chosen the syntax
for sets to be the same as that used by the system for outputting a set.)
3. The fact that Java has a HashMap class means that no Java programmer has to write (solution)
an implementation of hash tables from scratch—unless, of course, that programmer is a
computer science student.
For this exercise, you should write a hash table in which both the keys and the values
are of type String. (This is not an exercise in generic programming; do not try to write a
generic class.) Write an implementation of hash tables from scratch. Define the following
methods: get(key), put(key,value), remove(key), containsKey(key), and size().
Remember that every object, obj, has a method obj.hashCode() that can be used for
computing a hash code for the object, so at least you don’t have to define your own hash
Exercises 541
function. Do not use any of Java’s built-in generic types; create your own linked lists
using nodes as covered in Subsection 9.2.2. However, you are not required to increase the
size of the table when it becomes too full.
You should also write a short program to test your solution.
4. A predicate is a boolean-valued function with one parameter. Java has the parameter- (solution)
ized functional interface Predicate<T>, from package java.util.function, to represent
predicates. The definition of Predicate<T> could be:
public interface Predicate<T> {
public boolean test( T obj );
}
The idea is that an object that implements this interface knows how to “test” objects
of type T in some way. Java already has some methods that use predicates, such as the
removeIf(p) method that is defined for any Collection. (See Subsection 10.6.1). However,
this exercise asks you to write a few similar methods yourself. Define a class that contains
the following generic static methods for working with predicate objects. The name of the
class should be Predicates, in analogy with the standard class Collections that provides
various static methods for working with collections. You should not use the stream API
for this exercise.
public static <T> void remove(Collection<T> coll, Predicate<T> pred)
// Remove every object, obj, from coll for which pred.test(obj)
// is true. (This does the same thing as coll.removeIf(pred).)
public static <T> void retain(Collection<T> coll, Predicate<T> pred)
// Remove every object, obj, from coll for which
// pred.test(obj) is false. (That is, retain the
// objects for which the predicate is true.)
public static <T> List<T> collect(Collection<T> coll, Predicate<T> pred)
// Return a List that contains all the objects, obj,
// from the collection, coll, such that pred.test(obj)
// is true.
public static <T> int find(ArrayList<T> list, Predicate<T> pred)
// Return the index of the first item in list
// for which the predicate is true, if any.
// If there is no such item, return -1.
5. This is a short exercise in using the stream API. Suppose that the class Score is defined (solution)
as
class ScoreInfo {
String firstName;
String lastName;
int score;
Score( String lName, String fName, int s ) {
firstName = fName;
lastName = lName;
score = s;
}
}
Exercises 542
and that scoreData is an array of ScoreInfos containing information about the scores of
students on a test. Use the stream API to do each of the following tasks:
• print the number of students (without using scoreData.length)
• print the average score for all of the students
• print the number of students who got an A (score greater than or equal to 90)
• use the collect() stream operation to create a List<String> that contains the names
of students whose score was less than 70; the names should be in the form first name
followed by last name
• print the names from the List that was generated in the previous task
• print out the student’s names and scores, ordered last name
• print out the student’s names and scores, ordered by score
You can put all of the code in main() routine and include ScoreInfo as a static nested
class. Do not use any for loops or other control structures. Do everything using the
stream API. For testing your code, you can use this array:
private static ScoreInfo[] scoreData = new ScoreInfo[] {
new Score("Smith","John",70),
new Score("Doe","Mary",85),
new Score("Page","Alice",82),
new Score("Cooper","Jill",97),
new Score("Flintstone","Fred",66),
new Score("Rubble","Barney",80),
new Score("Smith","Judy",48),
new Score("Dean","James",90),
new Score("Russ","Joe",55),
new Score("Wolfe","Bill",73),
new Score("Dart","Mary",54),
new Score("Rogers","Chris",78),
new Score("Toole","Pat",51),
new Score("Khan","Omar",93),
new Score("Smith","Ann",95)
};
6. An example in Subsection 10.4.2 concerns the problem of making an index for a book. (solution)
A related problem is making a concordance for a document. A concordance lists every
word that occurs in the document, and for each word it gives the line number of every
line in the document where the word occurs. All the subroutines for creating an index
that were presented in Subsection 10.4.2 can also be used to create a concordance. The
only real difference is that the integers in a concordance are line numbers rather than page
numbers.
Write a program that can create a concordance. The document should be read from
an input file, and the concordance data should be written to an output file. You can use
the indexing subroutines from Subsection 10.4.2, modified to write the data to TextIO
instead of to System.out. (You will need to make these subroutines static.) The input
and output files should be selected by the user when the program is run. The sample
program WordCount.java, from Subsection 10.4.4, can be used as a model of how to use
files. That program also has a useful subroutine that reads one word from input.
Exercises 543
As you read the file, you want to take each word that you encounter and add it to the
concordance along with the current line number. Keeping track of the line numbers is one
of the trickiest parts of the problem. In an input file, the end of each line in the file is
marked by the newline character, ’\n’. Every time you encounter this character, you have
to add one to the line number. WordCount.java ignores ends of lines. Because you need
to find and count the end-of-line characters, your program cannot process the input file in
exactly the same way as does WordCount.java. Also, you will need to detect the end of
the file. The function TextIO.peek(), which is used to look ahead at the next character
in the input, returns the value TextIO.EOF at end-of-file, after all the characters in the
file have been read.
Because it is so common, don’t include the word “the” in your concordance. Also, do
not include words that have length less than 3.
7. The sample program SimpleInterpreter.java from Subsection 10.4.1 can carry out com- (solution)
mands of the form “let variable = expression” or “print expression”. That program can
handle expressions that contain variables, numbers, operators, and parentheses. Extend
the program so that it can also handle the standard mathematical functions sin, cos,
tan, abs, sqrt, and log. For example, the program should be able to evaluate an expres-
sion such as sin(3*x-7)+log(sqrt(y)), assuming that the variables x and y have been
given values. Note that the name of a function must be followed by an expression that is
enclosed in parentheses.
In the original program, a symbol table holds a value for each variable that has been
defined. In your program, you should add another type of symbol to the table to represent
standard functions. You can use the following nested enumerated type and class for this
purpose:
private enum Functions { SIN, COS, TAN, ABS, SQRT, LOG }
/**
* An object of this class represents one of the standard functions.
*/
private static class StandardFunction {
/**
* Tells which function this is.
*/
Functions functionCode;
/**
* Constructor creates an object to represent one of
* the standard functions
* @param code which function is represented.
*/
StandardFunction(Functions code) {
functionCode = code;
}
/**
* Finds the value of this function for the specified
* parameter value, x.
*/
double evaluate(double x) {
switch(functionCode) {
Exercises 544
case SIN:
return Math.sin(x);
case COS:
return Math.cos(x);
case TAN:
return Math.tan(x);
case ABS:
return Math.abs(x);
case SQRT:
return Math.sqrt(x);
default:
return Math.log(x);
}
}
} // end class StandardFunction
Add a symbol to the symbol table to represent each function. The key is the name
of the function and the value is an object of type StandardFunction that represents the
function. For example:
symbolTable.put("sin", new StandardFunction(Function.SIN));
In SimpleInterpreter.java, the symbol table is a map of type HashMap<String,Double>. It’s
not legal to use a StandardFunction as the value in such a map, so you will have to change
the type of the map. The map has to hold two different types of objects. The easy way
to make this possible is to create a map of type HashMap<String,Object>. (A better way
is to create a general type to represent objects that can be values in the symbol table,
and to define two subclasses of that class, one to represent variables and one to represent
standard functions, but for this exercise, you should do it the easy way.)
In your parser, when you encounter a word, you have to be able to tell whether it’s a
variable or a standard function. Look up the word in the symbol table. If the associated
object is non-null and is of type Double, then the word is a variable. If it is of type
StandardFunction, then the word is a function. Remember that you can test the type of
an object using the instanceof operator. For example: if (obj instanceof Double).
Quiz 545
Quiz on Chapter 10
(answers)
2. Why can’t you make an object of type LinkedList<int>? What should you do instead?
3. What is an iterator and why are iterators necessary for generic programming?
5. Interfaces such as List, Set, and Map define abstract data types. Explain what this means.
6. What is the fundamental property that distinguishes Sets from other types of Collections?
9. Modify the following Date class so that it implements the interface Comparable<Date>.
The ordering on objects of type Date should be the natural, chronological ordering.
class Date {
int month; // Month number in range 1 to 12.
int day; // Day number in range 1 to 31.
int year; // Year number.
Date(int m, int d, int y) {
month = m;
day = d;
year = y;
}
}
10. Suppose that syllabus is a variable of type TreeMap<Date,String>, where Date is the class
from the preceding exercise. Write a code segment that will write out the value string for
every key that is in the month of December, 2018.
11. Write a generic class Stack<T> that can be used to represent stacks of objects of type T.
The class should include methods push(), pop(), and isEmpty(). Inside the class, use
an ArrayList to hold the items on the stack.
12. Write a generic method, using a generic type parameter <T>, that replaces every occurrence
in an ArrayList<T> of a specified item with a specified replacement item. The list and the
two items are parameters to the method. Both items are of type T. Take into account
the fact that the item that is being replaced might be null. For a non-null item, use
equals() to do the comparison.
13. Suppose that words is an array of Strings. Explain what is done by the following code:
Quiz 546
long n = Arrays.stream(words)
.filter( w -> (w != null) )
.map( w -> w.toLowerCase() )
.distinct()
.count();
14. Use the stream API to print all the even integers from 2 to 20. Start with
IntStream.range and apply a filter operation.
15. Write a generic method countIf(c,t) with type parameter <T>, where the first parameter,
c, is of type Collection<T>, and the second parameter, p, is of type Predicate<T>. The
method should return the number of items in the collection for which the predicate is true.
Give two versions, one using a loop and the other using the stream API.
Chapter 11
Computer programs are only useful if they interact with the rest of the world in some
way. This interaction is referred to as input/output, or I/O (pronounced “eye-oh”). Up
until now, this book has concentrated on just one type of interaction: interaction with the
user, through either a graphical user interface or a command-line interface. But the user is
only one possible source of information and only one possible destination for information. We
have already encountered one other type of input/output, since TextIO can read data from files
and write data to files. However, Java has an input/output framework that provides much
more power and flexibility than does TextIO, and that covers other kinds of I/O in addition
to files. Most important, aside from files, is that it supports communication over network
connections. In Java, the most common type of input/output involving files and networks is
based on I/O streams, which are objects that support I/O commands that are similar to
those that you have already used. In fact, standard output (System.out) and standard input
(System.in) are examples of I/O streams. (Note that I/O streams are not streams in the sense
of the stream API that was covered in Section 10.6.)
Working with files and networks requires familiarity with exceptions, which were covered in
Section 8.3. Many of the subroutines that are used can throw checked exceptions, which require
mandatory exception handling. This generally means calling the subroutine in a try..catch
statement that can deal with the exception if one occurs. Effective network communication
also requires the use of threads, which will be covered in Chapter 12. We will look at the basic
networking API in this chapter, but we will return to the topic of threads and networking in
Section 12.4.
547
CHAPTER 11. I/O STREAMS, FILES, AND NETWORKING 548
such as “files” and “channels” also exist, but in this section we will look only at streams. Every
stream represents either a source of input or a destination to which output can be sent.
rather than character streams. However, you should prefer Readers and Writers rather than
InputStreams and OutputStreams when working with character data, even when working with
the standard ASCII character set.
The standard I/O stream classes discussed in this section are defined in the package java.io,
along with several supporting classes. You must import the classes from this package if you
want to use them in your program. That means either importing individual classes or putting
the directive “import java.io.*;” at the beginning of your source file. I/O streams are used
for working with files and for doing communication over a network. They can also be used
for communication between two concurrently running threads, and there are stream classes for
reading and writing data stored in the computer’s memory.
(Note: The Java API provides additional support for I/O in the package java.nio and its
subpackages, but they are not covered in this textbook. In general, java.nio gives programmers
efficient access to more advanced I/O techniques.)
The beauty of the stream abstraction is that it is as easy to write data to a file or to send
data over a network as it is to print information on the screen.
∗ ∗ ∗
The basic I/O classes Reader, Writer, InputStream, and OutputStream provide only very
primitive I/O operations. For example, the InputStream class declares an abstract instance
method
public int read() throws IOException
for reading one byte of data, as a number in the range 0 to 255, from an input stream. If the end
of the input stream is encountered, the read() method will return the value -1 instead. If some
error occurs during the input attempt, an exception of type IOException is thrown. Since IOEx-
ception is a checked exception, this means that you can’t use the read() method except inside
a try statement or in a subroutine that is itself declared with a “throws IOException” clause.
(Checked exceptions and mandatory exception handling were covered in Subsection 8.3.3.)
The InputStream class also defines methods for reading multiple bytes of data in one step into
an array of bytes, which can be a lot more efficient that reading individual bytes. However,
InputStream provides no convenient methods for reading other types of data, such as int or
double, from a stream. This is not a problem because you will rarely use an object of type
InputStream itself. Instead, you’ll use subclasses of InputStream that add more convenient input
methods to InputStream’s rather primitive capabilities. Similarly, the OutputStream class defines
a primitive output method for writing one byte of data to an output stream. The method is
defined as:
public void write(int b) throws IOException
The parameter is of type int rather than byte, but the parameter value is type-cast to type
byte before it is written; this effectively discards all but the eight low order bits of b. Again,
in practice, you will almost always use higher-level output operations defined in some subclass
of OutputStream.
The Reader and Writer classes provide the analogous low-level read and write methods.
As in the byte stream classes, the parameter of the write(c) method in Writer and the return
value of the read() method in Reader are of type int, but in these character-oriented classes,
the I/O operations read and write characters rather than bytes. The return value of read()
is -1 if the end of the input stream has been reached. Otherwise, the return value must be
type-cast to type char to obtain the character that was read. In practice, you will ordinarily use
higher level I/O operations provided by sub-classes of Reader and Writer, as discussed below.
CHAPTER 11. I/O STREAMS, FILES, AND NETWORKING 550
11.1.2 PrintWriter
One of the neat things about Java’s I/O package is that it lets you add capabilities to a stream
by “wrapping” it in another stream object that provides those capabilities. The wrapper object
is also a stream, so you can read from or write to it—but you can do so using fancier operations
than those available for basic streams.
For example, PrintWriter is a subclass of Writer that provides convenient methods for out-
putting human-readable character representations of all of Java’s basic data types. If you have
an object belonging to the Writer class, or any of its subclasses, and you would like to use
PrintWriter methods to output data to that Writer, all you have to do is wrap the Writer in a
PrintWriter object. You do this by constructing a new PrintWriter object, using the Writer as
input to the constructor. For example, if charSink is of type Writer, then you could say
PrintWriter printableCharSink = new PrintWriter(charSink);
In fact, the parameter to the constructor can also be an OutputStream or a File, and the
constructor will build a PrintWriter that can write to that output destination. (Files are covered
in the next section.) When you output data to the PrintWriter printableCharSink, using the
high-level output methods in PrintWriter, that data will go to exactly the same place as data
written directly to charSink. You’ve just provided a better interface to the same output
destination. For example, this allows you to use PrintWriter methods to send data to a file or
over a network connection.
For the record, if out is a variable of type PrintWriter, then the following methods are
defined:
• out.print(x) — prints the value of x, represented in the form of a string of characters,
to the output stream; x can be an expression of any type, including both primitive types
and object types. An object is converted to string form using its toString() method. A
null value is represented by the string “null”.
• out.println() — outputs an end-of-line to the output stream.
• out.println(x) — outputs the value of x, followed by an end-of-line; this is equivalent
to out.print(x) followed by out.println().
• out.printf(formatString, x1, x2, ...) — does formatted output of x1, x2, ... to
the output stream. The first parameter is a string that specifies the format of the output.
There can be any number of additional parameters, of any type, but the types of the
parameters must match the formatting directives in the format string. Formatted output
for the standard output stream, System.out, was introduced in Subsection 2.4.1, and
out.printf has the same functionality.
• out.flush() — ensures that characters that have been written with the above methods
are actually sent to the output destination. In some cases, notably when writing to a file
or to the network, it might be necessary to call this method to force the output to actually
appear at the destination.
Note that none of these methods will ever throw an IOException. Instead, the PrintWriter
class includes the method
public boolean checkError()
which will return true if any error has been encountered while writing to the stream. The
PrintWriter class catches any IOExceptions internally, and sets the value of an internal error flag
if one occurs. The checkError() method can be used to check the error flag. This allows you
CHAPTER 11. I/O STREAMS, FILES, AND NETWORKING 551
to use PrintWriter methods without worrying about catching exceptions. On the other hand, to
write a fully robust program, you should call checkError() to test for possible errors whenever
you use a PrintWriter.
this is the character set that is used internally by Java. UTF-8 is a way of encoding UNICODE
characters using 8 bits for common ASCII characters and longer codes for other characters.
ISO-8859-1, also known as “Latin-1,” is an 8-bit encoding that includes ASCII characters as
well as certain accented characters that are used in several European languages. Readers and
Writers use the default charset for the computer on which they are running, unless you specify
a different one. That can be done, for example, in a constructor such as
Writer charSink = new OutputStreamWriter( byteSink, "ISO-8859-1" );
Certainly, the existence of a variety of charset encodings has made text processing more
complicated—unfortunate for us English-speakers but essential for people who use non-Western
character sets. Ordinarily, you don’t have to worry about this, but it’s a good idea to be aware
that different charsets exist in case you run into textual data encoded in a non-default way.
processOneLineOfInput( line );
line = in.readLine();
}
}
catch (IOException e) {
}
This code segment reads and processes lines from standard input until an end-of-stream is
encountered. (An end-of-stream is possible even for interactive input. For example, on at least
some computers, typing a Control-D generates an end-of-stream on the standard input stream.)
The try..catch statement is necessary because the readLine method can throw an exception of
type IOException, which requires mandatory exception handling; an alternative to try..catch
would be to declare that the method that contains the code “throws IOException”. Also,
remember that BufferedReader, InputStreamReader, and IOException must be imported from the
package java.io.
Note that the main purpose of BufferedReader is not simply to make it easier to read lines
of text. Some I/O devices work most efficiently if data is read or written in large chunks,
instead of as individual bytes or characters. A BufferedReader reads a chunk of data, and stores
it in internal memory. The internal memory is known as a buffer . When you read from the
BufferedReader, it will take data from the buffer if possible, and it will only go back to its input
source for more data when the buffer is emptied. There is also a BufferedWriter class, and there
are buffered stream classes for byte streams as well.
∗ ∗ ∗
Previously in this book, we have used the non-standard class TextIO for input both from
users and from files. The advantage of TextIO is that it makes it fairly easy to read data values
of any of the primitive types. Disadvantages include the fact that TextIO can only read from
one input source at a time and that it does not follow the same pattern as Java’s built-in
input/output classes. (If you like the style of input used by TextIO, you might take a look at
my TextReader.java, which implements a similar style of input in a more object-oriented way.
TextReader was used in previous versions of this textbook but is not used in this version.)
and if charSource is of type Reader, you can create a Scanner for reading from charSource
with:
Scanner scanner = new Scanner( charSource );
When processing input, a scanner usually works with tokens. A token is a meaningful
string of characters that cannot, for the purposes at hand, be further broken down into smaller
meaningful pieces. A token can, for example, be an individual word or a string of characters
that represents a value of type double. In the case of a scanner, tokens must be separated by
“delimiters.” By default, the delimiters are whitespace characters such as spaces, tabs, and end-
of-line markers. In normal processing, whitespace characters serve simply to separate tokens
and are discarded by the scanner. A scanner has instance methods for reading tokens of various
types. Suppose that scanner is an object of type Scanner. Then we have:
• scanner.next() — reads the next token from the input source and returns it as a String.
• scanner.nextInt(), scanner.nextDouble(), and so on — read the next token from the
input source and tries to convert it to a value of type int, double, and so on. There are
methods for reading values of any of the primitive types.
• scanner.nextLine() — reads an entire line from the input source, up to the next end-of-
line, and returns the line as a value of type String. The end-of-line marker is read but is
not part of the return value. Note that this method is not based on tokens. An entire line
is read and returned, including any whitespace characters in the line. The return value
can be the empty string.
All of these methods can generate exceptions. If an attempt is made to read past the
end of input, an exception of type NoSuchElementException is thrown. Methods such as
scanner.getInt() will throw an exception of type InputMismatchException if the next to-
ken in the input does not represent a value of the requested type. The exceptions that can be
generated do not require mandatory exception handling.
The Scanner class has very nice look-ahead capabilities. You can query a scanner to de-
termine whether more tokens are available and whether the next token is of a given type. If
scanner is of type Scanner :
• scanner.hasNext() — returns a boolean value that is true if there is at least one more
token in the input source.
• scanner.hasNextInt(), scanner.hasNextDouble(), and so on — return a boolean value
that is true if there is at least one more token in the input source and that token represents
a value of the requested type.
• scanner.hasNextLine() — returns a boolean value that is true if there is at least one
more line in the input source.
Although the insistence on defining tokens only in terms of delimiters limits the usability
of scanners to some extent, they are easy to use and are suitable for many applications. With
so many input classes available—BufferedReader, TextIO, Scanner —you might have trouble de-
ciding which one to use! In general, I would recommend using a Scanner unless you have some
particular reason for preferring TextIO-style input. BufferedReader can be used as a lightweight
alternative when all that you want to do is read entire lines of text from the input source.
(It is possible to change the delimiter that is used by a Scanner, but the syntax uses some-
thing called “regular expressions.” Unfortunately, the syntax for regular expressions is rather
complicated, and they are not covered in this book. However, as an example, suppose you want
CHAPTER 11. I/O STREAMS, FILES, AND NETWORKING 555
tokens to be words that consist entirely of letters of the English alphabet. In that case, delim-
iters should include all non-letter characters. If you want a Scanner, scnr, to use that kind of
delimiter, you can say: scnr.useDelimiter("[^a-zA-Z]+"). After that, tokens returned by
scnr.next() will consist entirely of letters. The string "[^a-zA-Z]+" is a regular expression.
Regular expressions are an important tool for a working programmer. If you have a chance to
learn about them, you should do so.)
correctly to the stream. Because of this, ObjectOutputStreams are meant mainly for use with
“immutable” objects that can’t be changed after they are created. (Strings are an example of
this.) However, if you do need to write mutable objects to an ObjectOutputStream, and if it
is possible that you will write the same object more than once, you can ensure that the full,
correct version of the object will be written by calling the stream’s reset() method before
writing the object to the stream.
11.2 Files
The data and programs in a computer’s main memory survive only as long as the power
is on. For more permanent storage, computers use files, which are collections of data stored
on a hard disk, on a USB memory stick, on a CD-ROM, or on some other type of storage
device. Files are organized into directories (also called folders). A directory can hold other
directories, as well as files. Both directories and files have names that are used to identify them.
Programs can read data from existing files. They can create new files and can write data
to files. In Java, such input and output can be done using I/O streams. Human-readable
character data can be read from a file using an object belonging to the class FileReader, which
is a subclass of Reader. Similarly, data can be written to a file in human-readable format through
an object of type FileWriter, a subclass of Writer. For files that store data in machine format,
the appropriate I/O classes are FileInputStream and FileOutputStream. In this section, I will
only discuss character-oriented file I/O using the FileReader and FileWriter classes. However,
FileInputStream and FileOutputStream are used in an exactly parallel fashion. All these classes
are defined in the java.io package.
you will probably want to wrap your FileReader in a Scanner, in a BufferedReader, or in some
other wrapper class. (See the previous section for a discussion of BufferedReader and Scanner.)
To create a BufferedReader for reading from a file named data.dat, you could say:
BufferedReader data;
try {
data = new BufferedReader( new FileReader("data.dat") );
}
catch (FileNotFoundException e) {
... // handle the exception
}
Wrapping a Reader in a BufferedReader lets you easily read lines of text from the file, and the
buffering can make the input more efficient.
To use a Scanner to read from the file, you can construct the scanner in a similar way.
However, it is more common to construct it more directly from an object of type File (to be
covered below):
Scanner in;
try {
in = new Scanner( new File("data.dat") );
}
catch (FileNotFoundException e) {
... // handle the exception
}
Working with output files is no more difficult than this. You simply create an object
belonging to the class FileWriter. You will probably want to wrap this output stream in an object
of type PrintWriter. For example, suppose you want to write data to a file named “result.dat”.
Since the constructor for FileWriter can throw an exception of type IOException, you should use
a try..catch statement:
PrintWriter result;
try {
result = new PrintWriter(new FileWriter("result.dat"));
}
catch (IOException e) {
... // handle the exception
}
However, as with Scanner, it is more common to use a constructor that takes a File as parameter;
this will automatically wrap the File in a FileWriter before creating the PrintWriter:
PrintWriter result;
try {
result = new PrintWriter(new File("result.dat"));
}
catch (IOException e) {
... // handle the exception
}
CHAPTER 11. I/O STREAMS, FILES, AND NETWORKING 558
You can even use just a String as the parameter to the constructor, and it will be interpreted as
a file name (but you should remember that a String in the Scanner constructor does not name
a file; instead the file will read characters from the string itself).
If no file named result.dat exists, a new file will be created. If the file already exists, then
the current contents of the file will be erased and replaced with the data that your program
writes to the file. This will be done without any warning. To avoid overwriting a file that
already exists, you can check whether a file of the same name already exists before trying
to create the stream, as discussed later in this section. An IOException might occur in the
PrintWriter constructor if, for example, you are trying to create a file on a disk that is “write-
protected,” meaning that it cannot be modified.
When you are finished with a PrintWriter, you should call its flush() method, such as
“result.flush()”, to make sure that all the output has been sent to its destination. If you
forget to do this, you might find that some of the data that you have written to a file has not
actually shown up in the file.
After you are finished using a file, it’s a good idea to close the file, to tell the operating
system that you are finished using it. You can close a file by calling the close() method of
the associated PrintWriter, BufferedReader, or Scanner. Once a file has been closed, it is no
longer possible to read data from it or write data to it, unless you open it again as a new I/O
stream. (Note that for most I/O stream classes, including BufferedReader the close() method
can throw an IOException, which must be handled; however, PrintWriter and Scanner override
this method so that it cannot throw such exceptions.) If you forget to close a file, the file
will ordinarily be closed automatically when the program terminates or when the file object
is garbage collected, but it is better not to depend on this. Note that calling close() should
automatically call flush() before the file is closed. (I have seen that fail, but not recently.)
As a complete example, here is a program that will read numbers from a file named
data.dat, and will then write out the same numbers in reverse order to another file named
result.dat. It is assumed that data.dat contains only real numbers. The input file is read
using a Scanner. Exception-handling is used to check for problems along the way. Although the
application is not a particularly useful one, this program demonstrates the basics of working
with files.
import java.io.*;
import java.util.ArrayList;
import java.util.Scanner;
/**
* Reads numbers from a file named data.dat and writes them to a file
* named result.dat in reverse order. The input file should contain
* only real numbers.
*/
public class ReverseFileWithScanner {
public static void main(String[] args) {
Scanner data; // For reading the data.
PrintWriter result; // Character output stream for writing data.
ArrayList<Double> numbers; // An ArrayList for holding the data.
numbers = new ArrayList<Double>();
try { // Create the input stream.
data = new Scanner(new File("data.dat"));
CHAPTER 11. I/O STREAMS, FILES, AND NETWORKING 559
}
catch (FileNotFoundException e) {
System.out.println("Can’t find file data.dat!");
return; // End the program by returning from main().
}
try { // Create the output stream.
result = new PrintWriter("result.dat");
}
catch (FileNotFoundException e) {
System.out.println("Can’t open file result.dat!");
System.out.println("Error: " + e);
data.close(); // Close the input file.
return; // End the program.
}
while ( data.hasNextDouble() ) { // Read until end-of-file.
double inputNumber = data.nextDouble();
numbers.add( inputNumber );
}
// Output the numbers in reverse order.
for (int i = numbers.size()-1; i >= 0; i--)
result.println(numbers.get(i));
System.out.println("Done!");
data.close();
result.close();
} // end of main()
} // end class ReverseFileWithScanner
Note that this program will simply stop reading data from the file if it encounters anything
other than a number in the input. That will not be considered to be an error.
∗ ∗ ∗
As mentioned at the end of Subsection 8.3.2, the pattern of creating or opening a “resource,”
using it, and then closing the resource is a very common one, and the pattern is supported by
the syntax of the try..catch statement. Files are resources in this sense, as are Scanner,
PrintWriter, and all of Java’s I/O streams. All of these things define close() methods, and
it is good form to close them when you are finished using them. Since they all implement
the AutoCloseable interface, they are all resources in the sense required by try..catch. A
try..catch statement can be used to automatically close a resource when the try statement
ends, which eliminates the need to close it by hand in a finally clause. This assumes that you
will open the resource and use it in the same try..catch.
As an example, the sample program ReverseFileWithResources.java is another version of
the example we have been looking at. In this case, try..catch statements using the resource
pattern are used to read the data from a file and to write the data to a file. My original
program opened a file in one try statement and used it in another try statement. The resource
pattern requires that it all be done in one try, which requires some reorganization of the code
(and can sometimes make it harder to determine the exact cause of an exception). Here is the
try..catch statement from the sample program that opens the input file, reads from it, and
closes it automatically.
CHAPTER 11. I/O STREAMS, FILES, AND NETWORKING 560
When working on the command line, it’s safe to say that if you stick to using simple file names
only, and if the files are stored in the same directory with the program that will use them, then
you will be OK. Later in this section, we’ll look at a convenient way of letting the user specify
a file in a GUI program, which allows you to avoid the issue of path names altogether.
It is possible for a Java program to find out the absolute path names for two important
directories, the current directory and the user’s home directory. The names of these directories
are system properties, and they can be read using the function calls:
• System.getProperty("user.dir") — returns the absolute path name of the current
directory as a String.
• System.getProperty("user.home") — returns the absolute path name of the user’s home
directory as a String.
To avoid some of the problems caused by differences in path names between platforms, Java
has the class java.io.File. An object belonging to this class does not actually represent a
file! Precisely speaking, an object of type File represents a file name rather than a file as such.
The file to which the name refers might or might not exist. Directories are treated in the same
way as files, so a File object can represent a directory just as easily as it can represent a file.
A File object has a constructor, “new File(String)”, that creates a File object from a path
name. The name can be a simple name, a relative path, or an absolute path. For example,
new File("data.dat") creates a File object that refers to a file named data.dat, in the current
directory. Another constructor, “new File(File,String)”, has two parameters. The first is a
File object that refers to a directory. The second can be the name of the file in that directory
or a relative path from that directory to the file.
File objects contain several useful instance methods. Assuming that file is a variable of
type File, here are some of the methods that are available:
• file.exists() — This boolean-valued function returns true if the file named by the
File object already exists. You can use this method if you want to avoid overwriting the
contents of an existing file when you create a new output stream. The boolean function
file.canRead() returns true if the file exists and the program has permission to read the
file. And file.canWrite() is true if the program has permission to write to the file.
• file.isDirectory() — This boolean-valued function returns true if the File object
refers to a directory. It returns false if it refers to a regular file or if no file with the given
name exists.
• file.delete() — Deletes the file, if it exists. Returns a boolean value to indicate whether
the file was successfully deleted.
• file.list() — If the File object refers to a directory, this function returns an array of
type String[] containing the names of the files in that directory. Otherwise, it returns
null. The method file.listFiles() is similar, except that it returns an array of File
instead of an array of String.
Here, for example, is a program that will list the names of all the files in a directory specified
by the user. In this example, I have used a Scanner to read the user’s input:
import java.io.File;
import java.util.Scanner;
/**
* This program lists the files in a directory specified by
* the user. The user is asked to type in a directory name.
CHAPTER 11. I/O STREAMS, FILES, AND NETWORKING 562
call instance methods in the FileChooser object to set some properties of the dialog box. For
example, you can set the file name that is shown to the user as a default initial value for the
file.
A file dialog box can have an “owner,” which is a window, which in JavaFX means an object
of type Stage. Until the dialog box is dismissed by the user—either by canceling the dialog or
selecting a file—all interaction with the owner window is blocked. The owner can be specified
as a parameter to the method that opens the dialog. The owner can be null, which will mean
that no window is blocked.
There are two types of file dialog: an open file dialog that allows the user to specify an
existing file to be opened for reading data into the program; and a save file dialog lets the user
specify a file, which might or might not already exist, to be opened for output. A FileChooser
has two instance methods for showing the two kinds of dialog box on the screen. Suppose that
fileDialog is a variable of type FileChooser. Then the following methods are available:
• fileDialog.showOpenDialog(window) — shows an open file dialog box on the screen.
The parameter specifies the owner of the dialog box. This method does not return until
the user has selected a file or canceled the dialog without selecting a file. The method
returns a value of type File. The return value is null if the user canceled the dialog box.
Otherwise, it returns a File object representing the selected file.
• fileDialog.showSaveDialog(window) — shows a save file dialog box, with owner equal
to window. The parameter and return value are the same as for showOpenDialog(). If
the user selects a file that already exists, the system will automatically ask whether the
user wants to replace that file. So, you can safely go ahead and save the specified file.
• fileDialog.setTitle(title) — specifies a title to appear in the title bar of the dialog
box. The parameter is a string. This method must be called before showing the dialog
box.
• fileDialog.setInitialFileName(name) — sets the name that appears as a default name
in the input box for the file name. The parameter is a string. If the parameter is null,
the input box will be empty; that is also the default. This method must be called before
showing the dialog box.
• fileDialog.setInitialDirectory(directory) — sets the directory that is shown in
the dialog box when it first appears. The parameter is of type File. If the parameter is
null, the initial directory will be a system-dependent default (possibly the directory from
which the program was run). If the parameter is not null, it must be a File object that
represents a directory, not a regular file, or an error will occur. This method must be
called before showing the dialog box.
A typical program has “Save” and “Open” commands for working with files. When the
user selects a file for saving or opening, it can be a good idea to store the selected File object
in an instance variable. Later, that file can be used to initialize the directory and possibly the
file name the next time a file dialog box is created. If editFile is the instance variable that
records the selected file, and if it is non-null, then editFile.getName() is a String giving the
name of the file, and editFile.getParent() is a File representing the directory that contains
the file.
This leaves open one question: what to do when an error occurs while reading or writing
the selected file? The error should be caught, and the user should be informed that an error
occurred. In a GUI program, the natural way to do that is with another dialog box that shows
an error message to the user and has an “OK” button for dismissing the dialog. Dialog boxes
CHAPTER 11. I/O STREAMS, FILES, AND NETWORKING 564
were not covered in Chapter 6, but some common simple dialog boxes can be shown using
objects of type Alert, from package javafx.scene.control. (See Subsection 13.4.1 for more
about alerts.) Here is how to show an error message to the user:
Alert errorAlert = new Alert( Alert.AlertType.ERROR, message );
errorAlert.showAndWait();
Putting all this together, we can look at a typical subroutine that saves data to a file. The
file is selected using a FileChooser. In this example, the data is written in text form, using a
PrintWriter :
private void doSave() {
FileChooser fileDialog = new FileChooser();
if (editFile == null) {
// No file is being edited. Set file name to "filename.txt"
// and set the directory to the user’s home directory.
fileDialog.setInitialFileName("filename.txt");
fileDialog.setInitialDirectory(
new File( System.getProperty("user.home")) );
}
else {
// Get the file name and directory for the dialog from
// the file that is currently being edited.
fileDialog.setInitialFileName(editFile.getName());
fileDialog.setInitialDirectory(editFile.getParentFile());
}
fileDialog.setTitle("Select File to be Saved");
File selectedFile = fileDialog.showSaveDialog(mainWindow);
if ( selectedFile == null )
return; // User did not select a file.
// Note: User has selected a file AND, if the file exists, has
// confirmed that it is OK to erase the exiting file.
PrintWriter out;
try {
FileWriter stream = new FileWriter(selectedFile);
out = new PrintWriter( stream );
}
catch (Exception e) {
// Most likely, user doesn’t have permission to write the file.
Alert errorAlert = new Alert(Alert.AlertType.ERROR,
"Sorry, but an error occurred while\n" +
trying to open the file for output.");
errorAlert.showAndWait();
return;
}
try {
.
. // WRITE TEXT TO THE FILE, using the PrintWriter
.
out.flush(); // (not needed?; it’s probably done by out.close();
out.close();
if (out.checkError()) // (need to check for errors in PrintWriter)
throw new IOException("Error check failed.");
editFile = selectedFile;
CHAPTER 11. I/O STREAMS, FILES, AND NETWORKING 565
}
catch (Exception e) {
Alert errorAlert = new Alert(Alert.AlertType.ERROR,
"Sorry, but an error occurred while\n" +
"trying to write data to the file.");
errorAlert.showAndWait();
}
}
This general outline can easily be adapted to non-text files by using a different type of output
stream.
Reading data from a file is similar, and I won’t show the corresponding doOpen() method
here. You can find working subroutines for saving and opening text files in the sample program
TrivialEdit.java, which lets the user edit small text files. The file subroutines in that program
can be adapted to many GUI programs that work with files.
in the array of strings, args, which is a parameter to the main() routine. The program can
retrieve the command-line arguments from this array. (See Subsection 4.3.6.) For example, if
the program is named CopyFile and if the user runs the program with the command
java CopyFile work.dat oldwork.dat
then in the program, args[0] will be the string "work.dat" and args[1] will be the string
"oldwork.dat". The value of args.length tells the program how many command-line argu-
ments were specified by the user.
The program CopyFile.java gets the names of the files from the command-line arguments.
It prints an error message and exits if the file names are not specified. To add a little interest,
there are two ways to use the program. The command line can simply specify the two file names.
In that case, if the output file already exists, the program will print an error message and end.
This is to make sure that the user won’t accidently overwrite an important file. However, if the
command line has three arguments, then the first argument must be “-f” while the second and
third arguments are file names. The -f is a command-line option, which is meant to modify
the behavior of the program. The program interprets the -f to mean that it’s OK to overwrite
an existing program. (The “f” stands for “force,” since it forces the file to be copied in spite of
what would otherwise have been considered an error.) You can see in the source code how the
command line arguments are interpreted by the program:
import java.io.*;
/**
* Makes a copy of a file. The original file and the name of the
* copy must be given as command-line arguments. In addition, the
* first command-line argument can be "-f"; if present, the program
* will overwrite an existing file; if not, the program will report
* an error and end if the output file already exists. The number
* of bytes that are copied is reported.
*/
public class CopyFile {
public static void main(String[] args) {
String sourceName; // Name of the source file,
// as specified on the command line.
String copyName; // Name of the copy,
// as specified on the command line.
InputStream source; // Stream for reading from the source file.
OutputStream copy; // Stream for writing the copy.
boolean force; // This is set to true if the "-f" option
// is specified on the command line.
int byteCount; // Number of bytes copied from the source file.
/* Get file names from the command line and check for the
presence of the -f option. If the command line is not one
of the two possible legal forms, print an error message and
end this program. */
if (args.length == 3 && args[0].equalsIgnoreCase("-f")) {
sourceName = args[1];
copyName = args[2];
force = true;
}
CHAPTER 11. I/O STREAMS, FILES, AND NETWORKING 567
else if (args.length == 2) {
sourceName = args[0];
copyName = args[1];
force = false;
}
else {
System.out.println(
"Usage: java CopyFile <source-file> <copy-name>");
System.out.println(
" or java CopyFile -f <source-file> <copy-name>");
return;
}
/* Create the input stream. If an error occurs, end the program. */
try {
source = new FileInputStream(sourceName);
}
catch (FileNotFoundException e) {
System.out.println("Can’t find file \"" + sourceName + "\".");
return;
}
/* If the output file already exists and the -f option was not
specified, print an error message and end the program. */
File file = new File(copyName);
if (file.exists() && force == false) {
System.out.println(
"Output file exists. Use the -f option to replace it.");
return;
}
/* Create the output stream. If an error occurs, end the program. */
try {
copy = new FileOutputStream(copyName);
}
catch (IOException e) {
System.out.println("Can’t open output file \"" + copyName + "\".");
return;
}
/* Copy one byte at a time from the input stream to the output
stream, ending when the read() method returns -1 (which is
the signal that the end of the stream has been reached). If any
error occurs, print an error message. Also print a message if
the file has been copied successfully. */
byteCount = 0;
try {
while (true) {
int data = source.read();
if (data < 0)
break;
copy.write(data);
CHAPTER 11. I/O STREAMS, FILES, AND NETWORKING 568
byteCount++;
}
source.close();
copy.close();
System.out.println("Successfully copied " + byteCount + " bytes.");
}
catch (Exception e) {
System.out.println("Error occurred while copying. "
+ byteCount + " bytes copied.");
System.out.println("Error: " + e);
}
} // end main()
the file is read and entered into a TreeMap named phoneBook that represents the phone book
while the program is running. (See Subsection 10.3.1.) In order to store the phone book in a
file, some decision must be made about how the data in the phone book will be represented.
For this example, I chose a simple representation in which each line of the file contains one
entry consisting of a name and the associated phone number. A percent sign (’%’) separates
the name from the number. The following code at the beginning of the program will read the
phone book data file, if it exists and has the correct format:
File userHomeDirectory = new File( System.getProperty("user.home") );
File dataFile = new File( userHomeDirectory, ".phone book data" );
// A file named .phone book data in the user’s home directory.
if ( ! dataFile.exists() ) {
System.out.println("No phone book data file found. A new one");
System.out.println("will be created, if you add any entries.");
System.out.println("File name: " + dataFile.getAbsolutePath());
}
else {
System.out.println("Reading phone book data...");
try( Scanner scanner = new Scanner(dataFile) ) {
while (scanner.hasNextLine()) {
// Read one line from the file, containing one name/number pair.
String phoneEntry = scanner.nextLine();
int separatorPosition = phoneEntry.indexOf(’%’);
if (separatorPosition == -1)
throw new IOException("File is not a phonebook data file.");
name = phoneEntry.substring(0, separatorPosition);
number = phoneEntry.substring(separatorPosition+1);
phoneBook.put(name,number);
}
}
catch (IOException e) {
System.out.println("Error in phone book data file.");
System.out.println("File name: " + dataFile.getAbsolutePath());
System.out.println("This program cannot continue.");
System.exit(1);
}
}
The program then lets the user do various things with the phone book, including making
modifications. Any changes that are made are made only to the TreeMap that holds the data.
When the program ends, the phone book data is written to the file (if any changes have been
made while the program was running), using the following code:
if (changed) {
System.out.println("Saving phone directory changes to file " +
dataFile.getAbsolutePath() + " ...");
PrintWriter out;
try {
out = new PrintWriter( new FileWriter(dataFile) );
}
catch (IOException e) {
System.out.println("ERROR: Can’t open data file for output.");
return;
CHAPTER 11. I/O STREAMS, FILES, AND NETWORKING 570
}
for ( Map.Entry<String,String> entry : phoneBook.entrySet() )
out.println(entry.getKey() + "%" + entry.getValue() );
out.flush();
out.close();
if (out.checkError())
System.out.println("ERROR: Some error occurred while writing data file.");
else
System.out.println("Done.");
}
The net effect of this is that all the data, including the changes, will be there the next time the
program is run. I’ve shown you all the file-handling code from the program. If you would like
to see the rest of the program, see the source code.
Then, a list of type ArrayList<CurveData> is used to hold data for all of the curves that the user
has drawn.
Let’s think about how the data for a sketch could be saved to a text file/ The basic idea is
that all data necessary to reconstitute a sketch must be saved to the output file in some definite
format. The method that reads the file must follow exactly the same format as it reads the
data, and it must use the data to rebuild the data structures that represent the sketch while
the program is running.
When writing character data, all of the data has to be expressed, ultimately, in terms of
simple data values such as strings and primitive type values. A color, for example, can be
expressed in terms of three numbers giving the red, green, and blue components of the color.
The first (not very good) idea that comes to mind might be to just dump all the necessary data,
in some definite order, into the file. Suppose that out is a PrintWriter that is used to write to
the file. We could then say:
Color bgColor = getBackground(); // Write the background color to the file.
out.println( bgColor.getRed() );
out.println( bgColor.getGreen() );
out.println( bgColor.getBlue() );
out.println( curves.size() ); // Write the number of curves.
for ( CurveData curve : curves ) { // For each curve, write...
out.println( curve.color.getRed() ); // the color of the curve
out.println( curve.color.getGreen() );
out.println( curve.color.getBlue() );
out.println( curve.symmetric ? 0 : 1 ); // the curve’s symmetry property
out.println( curve.points.size() ); // the number of points on curve
for ( Point2D pt : curve.points ) { // the coordinates of each point
out.println( pt.getX() );
out.println( pt.getY() );
}
}
This works in the sense that the file-reading method can read the data and rebuild the data
structures. Suppose that the input method uses a Scanner named scanner to read the data
file. Then it could say:
Color newBackgroundColor; // Read the background Color.
double red = scanner.nextDouble();
double green = scanner.nextDouble();
double blue = scanner.nextDouble();
newBackgroundColor = Color.color(red,green,blue);
ArrayList<CurveData> newCurves = new ArrayList<>();
int curveCount = scanner.nextInt(); // The number of curves to be read.
for (int i = 0; i < curveCount; i++) {
CurveData curve = new CurveData();
double r = scanner.nextDouble(); // Read the curve’s color.
double g = scanner.nextDouble();
double b = scanner.nextDouble();
curve.color = Color.color(r,g,b);
int symmetryCode = scanner.nextInt(); // Read the curve’s symmetry property.
curve.symmetric = (symmetryCode == 1);
curveData.points = new ArrayList<>();
CHAPTER 11. I/O STREAMS, FILES, AND NETWORKING 572
xy-coordinates of each point on the curve. Again, the meaning is clear. Files in this format can
easily be created or edited by hand. In fact, the data file shown above was actually created in
a text editor rather than by the program. Furthermore, it’s easy to extend the format to allow
for additional options. Future versions of the program could add a “thickness” property to the
curves to make it possible to have curves with differing line widths. Shapes such as rectangles
and ovals could easily be added.
Outputting data in this format is easy. Suppose that out is a PrintWriter that is being used
to write the sketch data to a file. Then the output is be done with:
out.println("SimplePaintWithFiles 1.0"); // Version number.
out.println( "background " + backgroundColor.getRed() + " " +
backgroundColor.getGreen() + " " + backgroundColor.getBlue() );
for ( CurveData curve : curves ) {
out.println();
out.println("startcurve");
out.println(" color " + curve.color.getRed() + " " +
curve.color.getGreen() + " " + curve.color.getBlue() );
out.println( " symmetry " + curve.symmetric );
for ( Point2D pt : curve.points )
out.println( " coords " + pt.getX() + " " + pt.getY() );
out.println("endcurve");
}
In the program, this code is used in a doSave() method that is similar to the one that is
presented in Subsection 11.2.3. The method uses a file dialog box to allow the user to select
the output file.
Reading the data is somewhat harder, since the input routine has to deal with all the extra
words in the data. In my input routine, I decided to allow some variation in the order in which
the data occurs in the file. For example, the background color can be specified at the end of the
file, instead of at the beginning. It can even be left out altogether, in which case white will be
used as the default background color. This is possible because each item of data is labeled with
a word that describes its meaning; the labels can be used to drive the processing of the input.
Here is the complete method from SimplePaintWithFiles.java that reads data files created by
the doSave() method. It uses a Scanner to read items from the file:
private void doOpen() {
FileChooser fileDialog = new FileChooser();
fileDialog.setTitle("Select File to be Opened");
fileDialog.setInitialFileName(null); // No file is initially selected.
if (editFile == null)
fileDialog.setInitialDirectory(new File(System.getProperty("user.home")));
else
fileDialog.setInitialDirectory(editFile.getParentFile());
File selectedFile = fileDialog.showOpenDialog(window);
if (selectedFile == null)
return; // User canceled.
Scanner scanner;
try {
scanner = new Scanner( selectedFile );
}
catch (Exception e) {
Alert errorAlert = new Alert(Alert.AlertType.ERROR,
"Sorry, but an error occurred\nwhile trying to open the file.");
CHAPTER 11. I/O STREAMS, FILES, AND NETWORKING 574
errorAlert.showAndWait();
return;
}
try {
String programName = scanner.next();
if ( ! programName.equals("SimplePaintWithFiles") )
throw new IOException("File is not a SimplePaintWithFiles data file.");
double version = scanner.nextDouble();
if (version > 1.0)
throw new IOException("File requires a newer version of SimplePaintWithFiles.");
Color newBackgroundColor = Color.WHITE;
ArrayList<CurveData> newCurves = new ArrayList<CurveData>();
while (scanner.hasNext()) {
String itemName = scanner.next();
if (itemName.equalsIgnoreCase("background")) {
double red = scanner.nextDouble();
double green = scanner.nextDouble();
double blue = scanner.nextDouble();
newBackgroundColor = Color.color(red,green,blue);
}
else if (itemName.equalsIgnoreCase("startcurve")) {
CurveData curve = new CurveData();
curve.color = Color.BLACK;
curve.symmetric = false;
curve.points = new ArrayList<Point2D>();
itemName = scanner.next();
while ( ! itemName.equalsIgnoreCase("endcurve") ) {
if (itemName.equalsIgnoreCase("color")) {
double r = scanner.nextDouble();
double g = scanner.nextDouble();
double b = scanner.nextDouble();
curve.color = Color.color(r,g,b);
}
else if (itemName.equalsIgnoreCase("symmetry")) {
curve.symmetric = scanner.nextBoolean();
}
else if (itemName.equalsIgnoreCase("coords")) {
double x = scanner.nextDouble();
double y = scanner.nextDouble();
curve.points.add( new Point2D(x,y) );
}
else {
throw new Exception("Unknown term in input.");
}
itemName = scanner.next();
}
newCurves.add(curve);
}
else {
throw new Exception("Unknown term in input.");
}
}
scanner.close();
backgroundColor = newBackgroundColor;
CHAPTER 11. I/O STREAMS, FILES, AND NETWORKING 575
curves = newCurves;
redraw();
editFile = selectedFile;
window.setTitle("SimplePaint: " + editFile.getName());
}
catch (Exception e) {
Alert errorAlert = new Alert(Alert.AlertType.ERROR,
"Sorry, but an error occurred while\ntrying to read the data:\n"
+ e);
errorAlert.showAndWait();
}
}
The main reason for this long discussion of file formats has been to get you to think about
the problem of representing complex data in a form suitable for storing the data in a file. The
same problem arises when data must be transmitted over a network. There is no one correct
solution to the problem, but some solutions are certainly better than others. In Section 11.5,
we will look at one solution to the data representation problem that has become increasingly
common.
∗ ∗ ∗
In addition to being able to save sketch data in text form, SimplePaintWithFiles can also
save the picture itself as an image file that could be, for example, printed or put on a web page.
This is a preview of image-handling techniques that will be covered in Subsection 13.2.6, and
it uses techniques that I have not yet covered.
11.4 Networking
As far as a program is concerned, a network is just another possible source of input data,
and another place where data can be output. That does oversimplify things, because networks
are not as easy to work with as files are. But in Java, you can do network communication using
input streams and output streams, just as you can use such streams to communicate with the
user or to work with files. Nevertheless, opening a network connection between two computers
is a bit tricky, since there are two computers involved and they have to somehow agree to open a
connection. And when each computer can send data to the other, synchronizing communication
can be a problem. But the fundamentals are the same as for other forms of I/O.
One of the standard Java packages is called java.net. This package includes several
classes that can be used for networking. Two different styles of network I/O are supported.
One of these, which is fairly high-level, is based on the World Wide Web, and provides the
sort of network communication capability that is used by a Web browser when it downloads
pages for you to view. The main classes for this style of networking are java.net.URL and
java.net.URLConnection. An object of type URL is an abstract representation of a Univer-
sal Resource Locator , which is an address for an HTML document or other resource. A
URLConnection represents a network connection to such a resource.
The second style of I/O, which is more general and more important, views the network at
a lower level. It is based on the idea of a socket. A socket is used by a program to establish a
connection with another program on a network. Communication over a network involves two
sockets, one on each of the computers involved in the communication. Java uses a class called
java.net.Socket to represent sockets that are used for network communication. The term
“socket” presumably comes from an image of physically plugging a wire into a computer to
CHAPTER 11. I/O STREAMS, FILES, AND NETWORKING 576
to determine the type. The type might not be available until after the input stream has
been created, so you should generally call getContentType() after getInputStream(). The
string returned by getContentType() is in a format called a mime type. Mime types include
“text/plain”, “text/html”, “image/jpeg”, “image/png”, and many others. All mime types
contain two parts: a general type, such as “text” or “image”, and a more specific type within
that general category, such as “html” or “png”. If you are only interested in text data, for
example, you can check whether the string returned by getContentType() starts with “text”.
(Mime types were first introduced to describe the content of email messages. The name stands
for “Multipurpose Internet Mail Extensions.” They are now used almost universally to specify
the type of information in a file or other resource.)
Let’s look at a short example that uses all this to read the data from a URL. This subroutine
opens a connection to a specified URL, checks that the type of data at the URL is text, and
then copies the text onto the screen. Many of the operations in this subroutine can throw
exceptions. They are handled by declaring that the subroutine “throws IOException” and
leaving it up to the main program to decide what to do when an error occurs.
static void readTextFromURL( String urlString ) throws IOException {
/* Open a connection to the URL, and get an input stream
for reading data from the URL. */
URL url = new URL(urlString);
URLConnection connection = url.openConnection();
InputStream urlData = connection.getInputStream();
/* Check that the content is some type of text. Note:
connection.getContentType() method should be called
after getInputStream(). */
String contentType = connection.getContentType();
System.out.println("Stream opened with content type: " + contentType);
System.out.println();
if (contentType == null || contentType.startsWith("text") == false)
throw new IOException("URL does not seem to refer to a text file.");
System.out.println("Fetching context from " + urlString + " ...");
System.out.println();
/* Copy lines of text from the input stream to the screen, until
end-of-file is encountered (or an error occurs). */
BufferedReader in; // For reading from the connection’s input stream.
in = new BufferedReader( new InputStreamReader(urlData) );
while (true) {
String line = in.readLine();
if (line == null)
break;
System.out.println(line);
}
in.close();
} // end readTextFromURL()
A complete program that uses this subroutine can be found in the file FetchURL.java. When
you run the program, you can specify the URL on the command line; if not, you will be
prompted to enter the URL. For this program, a URL can begin with “http://” or “https://”
CHAPTER 11. I/O STREAMS, FILES, AND NETWORKING 578
for a URL that refers to a resource on the web, with “file://” for a URL that refers to a file on
your computer, or with “ftp://” for a URL that uses the “File Transfer Protocol.” If it does
not start with any of these, then “http://” is added to the start of the URL. Try the program
with URL math.hws.edu/javanotes to fetch the front page of this textbook from its website.
Try it with some bad inputs to see the various errors that can occur.
A client program has to have some way to specify which computer, among all those on
the network, it wants to communicate with. Every computer on the Internet has an IP ad-
dress which identifies it. Many computers can also be referred to by domain names such as
math.hws.edu or www.whitehouse.gov. (See Section 1.7.) Traditional (or IPv4 ) IP addresses
are 32-bit integers. They are usually written in the so-called “dotted decimal” form, such as
64.89.144.237, where each of the four numbers in the address represents an 8-bit integer in
the range 0 through 255. A new version of the Internet Protocol, IPv6 , is currently being
introduced. IPv6 addresses are 128-bit integers and are usually written in hexadecimal form
(with some colons and maybe some extra information thrown in). In actual use, IPv6 addresses
are still fairly rare.
A computer can have several IP addresses, and can have both IPv4 and IPv6 addresses.
Usually, one of these is the loopback address, which can be used when a program wants to
communicate with another program on the same computer. The loopback address has IPv4
address 127.0.0.1 and can also, in general, be referred to using the domain name localhost. In
addition, there can be one or more IP addresses associated with physical network connections.
Your computer probably has some utility for displaying your computer’s IP addresses. I have
written a small Java program, ShowMyNetwork.java, that does the same thing. When I run
ShowMyNetwork on my computer, the output is:
en1 : /192.168.1.47 /fe80:0:0:0:211:24ff:fe9c:5271%5
lo0 : /127.0.0.1 /fe80:0:0:0:0:0:0:1%1 /0:0:0:0:0:0:0:1%0
The first thing on each line is a network interface name, which is really meaningful only to the
computer’s operating system. The same line also contains the IP addresses for that interface.
In this example, lo0 refers to the loopback address, which has IPv4 address 127.0.0.1 as usual.
The most important number here is 192.168.1.47, which is the IPv4 address that can be used
for communication over the network. (The slashes at the start of each address are not part of
the actual address.) The other numbers in the output are IPv6 addresses.
Now, a single computer might have several programs doing network communication at the
same time, or one program communicating with several other computers. To allow for this
possibility, a network connection actually has a port number in combination with an IP
address. A port number is just a 16-bit positive integer. A server does not simply listen for
connections—it listens for connections on a particular port. A potential client must know both
the Internet address (or domain name) of the computer on which the server is running and
the port number on which the server is listening. A Web server, for example, generally listens
for connections on port 80; other standard Internet services also have standard port numbers.
(The standard port numbers are all less than 1024, and are reserved for particular services. If
you create your own server programs, you should use port numbers greater than 1024.)
When you construct a ServerSocket object, you have to specify the port number on which
the server will listen. The specification for the constructor is
public ServerSocket(int port) throws IOException
The port number must be in the range 0 through 65535, and should generally be greater than
1024. The constructor might throw a SecurityException if a port number smaller than 1024 is
specified. An IOException can also occur if, for example, the specified port number is already
in use. (A parameter value of 0 in this method tells the server socket to listen on any available
port.)
As soon as a ServerSocket is created, it starts listening for connection requests. The
accept() method in the ServerSocket class accepts such a request, establishes a connection
with the client, and returns a Socket that can be used for communication with the client. The
accept() method has the form
public Socket accept() throws IOException
When you call the accept() method, it will not return until a connection request is received (or
until some error occurs). The method is said to block while waiting for the connection. (While
the method is blocked, the program—or more exactly, the thread—that called the method
can’t do anything else. If there are other threads in the same program, they can proceed.) You
can call accept() repeatedly to accept multiple connection requests. The ServerSocket will
continue listening for connections until it is closed, using its close() method, or until some
error occurs, or until the program is terminated in some way.
Suppose that you want a server to listen on port 1728, and that you want it to continue to
accept connections as long as the program is running. Suppose that you’ve written a method
provideService(Socket) to handle the communication with one client. Then the basic form
of the server program would be:
try {
ServerSocket server = new ServerSocket(1728);
while (true) {
Socket connection = server.accept();
provideService(connection);
}
}
catch (IOException e) {
System.out.println("Server shut down with error: " + e);
}
On the client side, a client socket is created using a constructor in the Socket class. To
connect to a server on a known computer and port, you would use the constructor
public Socket(String computer, int port) throws IOException
The first parameter can be either an IP number or a domain name. This constructor will block
until the connection is established or until an error occurs.
Once you have a connected socket, no matter how it was created, you can use the Socket
methods getInputStream() and getOutputStream() to obtain streams that can be used for
communication over the connection. These methods return objects of type InputStream and
OutputStream, respectively. Keeping all this in mind, here is the outline of a method for
working with a client connection:
CHAPTER 11. I/O STREAMS, FILES, AND NETWORKING 581
/**
* Open a client connection to a specified server computer and
* port number on the server, and then do communication through
* the connection.
*/
void doClientConnection(String computerName, int serverPort) {
Socket connection;
InputStream in;
OutputStream out;
try {
connection = new Socket(computerName,serverPort);
in = connection.getInputStream();
out = connection.getOutputStream();
}
catch (IOException e) {
System.out.println(
"Attempt to create connection failed with error: " + e);
return;
}
.
. // Use the streams, in and out, to communicate with the server.
.
try {
connection.close();
// (Alternatively, you might depend on the server
// to close the connection.)
}
catch (IOException e) {
}
} // end doClientConnection()
All this makes network communication sound easier than it really is. (And if you think it
sounded hard, then it’s even harder.) If networks were completely reliable, things would be
almost as easy as I’ve described. The problem, though, is to write robust programs that can
deal with network and human error. I won’t go into detail. However, what I’ve covered here
should give you the basic ideas of network programming, and it is enough to write some simple
network applications. Let’s look at a few working examples of client/server programming.
not specified on the command line, then the user is prompted to enter one. Here is the complete
client program:
import java.net.*;
import java.util.Scanner;
import java.io.*;
/**
* This program opens a connection to a computer specified
* as the first command-line argument. If no command-line
* argument is given, it prompts the user for a computer
* to connect to. The connection is made to
* the port specified by LISTENING PORT. The program reads one
* line of text from the connection and then closes the
* connection. It displays the text that it read on
* standard output. This program is meant to be used with
* the server program, DateServer, which sends the current
* date and time on the computer where the server is running.
*/
public class DateClient {
public static final int LISTENING PORT = 32007;
public static void main(String[] args) {
String hostName; // Name of the server computer to connect to.
Socket connection; // A socket for communicating with server.
BufferedReader incoming; // For reading data from the connection.
/* Get computer name from command line. */
if (args.length > 0)
hostName = args[0];
else {
Scanner stdin = new Scanner(System.in);
System.out.print("Enter computer name or IP address: ");
hostName = stdin.nextLine();
}
/* Make the connection, then read and display a line of text. */
try {
connection = new Socket( hostName, LISTENING PORT );
incoming = new BufferedReader(
new InputStreamReader(connection.getInputStream()) );
String lineFromServer = incoming.readLine();
if (lineFromServer == null) {
// A null from incoming.readLine() indicates that
// end-of-stream was encountered.
throw new IOException("Connection was opened, " +
"but server did not send any data.");
}
System.out.println();
System.out.println(lineFromServer);
System.out.println();
incoming.close();
}
CHAPTER 11. I/O STREAMS, FILES, AND NETWORKING 583
catch (Exception e) {
System.out.println("Error: " + e);
}
} // end main()
*/
public class DateServer {
public static final int LISTENING PORT = 32007;
public static void main(String[] args) {
ServerSocket listener; // Listens for incoming connections.
Socket connection; // For communication with the connecting program.
/* Accept and process connections forever, or until some error occurs.
(Note that errors that occur while communicating with a connected
program are caught and handled in the sendDate() routine, so
they will not crash the server.) */
try {
listener = new ServerSocket(LISTENING PORT);
System.out.println("Listening on port " + LISTENING PORT);
while (true) {
// Accept next connection request and handle it.
connection = listener.accept();
sendDate(connection);
}
}
catch (Exception e) {
System.out.println("Sorry, the server has shut down.");
System.out.println("Error: " + e);
return;
}
} // end main()
/**
* The parameter, client, is a socket that is already connected to another
* program. Get an output stream for the connection, send the current time,
* and close the connection.
*/
private static void sendDate(Socket client) {
try {
System.out.println("Connection from " +
client.getInetAddress().toString() );
Date now = new Date(); // The current date and time.
PrintWriter outgoing; // Stream for sending data.
outgoing = new PrintWriter( client.getOutputStream() );
outgoing.println( now.toString() );
outgoing.flush(); // Make sure the data is actually sent!
client.close();
}
catch (Exception e){
System.out.println("Error: " + e);
}
} // end sendDate()
When you run DateServer in a command-line interface, it will sit and wait for connection
requests and report them as they are received. To make the DateServer service permanently
available on a computer, the program would be run as a daemon. A daemon is a program that
runs continually on a computer, independently of any user. The computer can be configured
to start the daemon automatically as soon as the computer boots up. It then runs in the
background, even while the computer is being used for other purposes. For example, a computer
that makes pages available on the World Wide Web runs a daemon that listens for requests
for web pages and responds by transmitting the pages. It’s just a souped-up analog of the
DateServer program! However, the question of how to set up a program as a daemon is not one
I want to go into here. For testing purposes, it’s easy enough to start the program by hand,
and, in any case, my examples are not really robust enough or full-featured enough to be run
as serious servers. (By the way, the word “daemon” is just an alternative spelling of “demon”
and is usually pronounced the same way.)
Note that after calling outgoing.println() to send a line of data to the client, the server
program calls outgoing.flush(). The flush() method is available in every output stream
class. Calling it ensures that data that has been written to the stream is actually sent to its
destination. You should generally call this function every time you use an output stream to
send data over a network connection. If you don’t do so, it’s possible that the stream will collect
data until it has a reasonably large batch of data to send. This is done for efficiency, but it can
impose unacceptable delays when the client is waiting for the transmission. It is even possible
that some of the data might remain untransmitted when the socket is closed, so it is especially
important to call flush() before closing the connection. This is one of those unfortunate cases
where different implementations of Java can behave differently. If you fail to flush your output
streams, it is possible that your network application will work on some types of computers but
not on others.
/**
* This program is one end of a simple command-line interface chat program.
* It acts as a server which waits for a connection from the CLChatClient
* program. The port on which the server listens can be specified as a
* command-line argument. If it is not, then the port specified by the
* constant DEFAULT PORT is used. Note that if a port number of zero is
* specified, then the server will listen on any available port.
* This program only supports one connection. As soon as a connection is
* opened, the listening socket is closed down. The two ends of the connection
* each send a HANDSHAKE string to the other, so that both ends can verify
* that the program on the other end is of the right type. Then the connected
* programs alternate sending messages to each other. The client always sends
* the first message. The user on either end can close the connection by
* entering the string "quit" when prompted for a message. Note that the first
* character of any string sent over the connection must be 0 or 1; this
* character is interpreted as a command.
*/
public class CLChatServer {
/**
* Port to listen on, if none is specified on the command line.
*/
static final int DEFAULT PORT = 1728;
/**
* Handshake string. Each end of the connection sends this string to the
* other just after the connection is opened. This is done to confirm that
* the program on the other side of the connection is a CLChat program.
*/
static final String HANDSHAKE = "CLChat";
/**
* This character is prepended to every message that is sent.
*/
static final char MESSAGE = ’0’;
/**
* This character is sent to the connected program when the user quits.
*/
static final char CLOSE = ’1’;
This program is a little more robust than DateServer. For one thing, it uses a handshake to
make sure that a client who is trying to connect is really a CLChatClient program. A handshake
is simply information sent between a client and a server as part of setting up a connection,
before any actual data is sent. In this case, each side of the connection sends a string to
the other side to identify itself. The handshake is part of the protocol that I made up for
communication between CLChatClient and CLChatServer. A protocol is a detailed specification
of what data and messages can be exchanged over a connection, how they must be represented,
and what order they can be sent in. When you design a client/server application, the design
of the protocol is an important consideration. Another aspect of the CLChat protocol is that
after the handshake, every line of text that is sent over the connection begins with a character
that acts as a command. If the character is 0, the rest of the line is a message from one user to
the other. If the character is 1, the line indicates that a user has entered the “quit” command,
and the connection is to be shut down.
CHAPTER 11. I/O STREAMS, FILES, AND NETWORKING 589
Remember that if you want to try out this program on a single computer, you can use
two command-line windows. In one, give the command “java CLChatServer” to start the
server. Then, in the other, use the command “java CLChatClient localhost” to connect to
the server that is running on the same machine. If the server is not listening on the default
port, you can add the port number as a second command line option to the program. If you
do not specify the connection information on the command line, you will be asked to enter it.
Every well-formed XML document follows a strict syntax. Here are some of the most
important syntax rules: Tag names and attribute names in XML are case sensitive. A name
must begin with a letter and can contain letters, digits and certain other characters. Spaces
and ends-of-line are significant only in textual content. Every tag must either be an empty
tag or have a matching end-tag. By “matching” here, I mean that elements must be properly
nested; if a tag is inside some element, then the matching end-tag must also be inside that
element. A document must have a root element, which contains all the other elements. The
root element in the above example has tag name simplepaint. Every attribute must have
a value, and that value must be enclosed in quotation marks; either single quotes or double
quotes can be used for this. The special characters < and &, if they appear in attribute values
or textual content, must be written as < and &. “<” and “&” are examples
of entities. The entities >, ", and ' are also defined, representing >, double
quote, and single quote. (Additional entities can be defined in a DTD.)
While this description will not enable you to understand everything that you might en-
counter in XML documents, it should allow you to design well-formed XML documents to
represent data structures used in Java programs.
from it. This could be difficult to do by hand. Fortunately, Java has a standard API for parsing
and processing XML Documents. (Actually, it has two, but we will only look at one of them.)
A well-formed XML document has a certain structure, consisting of elements containing
attributes, nested elements, and textual content. It’s possible to build a data structure in the
computer’s memory that corresponds to the structure and content of the document. Of course,
there are many ways to do this, but there is one common standard representation known as
the Document Object Model , or DOM. The DOM specifies how to build data structures
to represent XML documents, and it specifies some standard methods for accessing the data
in that structure. The data structure is a kind of tree whose structure mirrors the structure
of the document. The tree is constructed from nodes of various types. There are nodes to
represent elements, attributes, and text. (The tree can also contain several other types of node,
representing aspects of XML that we can ignore here.) Attributes and text can be processed
without directly manipulating the corresponding nodes, so we will be concerned almost entirely
with element nodes.
(The sample program XMLDemo.java lets you experiment with XML and the DOM. It has
a text area where you can enter an XML document. Initially, the input area contains the sample
XML document from this section. When you click a button named “Parse XML Input”, the
program will attempt to read the XML from the input box and build a DOM representation of
that document. If the input is not well-formed XML, an error message is displayed. If it is legal,
the program will traverse the DOM representation and display a list of elements, attributes,
and textual content that it encounters. The program uses a few techniques for processing XML
that I won’t discuss here.)
In Java, the DOM representation of an XML document file can be created with just two
statements. If selectedFile is a variable of type File that represents the XML file, and xmldoc
is of type Document, then
DocumentBuilder docReader
= DocumentBuilderFactory.newInstance().newDocumentBuilder();
xmldoc = docReader.parse(selectedFile);
will open the file, read its contents, and build the DOM representation. The classes Document-
Builder and DocumentBuilderFactory are both defined in the package javax.xml.parsers. The
method docReader.parse() does the actual work. It will throw an exception if it can’t read
the file or if the file does not contain a legal XML document. If it succeeds, then the value
returned by docReader.parse() is an object that represents the entire XML document. (This
is a very complex task! It has been coded once and for all into a method that can be used very
easily in any Java program. We see the benefit of using a standardized syntax.)
The structure of the DOM data structure is defined in the package org.w3c.dom, which
contains several data types that represent an XML document as a whole and the individual
nodes in a document. The “org.w3c” in the name refers to the World Wide Web Consortium,
W3C, which is the standards organization for the Web. DOM, like XML, is a general standard,
not just a Java standard. The data types that we need here are Document, Node, Element, and
NodeList. (They are defined as interfaces rather than classes, but that fact is not relevant
here.) We can use methods that are defined in these data types to access the data in the DOM
representation of an XML document.
An object of type Document represents an entire XML document. The return value of
docReader.parse()—xmldoc in the above example—is of type Document. We will only need
one method from this class: If xmldoc is of type Document, then
xmldoc.getDocumentElement()
CHAPTER 11. I/O STREAMS, FILES, AND NETWORKING 593
returns a value of type Element that represents the root element of the document. (Recall that
this is the top-level element that contains all the other elements.) In the sample XML document
from earlier in this section, the root element consists of the tag <simplepaint version="1.0">,
the end-tag </simplepaint>, and everything in between. The elements that are nested inside
the root element are represented by their own nodes, which are said to be children of the
root node. An object of type Element contains several useful methods. If element is of type
Element, then we have:
• element.getTagName() — returns a String containing the name that is used in the ele-
ment’s tag. For example, the name of a <curve> element is the string “curve”.
• element.getAttribute(attrName) — if attrName is the name of an attribute in the
element, then this method returns the value of that attribute. For the element,
<point x="83" y="42"/>, element.getAttribute("x") would return the string “83”.
Note that the return value is always a String, even if the attribute is supposed to represent
a numerical value. If the element has no attribute with the specified name, then the return
value is an empty string.
• element.getTextContent() — returns a String containing all of the textual content that
is contained in the element. Note that this includes text that is contained inside other
elements that are nested inside the element.
• element.getChildNodes() — returns a value of type NodeList that contains all the Nodes
that are children of the element. The list includes nodes representing other elements and
textual content that are directly nested in the element (as well as some other types of node
that I don’t care about here). The getChildNodes() method makes it possible to traverse
the entire DOM data structure by starting with the root element, looking at children of
the root element, children of the children, and so on. (There is a similar method that
returns the attributes of the element, but I won’t be using it here.)
• element.getElementsByTagName(tagName) — returns a NodeList that contains all the
nodes representing all elements that are nested inside element and which have the given
tag name. Note that this includes elements that are nested to any level, not just elements
that are directly contained inside element. The getElementsByTagName() method allows
you to reach into the document and pull out specific data that you are interested in.
An object of type NodeList represents a list of Nodes. Unfortunately, it does not use the
API defined for lists in the Java Collection Framework. Instead, a value, nodeList, of type
NodeList has two methods: nodeList.getLength() returns the number of nodes in the list,
and nodeList.item(i) returns the node at position i, where the positions are numbered 0,
1, . . . , nodeList.getLength() - 1. Note that the return value of nodeList.get() is of type
Node, and it might have to be type-cast to a more specific node type before it is used.
Knowing just this much, you can do the most common types of processing of DOM rep-
resentations. Let’s look at a few code fragments. Suppose that in the course of processing a
document you come across an Element node that represents the element
<background red=’1’ green=’0.6’ blue=’0.2’/>
This element might be encountered either while traversing the document with getChildNodes()
or in the result of a call to getElementsByTagName("background"). Our goal is to reconstruct
the data structure represented by the document, and this element represents part of that data.
In this case, the element represents a color, and the red, green, and blue components are given
by the attributes of the element. If element is a variable that refers to the node, the color can
be obtained by saying:
CHAPTER 11. I/O STREAMS, FILES, AND NETWORKING 594
curves = newCurves;
You can find the complete source code in SimplePaintWithXML.java.
∗ ∗ ∗
XML has developed into an extremely important technology, and some applications of it are
very complex. But there is a core of simple ideas that can be easily applied in Java. Knowing
just the basics, you can make good use of XML in your own Java programs.
Exercises 597
1. The sample program DirectoryList.java, given as an example in Subsection 11.2.2, will (solution)
print a list of files in a directory specified by the user. But some of the files in that
directory might themselves be directories. And the subdirectories can themselves contain
directories. And so on. Write a modified version of DirectoryList that will list all the
files in a directory and all its subdirectories, to any level of nesting. You will need a
recursive subroutine to do the listing. The subroutine should have a parameter of type
File. You will need the constructor from the File class that has the form
public File( File dir, String fileName )
// Constructs the File object representing a file
// named fileName in the directory specified by dir.
2. Write a program that will count the number of lines in each file that is specified on the (solution)
command line. Assume that the files are text files. Note that multiple files can be specified,
as in:
java LineCounts file1.txt file2.txt file3.txt
Write each file name, along with the number of lines in that file, to standard output. If an
error occurs while trying to read from one of the files, you should print an error message
for that file, but you should still process all the remaining files. Do not use TextIO to
process the files; use a Scanner or a BufferedReader to process each file.
3. For this exercise, you will write a network server program. The program is a simple file (solution)
server that makes a collection of files available for transmission to clients. When the server
starts up, it needs to know the name of the directory that contains the collection of files.
This information can be provided as a command-line argument. You can assume that the
directory contains only regular files (that is, it does not contain any sub-directories). You
can also assume that all the files are text files.
When a client connects to the server, the server first reads a one-line command from
the client. The command can be the string “INDEX”. In this case, the server responds by
sending a list of names of all the files that are available on the server. Or the command can
be of the form “GET <filename>”, where <filename> is a file name. The server checks
whether the requested file actually exists. If so, it first sends the word “OK” as a message
to the client. Then it sends the contents of the file and closes the connection. Otherwise,
it sends a line beginning with the word “ERROR” to the client and closes the connection.
(The error response can include an error message on the rest of the line.)
Your program should use a subroutine to handle each request that the server receives.
It should not stop after handling one request; it should remain open and continue to
accept new requests. See the DirectoryList example in Subsection 11.2.2 for help with the
problem of getting the list of files in the directory.
4. Write a client program for the server from Exercise 11.3. Design a user interface that will (solution)
let the user do at least two things: (1) Get a list of files that are available on the server
and display the list on standard output; and (2) Get a copy of a specified file from the
server and save it to a local file (on the computer where the client is running).
Exercises 598
6. The sample program Checkers.java from Subsection 7.5.3 lets two players play checkers. (solution)
It would be nice if, in the middle of a game, the state of the game could be saved to a file.
Later, the file could be read back into the file to restore the game and allow the players to
continue. Add the ability to save and load files to the checkers program. Design a simple
text-based format for the files. Here is a picture of my solution to this exercise, showing
that Load and Save buttons have been added:
It’s a little tricky to completely restore the state of a game. The program has a
variable board of type CheckersData that stores the current contents of the board, and it
has a variable currentPlayer of type int that indicates whether Red or Black is currently
moving. This data must be stored in the file when a file is saved. When a file is read into the
program, you should read the data into two local variables newBoard of type CheckersData
and newCurrentPlayer of type int. Once you have successfully read all the data from
the file, you can use the following code to set up the program state correctly. This code
assumes that you have introduced two new variables saveButton and loadButton of type
Button to represent the “Save Game” and “Load Game” buttons:
Exercises 599
Quiz on Chapter 11
(answers)
1. In Java, input/output is done using I/O streams. I/O streams are an abstraction. Explain
what this means and why it is important.
2. Java has two types of I/O stream: character streams and byte streams. Why? What is
the difference between the two types of streams?
5. The package java.io includes a class named URL. What does an object of type URL
represent, and how is it used?
8. What is a socket?
11. What is it about XML that makes it suitable for representing almost any type of data?
12. Write a complete program that will display the first ten lines from a text file. The lines
should be written to standard output, System.out. The file name is given as the command-
line argument args[0]. You can assume that the file contains at least ten lines. Don’t
bother to make the program robust. Do not use TextIO to process the file; read from the
file using methods covered in this chapter.
Chapter 12
In the classic programming model, there is a single central processing unit that reads
instructions from memory and carries them out, one after the other. The purpose of a program
is to provide the list of instructions for the processor to execute. This is the only type of
programming that we have considered so far.
However, this model of programming has limitations. Modern computers have multiple
processors, making it possible for them to perform several tasks at the same time. To use the
full potential of all those processors, you will need to write programs that can do parallel
processing . For Java programmers, that means learning about threads. A single thread is
similar to the programs that you have been writing up until now, but more than one thread
can be running at the same time, “in parallel.” What makes things more interesting—and
more difficult—than single-threaded programming is the fact that the threads in a parallel
program are rarely completely independent of one another. They usually need to cooperate
and communicate. Learning to manage and control cooperation among threads is the main
hurdle that you will face in this chapter.
There are several reasons to use parallel programming. One is simply to do computations
more quickly by setting several processors to work on them simultaneously. Just as important,
however, is to use threads to deal with “blocking” operations, where a process can’t proceed
until some event occurs. In the previous chapter, for example, we saw how programs can block
while waiting for data to arrive over a network connection. Threads make it possible for one
part of a program to continue to do useful work even while another part is blocked, waiting for
some event to occur. In this context, threads are a vital programming tool even for a computer
that has only a single processing unit.
As Java has developed, new features have been added to the language that make it possible
to do parallel programming without using threads directly. We have already seen one of these:
parallel streams in the stream API from Section 10.6. We will encounter several more in this
chapter. But to use these higher level language features safely and efficiently, you still need to
know something about the hazards and benefits of threads.
601
CHAPTER 12. THREADS AND MULTIPROCESSING 602
have more than one processing unit, and such computers can literally work on several tasks
simultaneously. It is likely that from now on, most of the increase in computing power will
come from adding additional processors to computers rather than from increasing the speed of
individual processors. To use the full power of these multiprocessing computers, a programmer
must do parallel programming , which means writing a program as a set of several tasks that
can be executed simultaneously. Even on a single-processor computer, parallel programming
techniques can be useful, since some problems can be tackled most naturally by breaking the
solution into a set of simultaneous tasks that cooperate to solve the problem.
In Java, a single task is called a thread . The term “thread” refers to a “thread of control” or
“thread of execution,” meaning a sequence of instructions that are executed one after another—
the thread extends through time, connecting each instruction to the next. In a multithreaded
program, there can be many threads of control, weaving through time in parallel and forming
the complete fabric of the program. (Ok, enough with the metaphor, already!) Every Java
program has at least one thread; when the Java virtual machine runs your program, it creates a
thread that is responsible for executing the main routine of the program. This main thread can
in turn create other threads that can continue even after the main thread has terminated. In a
GUI program, there is at least one additional thread, which is responsible for handling events
and drawing components on the screen. In JavaFX, that thread is the application thread, which
does all the work involved with handling events and drawing components on the screen.
Unfortunately, parallel programming is even more difficult than ordinary, single-threaded
programming. When several threads are working together on a problem, a whole new category
of errors is possible. This just means that techniques for writing correct and robust programs are
even more important for parallel programming than they are for normal programming. On the
other hand, fortunately, Java has a nice thread API that makes basic uses of threads reasonably
easy. It also has a variety of standard classes to help with some of the more tricky parts or to
hide them entirely. It won’t be until midway through Section 12.3 that you’ll learn about the
low-level techniques that are necessary to handle the trickiest parts of parallel programming.
In fact, a programmer can do a lot with threads without ever learning about the low-level stuff.
public void run() { // The run method prints a message to standard output.
System.out.println("Greetings from thread ’" + name + "’!");
}
}
To use a NamedThread, you must of course create an object belonging to this class. For
example,
NamedThread greetings = new NamedThread("Fred");
However, creating the object does not automatically start the thread running or cause its run()
method to be executed. To do that, you must call the start() method in the thread object.
For the example, this would be done with the statement
greetings.start();
The purpose of the start() method is to create the new thread of control that will execute the
Thread object’s run() method. The new thread runs in parallel with the thread in which the
start() method was called, along with any other threads that already existed. The start()
method returns immediately after starting the new thread of control, without waiting for
the thread to terminate. This means that the code in the thread’s run() method executes
at the same time as the statements that follow the call to the start() method. Consider this
code segment:
NamedThread greetings = new NamedThread("Fred");
greetings.start();
System.out.println("Thread has been started");
After greetings.start() is executed, there are two threads. One of them will print “Thread
has been started” while the other one wants to print “Greetings from thread ’Fred’ !”. It is
important to note that these messages can be printed in either order. The two threads run
simultaneously and will compete for access to standard output, so that they can print their
messages. Whichever thread happens to be the first to get access will be the first to print its
message. In a normal, single-threaded program, things happen in a definite, predictable order
from beginning to end. In a multi-threaded program, there is a fundamental indeterminacy.
You can’t be sure what order things will happen in. This indeterminacy is what makes parallel
programming so difficult!
Note that calling greetings.start() is very different from calling greetings.run().
Calling greetings.run() would execute the run() method in the same thread, rather than
creating a new thread. This means that all the work of the run() method will be done before
the computer moves on to the statements that follow the call to greetings.run(). There is
no parallelism and no indeterminacy.
CHAPTER 12. THREADS AND MULTIPROCESSING 604
return start
thread
start
thread
call
subroutine
END
END
return
END
This discussion has assumed that the computer on which the program is running has more
than one processing unit, so that it is possible for the original thread and the newly created
thread to literally be executed at the same time. However, it’s possible to create multiple
threads even on a computer that has only one processor (and, more generally, it is possible to
create many more threads than there are processors, on any computer). In that case, the two
threads will compete for time on the processor. However, there is still indeterminacy because
the processor can switch from one thread to another at unpredictable times. In fact, from
the point of view of the programmer, there is no difference between programming for a single-
processor computer and programming for a multi-processor computer, and we will mostly ignore
the distinction from now on.
∗ ∗ ∗
I mentioned that there are two ways to program a thread. The first way was to de-
fine a subclass of Thread. The second is to define a class that implements the interface
java.lang.Runnable. The Runnable interface defines a single method, public void run().
Given a Runnable, it is possible to create a Thread whose task is to execute the Runnable’s
run() method.
The Thread class has a constructor that takes a Runnable as its parameter. When an object
that implements the Runnable interface is passed to that constructor, the run() method of the
thread will simply call the run() method from the Runnable, and calling the thread’s start()
method will create a new thread of control in which the Runnable’s run() method is executed.
For example, as an alternative to the NamedThread class, we could define the class:
public class NamedRunnable implements Runnable {
private String name; // The name of this Runnable.
public NamedRunnable(String name) { // Constructor gives name to object.
this.name = name;
}
public void run() { // The run method prints a message to standard output.
System.out.println("Greetings from runnable ’" + name +"’!");
}
}
To use this version of the class, we would create a NamedRunnable object and use that object
to create an object of type Thread:
CHAPTER 12. THREADS AND MULTIPROCESSING 605
∗ ∗ ∗
To help you understand how multiple threads are executed in parallel, we consider the
sample program ThreadTest1.java. This program creates several threads. Each thread performs
exactly the same task. The task is to count the number of integers less than 5000000 that are
prime. (The particular task that is done is not important for our purposes here, as long as it
is something that takes a non-trivial amount of time. This is a demo program; it would be
silly in a real program to have multiple threads that do the same thing, and the method that
is used for counting the primes is not efficient.) This computation should take no more than
a few seconds on a modern computer. The threads that perform this task are defined by the
following static nested class:
/**
* When a thread belonging to this class is run it will count the
* number of primes between 2 and 5000000. It will print the result
* to standard output, along with its ID number and the elapsed
* time between the start and the end of the computation.
*/
private static class CountPrimesThread extends Thread {
int id; // An id number for this thread; specified in the constructor.
public CountPrimesThread(int id) {
this.id = id;
}
public void run() {
long startTime = System.currentTimeMillis();
int count = countPrimes(2,5000000); // Counts the primes.
long elapsedTime = System.currentTimeMillis() - startTime;
System.out.println("Thread " + id + " counted " +
count + " primes in " + (elapsedTime/1000.0) + " seconds.");
}
}
The main program asks the user how many threads to run, and then creates and starts the
specified number of threads:
CHAPTER 12. THREADS AND MULTIPROCESSING 606
The second line was printed immediately after the first. At this point, the main program
has ended but the eight threads continued to run. After a pause of about 12.5 seconds, all eight
threads completed at about the same time. The order in which the threads complete is not the
same as the order in which they were started, and the order is indeterminate. That is, if the
program is run again, the order in which the threads complete will probably be different.
On this computer, eight threads took about twice as long as one thread. This is because
the computer has four processors. With eight threads running on four processors—one-half
processor per thread—each thread was only actively being executed for about half of the time,
so it took twice as long to finish its task. On a single-processor computer, eight threads would
take about eight times as long as one thread. On a computer with eight or more processors,
eight threads might take no more time than a single thread. Because of overhead and other
reasons, the actual speedup will probably be a little smaller than this analysis indicates, but
on a multiprocessor machine, you should see a definite speedup. What happens when you run
the program on your own computer? How many processors do you have?
Whenever there are more threads to be run than there are processors to run them, the
computer divides its attention among all the runnable threads by switching rapidly from one
thread to another. That is, each processor runs one thread for a while then switches to another
thread and runs that one for a while, and so on. Typically, these “context switches” occur
CHAPTER 12. THREADS AND MULTIPROCESSING 607
about 100 times or more per second. The result is that the computer makes progress on all
the tasks, and it looks to the user as if all the tasks are being executed simultaneously. This
is why in the sample program, in which each thread has the same amount of work to do, all
the threads complete at about the same time: Over any time period longer than a fraction of
a second, the computer’s time is divided approximately equally among all the threads.
that executes the method—has been interrupted. It also has the unusual property of clearing
the interrupted status of the thread, so you only get one chance to check for an interruption.
In your own programs, your threads are not going to be interrupted unless you interrupt them.
So most often, you are not likely to need to do anything in response to an InterruptedException
(except to catch it).
Sometimes, it’s necessary for one thread to wait for another thread to die. This is done with
the join() method from the Thread class. Suppose that thrd is a Thread. Then, if another
thread calls thrd.join(), that other thread will go to sleep until thrd terminates. If thrd is
already dead when thrd.join() is called, then it simply has no effect. The join() method can
throw an InterruptedException, which must be handled as usual. As an example, the following
code starts several threads, waits for them all to terminate, and then outputs the elapsed time:
CountPrimesThread[] worker = new CountPrimesThread[numberOfThreads];
long startTime = System.currentTimeMillis();
for (int i = 0; i < numberOfThreads; i++) {
worker[i] = new CountPrimesThread();
worker[i].start();
}
for (int i = 0; i < numberOfThreads; i++) {
try {
worker[i].join(); // Wait until worker[i] finishes, if it hasn’t already.
}
catch (InterruptedException e) {
}
}
// At this point, all the worker threads have terminated.
long elapsedTime = System.currentTimeMillis() - startTime;
System.out.println("Total elapsed time: " + (elapsedTime/1000.0) + " seconds");
An observant reader will note that this code assumes that no InterruptedException will occur.
To be absolutely sure that the thread worker[i] has terminated in an environment where
InterruptedExceptions are possible, you would have to do something like:
while (worker[i].isAlive()) {
try {
worker[i].join();
}
catch (InterruptedException e) {
}
}
Another version of the join() method takes an integer parameter that specifies the maxi-
mum number of milliseconds to wait. A call to thrd.join(m) will wait until either thrd has
terminated or until m milliseconds have elapsed (or until the waiting thread is interrupted).
This can be used to allow a thread to wake up occasionally to perform some task while it is
waiting. Here, for example, is a code segment that will start a thread, thrd, and then will
output a period every two seconds as long as thrd continues to run:
System.out.print("Running the thread ");
thrd.start();
while (thrd.isAlive()) {
try {
thrd.join(2000);
System.out.print(".");
CHAPTER 12. THREADS AND MULTIPROCESSING 609
}
catch (InterruptedException e) {
}
}
System.out.println(" Done!");
∗ ∗ ∗
Threads have two properties that are occasionally useful: a daemon status and a priority.
A Thread thrd can be designated as a daemon thread by calling thrd.setDaemon(true).
This must be done before the thread is started, and it can throw an exception of type Secu-
rityException if the calling thread is not allowed to modify thrd’s properties. This has only
one effect: The Java Virtual Machine will exit as soon as there are no non-daemon threads
that are still alive. That is, the fact that a daemon thread is still alive is not enough to keep
the Java Virtual Machine running. A daemon thread might exist, for example, only to provide
some service to other, non-daemon threads. When there are no more non-daemon threads,
there will be no further call for the daemon thread’s services, so the program might as well shut
down. (A call to System.exit() forces the JVM to terminate, even if there are non-daemon
threads still running.)
The priority of a thread is a more important property. Every thread has a priority , specified
as an integer. A thread with a greater priority value will be run in preference to a thread with a
smaller priority. For example, computations that can be done in the background, when no more
important thread has work to do, can be run with a low priority. In the next section, we will see
how this can be useful in GUI programs. If thrd is of type Thread, then thrd.getPriority()
returns the integer that specifies thrd’s priority, and thrd.setPriority(p) can be used to set
its priority to a given integer, p.
Priorities cannot be arbitrary integers, and thrd.setPriority() will throw an IllegalArgu-
mentException if the specified priority is not in the legal range for the thread. The range of legal
priority values can differ from one computer to another. The range of legal values is specified
by the constants Thread.MIN PRIORITY and Thread.MAX PRIORITY, but a given thread might
be further restricted to values less than Thread.MAX PRIORITY. The default priority is given
by Thread.NORM PRIORITY. To set thrd to run with a priority value just below the normal
priority, you can call
thrd.setPriority( Thread.NORM PRIORITY - 1 );
Note that thrd.setPriority() can also throw an exception of type SecurityException, if
the thread that calls the method is not allowed to set the priority of thrd to the specified value.
Finally, I’ll note that the static method Thread.currentThread() returns the current
thread. That is, the return value of this method is the thread that executed the method.
This allows a thread to get a reference to itself, so that it can modify its own proper-
ties. For example, you can determine the priority of the currently running thread by calling
Thread.currentThread().getPriority().
sharing the use of just one measuring cup, and imagine that Cook A fills the measuring cup
with milk, only to have Cook B grab the cup before Cook A has a chance to empty the milk
into his bowl. There has to be some way for Cook A to claim exclusive rights to the cup while
he performs the two operations: Add-Milk-To-Cup and Empty-Cup-Into-Bowl.
Something similar happens with threads, even with something as simple as adding one to a
counter. The statement
count = count + 1;
is actually a sequence of three operations:
Step 1. Get the value of count
Step 2. Add 1 to the value.
Step 3. Store the new value in count
Suppose that each of several threads performs these three steps. Remember that it’s possible
for two threads to run at the same time, and even if there is only one processor, it’s possible
for that processor to switch from one thread to another at any point. Suppose that while one
thread is between Step 2 and Step 3, another thread starts executing the same sequence of
steps. Since the first thread has not yet stored the new value in count, the second thread
reads the old value of count and adds one to that old value. Both threads have computed the
same new value for count, and both threads then go on to store that value back into count
by executing Step 3. After both threads have done so, the value of count has gone up only
by 1 instead of by 2! This type of problem is called a race condition. This occurs when one
thread is in the middle of a multi-step operation, and another thread can change some value or
condition that the first thread is depending upon. (The first thread is “in a race” to complete
all the steps before it is interrupted by another thread.)
Another example of a race condition can occur in an if statement. Consider the following
statement, which is meant to avoid a division-by-zero error:
if ( A != 0 ) {
B = C / A;
}
Suppose that this statement is executed by some thread. If the variable A is shared by one
or more other threads, and if nothing is done to guard against the race condition, then it is
possible that one of those other threads will change the value of A to zero between the time
that the first thread checks the condition A != 0 and the time that it does the division. This
means that the thread can end up dividing by zero, even though it just checked that A was not
zero!
To fix the problem of race conditions, there has to be some way for a thread to get exclusive
access to a shared resource. This is not a trivial thing to implement, but Java provides a high-
level and relatively easy-to-use approach to exclusive access. It’s done with synchronized
methods and with the synchronized statement. These are used to protect shared resources
by making sure that only one thread at a time will try to access the resource. Synchronization in
Java actually provides only mutual exclusion, which means that exclusive access to a resource
is only guaranteed if every thread that needs access to that resource uses synchronization.
Synchronization is like a cook leaving a note that says, “I’m using the measuring cup.” This
will get the cook exclusive access to the cup—but only if all the cooks agree to check the note
before trying to grab the cup.
Because this is a difficult topic, I will start with a simple example. Suppose that we want
to avoid the race condition that occurs when several threads all want to add 1 to a counter. We
CHAPTER 12. THREADS AND MULTIPROCESSING 611
can do this by defining a class to represent the counter and by using synchronized methods in
that class. A method is declared to be synchronized by adding the reserved word synchronized
as a modifier to the definition of the method:
public class ThreadSafeCounter {
private int count = 0; // The value of the counter.
synchronized public void increment() {
count = count + 1;
}
synchronized public int getValue() {
return count;
}
}
If tsc is of type ThreadSafeCounter, then any thread can call tsc.increment() to add 1 to the
counter in a completely safe way. The fact that tsc.increment() is synchronized means that
only one thread can be in this method at a time; once a thread starts executing this method, it
is guaranteed that it will finish executing it before another thread is allowed to access count.
There is no possibility of a race condition. Note that the guarantee depends on the fact that
count is a private variable. This forces all access to tsc.count to occur in the synchronized
methods that are provided by the class. If count were public, it would be possible for a
thread to bypass the synchronization by, for example, saying tsc.count++. This could change
the value of count while another thread is in the middle of tsc.increment(). Remember
that synchronization by itself does not guarantee exclusive access; it only guarantees mutual
exclusion among all the threads that are synchronized.
However, the ThreadSafeCounter class does not prevent all possible race conditions that
might arise when using a counter. Consider the if statement:
if ( tsc.getValue() == 10 ) {
doSomething();
}
where doSomething() is some method that requires the value of the counter to be 10. There is
still a race condition here, which occurs if a second thread increments the counter between the
time the first thread tests tsc.getValue() == 10 and the time it executes doSomething().
The first thread needs exclusive access to the counter during the execution of the whole if
statement. (The synchronization in the ThreadSafeCounter class only gives it exclusive access
during the time it is evaluating tsc.getValue().) We can solve the race condition by putting
the if statement in a synchronized statement:
synchronized(tsc) {
if ( tsc.getValue() == 10 )
doSomething();
}
Note that the synchronized statement takes an object—tsc in this case—as a kind of param-
eter. The syntax of the synchronized statement is:
synchronized( hobject i ) {
hstatements i
}
CHAPTER 12. THREADS AND MULTIPROCESSING 612
In Java, mutual exclusion is always associated with an object; we say that the synchronization
is “on” that object. For example, the if statement above is “synchronized on tsc.” A synchro-
nized instance method, such as those in the class ThreadSafeCounter, is synchronized on the
object that contains the instance method. In fact, adding the synchronized modifier to the
definition of an instance method is pretty much equivalent to putting the body of the method
in a synchronized statement of the form synchronized(this) {...}. It is also possible to
have synchronized static methods; a synchronized static method is synchronized on the special
class object that represents the class containing the static method.
The real rule of synchronization in Java is this: Two threads cannot be synchronized
on the same object at the same time; that is, they cannot simultaneously be executing
code segments that are synchronized on that object. If one thread is synchronized on an object,
and a second thread tries to synchronize on the same object, the second thread is forced to
wait until the first thread has finished with the object. This means that it is not only true
that two threads cannot be executing the same synchronized method at the same time, but
in fact two threads cannot be executing two different methods at the same time, if the two
methods are synchronized on the same object. This is implemented using something called a
synchronization lock . Every object has a synchronization lock, and that lock can be “held”
by only one thread at a time. To enter a synchronized statement or synchronized method, a
thread must obtain the associated object’s lock. If the lock is available, then the thread obtains
the lock and immediately begins executing the synchronized code. It releases the lock after it
finishes executing the synchronized code. If Thread A tries to obtain a lock that is already held
by Thread B, then Thread A has to wait until Thread B releases the lock. In fact, Thread A
will go to sleep, and will not be awoken until the lock becomes available.
The discussion of invariants in Subsection 8.2.3 mentioned that reasoning about invariants
becomes much more complicated when threads are involved. The problem is race conditions. We
would like our ThreadSafeCounter class to have the class invariant that “the value of count is the
number of times that increment() has been called.” In a single-threaded program, that would
be true even without synchronization. However, in a multithreaded program, synchronization
is needed to ensure that the class invariant is really invariant.
∗ ∗ ∗
As a simple example of shared resources, we return to the prime-counting problem. In this
case, instead of having every thread perform exactly the same task, we’ll do some real parallel
processing. The program will count the prime numbers in a given range of integers, and it will
do so by dividing the work up among several threads. Each thread will be assigned a part of
the full range of integers, and it will count the primes in its assigned part. At the end of its
computation, the thread has to add its count to the overall total of primes in the entire range.
The variable that represents the total is shared by all the threads, since each thread has to add
a number to the total. If each thread just says
total = total + count;
then there is a (small) chance that two threads will try to do this at the same time and that the
final total will be wrong. To prevent this race condition, access to total has to be synchronized.
My program uses a synchronized method to add the counts to the total. This method is called
once by each thread, and it is the only method in which the value of total is changed:
synchronized private static void addToTotal(int x) {
total = total + x;
System.out.println(total + " primes found so far.");
}
CHAPTER 12. THREADS AND MULTIPROCESSING 613
The source code for the program can be found in ThreadTest2.java. This program counts
the primes in the range 3000001 to 6000000. (The numbers are rather arbitrary.) The main()
routine in this program creates between 1 and 5 threads and assigns part of the job to each
thread. It waits for all the threads to finish, using the join() method as described above. It
then reports the total number of primes found, along with the elapsed time. Note that join()
is required here, since it doesn’t make sense to report the number of primes until all of the
threads have finished. If you run the program on a multiprocessor computer, it should take less
time for the program to run when you use more than one thread.
∗ ∗ ∗
Synchronization can help to prevent race conditions, but it introduces the possibility of
another type of error, deadlock . A deadlock occurs when a thread waits forever for a resource
that it will never get. In the kitchen, a deadlock might occur if two very simple-minded cooks
both want to measure a cup of milk at the same time. The first cook grabs the measuring cup,
while the second cook grabs the milk. The first cook needs the milk, but can’t find it because
the second cook has it. The second cook needs the measuring cup, but can’t find it because
the first cook has it. Neither cook can continue and nothing more gets done. This is deadlock.
Exactly the same thing can happen in a program, for example if there are two threads (like the
two cooks) both of which need to obtain locks on the same two objects (like the milk and the
measuring cup) before they can proceed. Deadlocks can easily occur, unless great care is taken
to avoid them.
In a program, an operation can be atomic even if it is not literally atomic on the machine
language level. An operation can be considered atomic if it can never be seen by any thread as
partly completed. For example, we can say that the ThreadSafeCounter class that was defined
above has an atomic increment operation. Synchronization can be seen as a way of ensuring
that operations are, at least effectively, atomic.
Still, it would be nice to have atomic operations that don’t require synchronization, espe-
cially since such operations might be implemented very efficiently on the hardware level.
Java has a package java.util.concurrent.atomic that defines classes that implement
atomic operations on several simple variable types. We will look at the class AtomicInteger,
which defines some atomic operations on an integer value, including atomic add, increment,
and decrement. Suppose, for example, that we want to add some integer values that are being
produced by a number of different threads. We can do that with an AtomicInteger such as
private static AtomicInteger total = new AtomicInteger();
The total is created with an initial value of zero. When a thread wants to add a value to
the total, it can use the method total.addAndGet(x), which adds x to the total and returns
the new value of total after x has been added. This is an atomic operation, which cannot be
interrupted, so we can be sure that the value of total will be correct in the end. The sample
program ThreadTest3.java is a small variation on ThreadTest2.java that uses an AtomicInteger
instead of synchronization to safely add up values from several threads.
AtomicInteger has similar methods for adding one to the total and subtracting one
from the total: total.incrementAndGet() and total.decrementAndGet(). The method
total.getAndSet(x) sets the value of the total to x and returns the previous value that x
is replacing. All of these operations are done atomically (either because they use atomic ma-
chine language instructions or because they use synchronization internally).
I should close with my usual warning: Using an atomic variable does not automatically
solve all race conditions involving that variable. For example, in the code
int currentTotal = total.addAndGet(x);
System.out.println("Current total is " + currentTotal);
it is possible that by the time the output statement is executed, the total has been changed by
another thread so that currentTotal is no longer the current value of total!
Suppose that we wanted to do something similar with a thread. That is, we would like to
call some subroutine at periodic intervals, say 30 times per second. The run() method of the
thread would have to execute a loop in which the thread sleeps for 30 milliseconds, then wakes
up to call the subroutine. This could be implemented in a nested class as follows using the
method Thread.sleep() that was discussed in Subsection 12.1.2:
private class Animator extends Thread {
public void run() {
while (true) {
try {
Thread.sleep(30);
}
catch (InterruptedException e) {
}
callSubroutine();
}
}
}
To use this class, you would create an object belonging to it and call its start() method. As
it stands, there would be no way to stop the thread once it is started. One way to make that
possible would be to end the loop when a volatile boolean variable, terminate, becomes
true, as discussed in Subsection 12.1.4. Since thread objects can only be executed once, in
order to restart the animation after it has been stopped in this way, it would be necessary to
create a new thread. In the next section, we’ll see some more versatile techniques for controlling
threads.
There is a subtle difference between using threads and using timers for animation. The
thread that is used by a JavaFX AniamtionTimer does nothing but call its handle() routine
repeatedly. That method is actually executed in the JavaFX application thread, which also
handles repainting of components and responses to user actions. This is important because
JavaFX is not thread-safe. That is, it does not use synchronization to avoid race conditions
among threads trying to access GUI components and their state variables. As long as everything
is done in the application thread, there is no problem. A problem can arise when another thread
tries to manipulate components or the variables that are also used in the GUI thread. Using
syncrhonization would be a solution in some cases. The best solution is probably to use an
AnimationTimer, if that is possible. But if you really need to use a separate thread, you might
be able to use Platform.runLater().
Platform.runLater(r) is a static method in class Platform, from package javafx.application.
The parameter is an object of type Runnable, the same interface that is used when creating
threads. Platform.runLater(r) can be called from any thread. Its purpose is to submit r to
be run on the JavaFX application thread. Platform.runLater(r) returns immediately, with-
out waiting for r to be run; r.run() will be called by the application thread at some future
time (which means within a fraction of a second, and perhaps almost immediately, unless the
computer is excessively busy). Runnables are executed in the order in which they are submitted.
Since the Runnable is called on the application thread, it can safely operate on the GUI, without
synchronization. It is often convenient to specify the parameter to Platform.runLater() as a
lambda expression of type Runnable. I will use Platform.runLater() in several examples in
this chapter and the next.
As an example, the sample program RandomArtWithThreads.java uses a thread to drive
a very simple animation. In this example, the thread does nothing except to call a redraw()
CHAPTER 12. THREADS AND MULTIPROCESSING 617
method every two seconds. The method redraws the content of a canvas. Platform.runLater()
is used to execute redraw() on the application thread. The user can click a button to start and
stop the animation. A new thread is created each time the animation is started. A volatile
boolean variable, running, is set to false when the user stops the animation, as a signal to the
thread to stop, as discussed in Subsection 12.1.4. The thread is defined by the following class:
private class Runner extends Thread {
public void run() {
while (running) {
Platform.runLater( () -> redraw() );
try {
Thread.sleep(2000); // Wait two seconds between redraws.
}
catch (InterruptedException e) {
}
}
}
}
value of the signal variable, running. If running is false, the delay() method throws the
exception that will cause the recursive algorithm, and eventually the animation thread itself,
to terminate. Here, then, is the delay() method:
private void delay(int millis) {
if (! running)
throw new ThreadTerminationException();
try {
Thread.sleep(millis);
}
catch (InterruptedException e) {
}
if (! running) // Check again, in case it changed during the sleep period.
throw new ThreadTerminationException();
}
The ThreadTerminationException is caught in the thread’s run() method:
/**
* This class defines the thread that runs the recursive
* QuickSort algorithm. The thread begins by randomizing the
* hue array. It then calls quickSort() to sort the entire array.
* If quickSort() is aborted by a ThreadTerminationException,
* which would be caused by the user clicking the Finish button,
* then the thread will restore the array to sorted order before
* terminating, so that whether or not the quickSort is aborted,
* the array ends up sorted. In any case, in the end, it
* resets the text on the button to "Start".
*/
private class Runner extends Thread {
public void run() {
for (int i = 0; i < hue.length; i++) {
// fill hue array with indices in order
hue[i] = i;
}
for (int i = hue.length-1; i > 0; i--) {
// Randomize the order of the hues.
int r = (int)((i+1)*Math.random());
int temp = hue[r];
hue[r] = hue[i];
// The last assignment that needs to be done in this
// loop is hue[i] = temp. The value of hue[i] will
// not change after this, so the assignment is done
// by calling a method setHue(i,temp) that will change
// the value in the array and also use Platform.runLater()
// to change the color of the i-th color bar in the canvas.
setHue(i,temp);
}
try {
delay(1000); // Wait one second before starting the sort.
quickSort(0,hue.length-1); // Sort the whole array.
}
catch (ThreadTerminationException e) { // User aborted quickSort.
// Put the colors back into sorted order. The drawSorted()
// method draws all of the color bars in sorted order.
CHAPTER 12. THREADS AND MULTIPROCESSING 619
the animation thread is sleeping between frames for the JavaFX application thread to do any
necessary redrawing of the GUI and to handle any other events.
When a thread is used for background computation, however, we want to keep the computer
as busy as possible working on the computation. The thread will compete for processor time
with the application thread; if you are not careful, event-handling—redrawing in particular—
can still be delayed. Fortunately, you can use thread priorities to avoid the problem. By setting
the computation thread to run at a lower priority than the event-handling thread, you make
sure that events will be processed as quickly as possible, while the computation thread will get
all the extra processing time. Since event handling generally uses very little time, this means
that most of the processing time goes to the background computation, but the interface is still
very responsive. (Thread priorities were discussed in Subsection 12.1.2.)
The sample program BackgroundComputationDemo.java is an example of background pro-
cessing. This program creates an image that takes a while to compute because it takes some
computation to compute the color of each pixel in the image. The image itself is a piece of
a mathematical object known as the Mandelbrot set. We will use the same image in several
examples in this chapter.
In outline, BackgroundComputationDemo is similar to the QuicksortThreadDemo discussed
above. The computation is done in a thread defined by a nested class, Runner. A volatile
boolean variable, running, is used to control the thread: If the value of running is set to
false, the thread should terminate. The sample program has a button that the user clicks
to start and to abort the computation. The difference is that the thread in this case is meant
to run continuously, without sleeping. To allow the user to see that progress is being made in
the computation (always a good idea), every time the thread computes a row of pixels, it uses
Platform.runLater() to copy those pixels to the image that is shown on the screen. The
user sees the image being built up line-by-line.
When the computation thread is created in response to the “Start” button, we need to set
it to run at a priority lower than the JavaFX application thread. The code that creates the
thread is itself running in the application thread, so we can use a priority that is one less than
the priority of the thread that is executing the code. Note that the priority is set inside a
try..catch statement. If an error occurs while trying to set the thread priority, the program
will still work, though perhaps not as smoothly as it would if the priority was correctly set.
Here is how the thread is created and started:
runner = new Runner();
try {
runner.setPriority( Thread.currentThread().getPriority() - 1 );
}
catch (Exception e) {
System.out.println("Error: Can’t set thread priority: " + e);
}
running = true; // Set the signal before starting the thread!
runner.start();
Although BackgroundComputationDemo works OK, there is one problem: The goal is to get
the computation done as quickly as possible, using all available processing time. The program
accomplishes that goal on a computer that has only one processor. But on a computer that
has several processors, we are still using only one of those processors for the computation.
(And in fact, in that case the business about thread priority is not all that relevant, because
the animation thread and the application thread can run both run simultaneously, on different
processors.) It would be nice to get all of the processors working on the computation. To
CHAPTER 12. THREADS AND MULTIPROCESSING 621
do that, we need real parallel processing, with several computation threads. We turn to that
problem next.
You should try out the program. On a multi-processor computer, the computation will complete
more quickly when using several threads than when using just one. Note that when using one
thread, this program has the same behavior as the previous example program.
The approach used in this example for dividing up the problem among threads is not optimal.
We will see in the next section how it can be improved. However, MultiprocessingDemo1 makes
a good first example of multiprocessing.
When the user clicks the “Start” button, the program has to create and start the specified
number of threads, and it has to assign a segment of the image to each thread. Here is how
this is done:
workers = new Runner[threadCount]; // Holds the computation threads.
int rowsPerThread; // How many rows of pixels should each thread compute?
rowsPerThread = height / threadCount; // (height = vertical size of image)
running = true; // Set the signal before starting the threads!
threadsCompleted = 0; // Records how many of the threads have terminated.
for (int i = 0; i < threadCount; i++) {
int startRow; // first row computed by thread number i
CHAPTER 12. THREADS AND MULTIPROCESSING 622
properly counted, and threadsRunning will never become zero. The program would hang in a
kind of deadlock. The problem would occur only very rarely, since it depends on exact timing.
But in a large program, problems of this sort can be both very serious and very hard to debug.
Proper synchronization makes the error impossible.
course, there will still be variation in the time that the various subtasks require. One processor
might complete several subproblems while another works on one particularly difficult case. And
a slow or busy processor might complete only one or two subproblems while another processor
finishes five or six. Each processor can work at its own pace. As long as the subproblems are
fairly small, most of the processors can be kept busy until near the end of the computation.
This is known as load balancing : the computational load is balanced among the available
processors in order to keep them all as busy as possible. Of course, some processors will still
finish before others, but not by longer than the time it takes to complete the longest subtask.
While the subproblems should be small, they should not be too small. There is some
computational overhead involved in creating the subproblems and assigning them to processors.
If the subproblems are very small, this overhead can add significantly to the total amount of
work that has to be done. In my example program, the task is to compute a color for each
pixel in an image. For dividing that task up into subtasks, one possibility would be to have
each subtask compute just one pixel. But the subtasks produced in that way are probably too
small. So, instead, each subtask in my program will compute the colors for one row of pixels.
Since there are several hundred rows of pixels in the image, the number of subtasks will be
fairly large, while each subtask will also be fairly large. The result is fairly good load balancing,
with a reasonable amount of overhead.
Note, by the way, that the problem that we are working on is a very easy one for parallel
programming. When we divide the problem of calculating an image into subproblems, all
the subproblems are completely independent. It is possible to work on any number of them
simultaneously, and they can be done in any order. Things get a lot more complicated when
some subtasks produce results that are required by other subtasks. In that case, the subtasks are
not independent, and the order in which the subtasks are performed is important. Furthermore,
there has to be some way for results from one subtask to be shared with other tasks. When the
subtasks are executed by different threads, this raises all the issues involved in controlling access
of threads to shared resources. So, in general, decomposing a problem for parallel processing
is much more difficult than it might appear from our relatively simple example. But for the
most part, that’s a topic for a course in parallel computing, not an introductory programming
course.
amount of time while waiting for some event or for access to some resource, you want to have
extra threads around for the processor to run while other threads are blocked. We’ll encounter
exactly that situation when we turn to using threads with networking in Section 12.4.
When several threads are available for performing tasks, those threads are called a thread
pool . Thread pools are used to avoid creating a new thread to perform each task. Instead,
when a task needs to be performed, it can be assigned to any idle thread in the “pool.”
Once all the threads in the thread pool are busy, any additional tasks will have to wait until
one of the threads becomes idle. This is a natural application for a queue: Associated with
the thread pool is a queue of waiting tasks. As tasks become available, they are added to the
queue. Every time that a thread finishes a task, it goes to the queue to get another task to
work on.
Note that there is only one task queue for the thread pool. All the threads in the pool
use the same queue, so the queue is a shared resource. As always with shared resources, race
conditions are possible and synchronization is essential. Without synchronization, for example,
it is possible that two threads trying to get items from the queue at the same time will end up
retrieving the same item. (See if you can spot the race conditions in the dequeue() method in
Subsection 9.3.2.)
Java has a built-in class to solve this problem: ConcurrentLinkedQueue. This class and others
that can be useful in parallel programming are defined in the package java.util.concurrent.
It is a parameterized class so that to create a queue that can hold objects of type Runnable,
you can say
ConcurrentLinkedQueue<Runnable> queue = new ConcurrentLinkedQueue<>();
This class represents a queue, implemented as a linked list, in which operations on the queue are
properly synchronized. The operations on a ConcurrentLinkedQueue are not exactly the queue
operations that we are used to. The method for adding a new item, x, to the end of queue is
queue.add(x). The method for removing an item from the front of queue is queue.poll().
The queue.poll() method returns null if the queue is empty; thus, poll() can be used to
test whether the queue is empty and to retrieve an item if it is not. It makes sense to do
things in this way because testing whether the queue is non-empty before taking an item from
the queue involves a race condition: Without synchronization, it is possible for another thread
to remove the last item from the queue between the time when you check that the queue is
non-empty and the time when you try to take the item from the queue. By the time you try to
get the item, there’s nothing there! On the other hand, queue.poll() is an “atomic” operation
(Subsection 12.1.5).
∗ ∗ ∗
To use ConcurrentLinkedQueue in our image-computing example, we can use the queue along
with a thread pool. To begin the computation of the image, we create all the tasks that make
up the image and add them to the queue. Then, we can create and start the worker threads that
will execute the tasks. Each thread will run in a loop in which it gets one task from the queue,
by calling the queue’s poll() method, and carries out that task. Since the task is an object of
type Runnable, it is only necessary for the thread to call the task’s run() method. When the
poll() method returns null, the queue is empty and the thread can terminate because all the
tasks have been assigned to threads.
The sample program MultiprocessingDemo2.java implements this idea. It uses a queue,
taskQueue, of type ConcurrentLinkedQueue<Runnable> to hold the tasks. In addition, in order
to allow the user to abort the computation before it finishes, it uses the volatile boolean
CHAPTER 12. THREADS AND MULTIPROCESSING 626
variable running to signal the thread when the user aborts the computation. The thread
should terminate when this variable is set to false, even if there are still tasks remaining in
the queue. The threads are defined by a nested class named WorkerThread. It is quite short
and simple to write:
private class WorkerThread extends Thread {
public void run() {
try {
while (running) {
Runnable task = taskQueue.poll(); // Get a task from the queue.
if (task == null)
break; // (because the queue is empty)
task.run(); // Execute the task;
}
}
finally {
threadFinished(); // Records fact that this thread has terminated.
// Done in finally to make sure it gets called.
}
}
}
The program uses a nested class named MandelbrotTask to represent the task of computing
one row of pixels in the image. This class implements the Runnable interface. Its run() method
does the actual work, that is, compute the color of each pixel, and apply the colors to the image.
Here is what the program does to start the computation (with a few details omitted):
taskQueue = new ConcurrentLinkedQueue<Runnable>(); // Create the queue.
for (int row = 0; row < height; row++) { // height is number of rows in image
MandelbrotTask task;
task = ... ; // Create a task to compute one row of the image.
taskQueue.add(task); // Add the task to the queue.
}
int threadCount = ... ; // Number of threads in the pool (selected by user).
workers = new WorkerThread[threadCount];
running = true; // Set the signal before starting the threads!
threadsRemaining = workers; // Records how many threads are still running.
for (int i = 0; i < threadCount; i++) {
workers[i] = new WorkerThread();
try {
workers[i].setPriority( Thread.currentThread().getPriority() - 1 );
}
catch (Exception e) {
}
workers[i].start();
}
Note that it is important that the tasks be added to the queue before the threads are started.
The threads see an empty queue as a signal to terminate. If the queue is empty when the
threads are started, they might see an empty queue and terminate immediately after being
started, without performing any tasks!
You should try out MultiprocessingDemo2. It computes the same image as
MultiprocessingDemo1, but the rows of pixels are not computed in the same order as in
CHAPTER 12. THREADS AND MULTIPROCESSING 627
that program (if there is more than one thread). If you look carefully, you might see that the
rows of pixels are not added to the image in strict order from top to bottom. This is because
it is possible for one thread to finish row number i+1 while another thread is still working on
row i, or even earlier rows. (The effect might be more apparent if you use more threads than
you have processors. Try it with 20 threads.)
Consumer
3
Producer
1 Consumer
Blocking Queue
3
Producer
2 Consumer
3
Producer
3 Consumer
3
We are talking parallel processing, so we need a synchronized queue, but we need more than
that. When the queue is empty, we need a way to have consumers wait until an item appears in
the queue. If the queue becomes full, we need a way to have producers wait until a space opens
up in the queue. In our application, the producers and consumers are threads. A thread that
is suspended, waiting for something to happen, is said to be blocked, and the type of queue
that we need is called a blocking queue. In a blocking queue, the operation of dequeueing an
item from the queue can block if the queue is empty. That is, if a thread tries to dequeue an
item from an empty queue, the thread will be suspended until an item becomes available; at
that time, it will wake up, retrieve the item, and proceed. Similarly, if the queue has a limited
capacity, a producer that tries to enqueue an item can block if there is no space in the queue.
CHAPTER 12. THREADS AND MULTIPROCESSING 628
Java has two classes that implement blocking queues: LinkedBlockingQueue and ArrayBlock-
ingQueue. These are parameterized types to allow you to specify the type of item that the queue
can hold. Both classes are defined in the package java.util.concurrent and both implement
an interface called BlockingQueue. If bqueue is a blocking queue belonging to one of these
classes, then the following operations are defined:
• bqueue.take() – Removes an item from the queue and returns it. If the queue is empty
when this method is called, the thread that called it will block until an item becomes
available. This method throws an InterruptedException if the thread is interrupted while
it is blocked.
• bqueue.put(item) – Adds the item to the queue. If the queue has a limited capacity
and is full, the thread that called it will block until a space opens up in the queue. This
method throws an InterruptedException if the thread is interrupted while it is blocked.
• bqueue.add(item) – Adds the item to the queue, if space is available. If the queue has
a limited capacity and is full, an IllegalStateException is thrown. This method does not
block.
• bqueue.clear() – Removes all items from the queue and discards them.
Java’s blocking queues define many additional methods (for example, bqueue.poll(500) is
similar to bqueue.take(), except that it will not block for longer than 500 milliseconds), but
the four listed here are sufficient for our purposes. Note that I have listed two methods for adding
items to the queue: bqueue.put(item) blocks if there is not space available in the queue and is
most appropriate for use with blocking queues that have a limited capacity; bqueue.add(item)
does not block and is most appropriate for use with blocking queues that have an unlimited
capacity.
An ArrayBlockingQueue has a maximum capacity that is specified when it is constructed.
For example, to create a blocking queue that can hold up to 25 objects of type ItemType, you
could say:
ArrayBlockingQueue<ItemType> bqueue = new ArrayBlockingQueue<>(25);
With this declaration, bqueue.put(item) will block if bqueue already contains 25 items, while
bqueue.add(item) will throw an exception in that case. Recall that this ensures that items are
not produced indefinitely at a rate faster than they can be consumed. A LinkedBlockingQueue
is meant for creating blocking queues with unlimited capacity. For example,
LinkedBlockingQueue<ItemType> bqueue = new LinkedBlockingQueue<>();
creates a queue with no upper limit on the number of items that it can contain. In
this case, bqueue.put(item) will never block and bqueue.add(item) will never throw an
IllegalStateException. You would use a LinkedBlockingQueue when you want to avoid block-
ing of producers, and you have some other way of ensuring that the queue will not grow to
arbitrary size. For both types of blocking queue, bqueue.take() will block if the queue is
empty.
∗ ∗ ∗
The sample program MultiprocessingDemo3.java uses a LinkedBlockingQueue in place of the
ConcurrentLinkedQueue in the previous version, MultiprocessingDemo2.java. In this example,
the queue holds tasks, that is, items of type Runnable, and the queue is declared as an instance
variable named taskQueue:
LinkedBlockingQueue<Runnable> taskQueue;
CHAPTER 12. THREADS AND MULTIPROCESSING 629
When the user clicks the “Start” button and it’s time to compute an image, all of the tasks that
make up the computation are put into this queue. This is done by calling taskQueue.add(task)
for each task. It’s important that this can be done without blocking, since the tasks are created
in the event-handling thread, and we don’t want to block that. The queue in this program
cannot grow indefinitely because the program only works on one image at a time, and there are
only a few hundred tasks per image.
Just as in the previous version of the program, worker threads belonging to a thread pool
will remove tasks from the queue and carry them out. However, in this case, the threads are
created once at the beginning of the program—actually, the first time the “Start” button is
pressed—and the same threads are reused for any number of images. When there are no tasks to
execute, the task queue is empty and the worker threads will block until tasks become available.
Each worker thread runs in an infinite loop, processing tasks forever, but it will spend a lot
of its time blocked, waiting for a task to be added to the queue. Here is the inner class that
defines the worker threads:
/**
* This class defines the worker threads that make up the thread pool.
* A WorkerThread runs in a loop in which it retrieves a task from the
* taskQueue and calls the run() method in that task. Note that if
* the queue is empty, the thread blocks until a task becomes available
* in the queue. The constructor starts the thread, so there is no
* need for the main program to do so. The thread will run at a priority
* that is one less than the priority of the thread that calls the
* constructor.
*
* A WorkerThread is designed to run in an infinite loop. It will
* end only when the Java virtual machine exits. (This assumes that
* the tasks that are executed don’t throw exceptions, which is true
* in this program.) The constructor sets the thread to run as
* a daemon thread; the Java virtual machine will exit automatically when
* the only threads are daemon threads, so the existence of the thread
* pool will not stop the JVM from exiting.
*/
private class WorkerThread extends Thread {
WorkerThread() {
try {
setPriority( Thread.currentThread().getPriority() - 1);
}
catch (Exception e) {
}
try {
setDaemon(true);
}
catch (Exception e) {
}
start(); // Thread starts as soon as it is constructed.
}
public void run() {
while (true) {
try {
Runnable task = taskQueue.take(); // wait for task if necessary
task.run();
CHAPTER 12. THREADS AND MULTIPROCESSING 630
}
catch (InterruptedException e) {
}
}
}
}
We should look more closely at how the thread pool works. The worker threads are created and
started before there is any task to perform. Each thread immediately calls taskQueue.take().
Since the task queue is empty, all the worker threads will block as soon as they are started. To
start the computation of an image, the event-handling thread will create tasks and add them
to the queue. As soon as this happens, worker threads will wake up and start processing tasks,
and they will continue doing so until the queue is emptied. (Note that on a multi-processor
computer, some worker threads can start processing even while the event thread is still adding
tasks to the queue.) When the queue is empty, the worker threads will go back to sleep until
processing starts on the next image.
∗ ∗ ∗
An interesting point in this program is that we want to be able to abort the computation
before it finishes, but we don’t want the worker threads to terminate when that happens. When
the user clicks the “Abort” button, the program calls taskQueue.clear(), which prevents any
more tasks from being assigned to worker threads. However, some tasks are most likely already
being executed when the task queue is cleared. Those tasks will complete after the computation
in which they are subtasks has supposedly been aborted. When those subtasks complete, we
don’t want their output to be applied to the image.
My solution is to assign a job number to each computation job. The job number of the
current job is stored in an instance variable named jobNum, and each task object has an instance
variable that tells which task that job is part of. When a job ends—either because the job
finishes on its own or because the user aborts it—the value of jobNum is incremented. When a
task completes, the job number stored in the task object is compared to jobNum. If they are
equal, then the task is part of the current job, and its output is applied to the image. If they
are not equal, then the task was part of a previous job, and its output is discarded.
It’s important that access to jobNum be properly synchronized. Otherwise, one thread might
check the job number just as another thread is incrementing it, and output meant for an old
job might sneak through after that job has been aborted. In the program, all the methods that
access or change jobNum are synchronized. You can read the source code to see how it works.
∗ ∗ ∗
One more point about MultiprocessingDemo3. . . . I have not provided any way to terminate
the worker threads in this program. They will continue to run until the Java Virtual Machine
exits. To allow thread termination before that, we could use a volatile signaling variable,
running, and set its value to false when we want the worker threads to terminate. The run()
methods for the threads would be replaced by
public void run() {
while ( running ) {
try {
Runnable task = taskQueue.take();
task.run();
}
catch (InterruptedException e) {
CHAPTER 12. THREADS AND MULTIPROCESSING 631
}
}
}
However, if a thread is blocked in taskQueue.take(), it will not see the new value of
running until it becomes unblocked. To ensure that that happens, it is necessary to call
worker.interrupt() for each worker thread worker, just after setting runner to false.
If a worker thread is executing a task when running is set to false, the thread will not
terminate until that task has completed. If the tasks are reasonably short, this is not a problem.
If tasks can take longer to execute than you are willing to wait for the threads to terminate,
then each task must also check the value of running periodically and exit when that value
becomes false.
represents a value of type T that might not be available until some future time. The method
executor.submit(c) returns a Future that represents the result of the future computation.
A Future, v, defines several methods, including v.isDone(), which is a boolean-valued
function that can be called to check whether the result is available; and v.get(), which will
retrieve the value of the future. The method v.get() will block until the value is available. It
can also generate exceptions and needs to be called in a try..catch statement.
As an example, ThreadTest4.java uses Callables, Futures, and an ExecutorService to count
the number of primes in a certain range of integers. (This is the same rather useless compu-
tation that was done by ThreadTest2.java in Subsection 12.1.3.) In this program, a subtask
counts the primes in a subrange of integers. The subtasks are represented by objects of type
Callable<Integer>, defined by this nested class:
/**
* An object belonging to this class will count primes in a specified range
* of integers. The range is from min to max, inclusive, where min and max
* are given as parameters to the constructor. The counting is done in
* the call() method, which returns the number of primes that were found.
*/
private static class CountPrimesTask implements Callable<Integer> {
int min, max;
public CountPrimesTask(int min, int max) {
this.min = min;
this.max = max;
}
public Integer call() {
int count = countPrimes(min,max); // does the counting
return count;
}
}
All the subtasks are submitted to a thread pool implemented as an ExecutorService, and the
Futures that are returned are saved in an array list. In outline:
int processors = Runtime.getRuntime().availableProcessors();
ExecutorService executor = Executors.newFixedThreadPool(processors);
ArrayList<Future<Integer>> results = new ArrayList<>();
for (int i = 0; i < numberOfTasks; i++) {
CountPrimesTask oneTask = . . . ;
Future<Integer> oneResult = executor.submit( oneTask );
results.add(oneResult); // Save the Future representing the (future) result.
}
The integers that are output by all the subtasks need to be added up to give a final result. The
outputs of the subtasks are obtained using the get() methods of the Futures in the list. Since
get() blocks until the result is available, the process completes only when all subtasks have
finished:
int total = 0;
for ( Future<Integer> res : results) {
try {
total += res.get(); // Waits for task to complete!
}
CHAPTER 12. THREADS AND MULTIPROCESSING 633
both methods to be running at the same time? In fact, lock.wait() is a special case: When
a thread calls lock.wait(), it gives up the lock that it holds on the synchronization object.
This gives another thread a chance to execute the synchronized(lock) block that contains the
lock.notify() statement. After the second thread exits from this block, the lock is returned
to the consumer thread so that it can continue.
In the full producer/consumer pattern, multiple results are produced by one or more pro-
ducer threads and are consumed by one or more consumer threads. Instead of having just one
sharedResult object, we keep a list of objects that have been produced but not yet consumed.
Let’s see how this might work in a very simple class that implements the three operations on a
LinkedBlockingQueue<Runnable> that are used in MultiprocessingDemo3:
import java.util.LinkedList;
public class MyLinkedBlockingQueue {
private LinkedList<Runnable> taskList = new LinkedList<Runnable>();
public void clear() {
synchronized(taskList) {
taskList.clear();
}
}
public void add(Runnable task) {
synchronized(taskList) {
taskList.addLast(task);
taskList.notify();
}
}
public Runnable take() throws InterruptedException {
synchronized(taskList) {
while (taskList.isEmpty())
taskList.wait();
return taskList.removeFirst();
}
}
}
An object of this class can be used as a direct replacement for the taskQueue in
MultiprocessingDemo3.
In this class, I have chosen to synchronize on the taskList object, but any object could be
used. In fact, I could simply use synchronized methods, which is equivalent to synchronizing
on this. (Note that you might see a call to wait() or notify() in a synchronized instance
method, with no reference to the object that is being used. Remember that wait() and
notify() in that context really mean this.wait() and this.notify().)
By the way, it is essential that the call to taskList.clear() be synchronized on the same
object, even though it doesn’t call wait() or notify(). Otherwise, there is a race condition
that can occur: The list might be cleared just after the take() method checks that taskList
is non-empty and before it removes an item from the list. In that case, the list is empty again
by the time taskList.removeFirst() is called, resulting in an error.
∗ ∗ ∗
CHAPTER 12. THREADS AND MULTIPROCESSING 636
It is possible for several threads to be waiting for notification. A call to obj.notify() will
wake only one of the threads that is waiting on obj. If you want to wake all threads that
are waiting on obj, you can call obj.notifyAll(). obj.notify() works OK in the above
example because only consumer threads can be blocked. We only need to wake one consumer
thread when a task is added to the queue, and it doesn’t matter which consumer gets the task.
But consider a blocking queue with limited capacity, where producers and consumers can both
block. When an item is added to the queue, we want to make sure that a consumer thread is
notified, not just another producer. One solution is to call notifyAll() instead of notify(),
which will notify all threads including any waiting consumer.
I should also mention a possible confusion about the name of the method obj.notify().
This method does not notify obj of anything! It notifies a thread that has called obj.wait()
(if there is such a thread). Similarly, in obj.wait(), it’s not obj that is waiting for something;
it’s the thread that calls the method.
And a final note on wait: There is another version of wait() that takes a number of
milliseconds as a parameter. A thread that calls obj.wait(milliseconds) will wait only up
to the specified number of milliseconds for a notification. If a notification doesn’t occur during
that period, the thread will wake up and continue without the notification. In practice, this
feature is most often used to let a waiting thread wake periodically while it is waiting in order
to perform some periodic task, such as causing a message “Waiting for computation to finish”
to blink.
∗ ∗ ∗
Let’s look at an example that uses wait() and notify() to allow one thread to control
another. The sample program TowersOfHanoiGUI.java solves the Towers Of Hanoi puzzle
(Subsection 9.1.2), with control buttons that allow the user to control the execution of the
algorithm. The user can click a “Next Step” button to execute just one step in the solution,
which moves a single disk from one pile to another. Clicking “Run” lets the algorithm run
automatically on its own; the text on the button changes from “Run” to “Pause”, and clicking
“Pause” stops the automatic execution. There is also a “Start Over” button that aborts the
current solution and puts the puzzle back into its initial configuration. Here is a picture of the
program in the middle of a solution, including the buttons:
In this program, there are two threads: a thread that runs a recursive algorithm to solve the
puzzle, and the event-handling thread that reacts to user actions. When the user clicks one of
the buttons, a method is called in the event-handling thread. But it’s actually the thread that
is running the recursion that has to respond by, for example, doing one step of the solution or
starting over. The event-handling thread has to send some sort of signal to the solution thread.
This is done by setting the value of a variable that is shared by both threads. The variable is
named status, and its possible values are the constants GO, PAUSE, STEP, and RESTART.
CHAPTER 12. THREADS AND MULTIPROCESSING 637
When the event-handling thread changes the value of this variable, the solution thread
should see the new value and respond. When status equals PAUSE, the solution thread is
paused, waiting for the user to click “Run” or “Next Step”. This is the initial state, when
the program starts. If the user clicks “Next Step”, the event-handling thread sets the value of
status to “STEP”; the solution thread should see the new value and respond by executing one
step of the solution and then resetting status to PAUSE. If the user clicks “Run”, status is
set to GO, which should cause the solution thread to run automatically. When the user clicks
“Pause” while the solution is running, status is reset to PAUSE, and the solution thread should
return to its paused state. If the user clicks “Start Over”, the event-handling thread sets status
to RESTART, and the solution thread should respond by ending the current recursive solution.
The main point for us is that when the solution thread is paused, it is sleeping. It won’t see a
new value for status unless it wakes up! To make that possible, the program uses wait() in the
solution thread to put that thread to sleep, and it uses notify() in the event-handling thread
to wake up the solution thread whenever it changes the value of status. Here are the methods
that respond to clicks on the buttons. When the user clicks a button, the corresponding method
changes the value of status and calls notify() to wake up the solution thread:
synchronized private void doStopGo() {
if (status == GO) { // Animation is running. Pause it.
status = PAUSE;
nextStepButton.setDisable(false);
runPauseButton.setText("Run");
}
else { // Animation is paused. Start it running.
status = GO;
nextStepButton.setDisable(true); // Disabled when animation is running
runPauseButton.setText("Pause");
}
notify(); // Wake up the thread so it can see the new status value!
}
synchronized private void doNextStep() {
status = STEP;
notify();
}
synchronized private void doRestart() {
status = RESTART;
notify();
}
These methods are synchronized to allow the calls to notify(). Remember that the notify()
method in an object can only be called by a thread that holds that object’s synchronization
lock. In this case, the synchronization object is this. Synchronization is also necessary because
of race conditions that arise due to the fact that the value of status can also be changed by
the solution thread.
The solution thread calls a method named checkStatus() to check the value of status.
This method calls wait() if the status is PAUSE, which puts the solution thread to sleep until
the event-handling thread calls notify(). Note that if the status is RESTART, checkStatus()
throws an IllegalStateException:
synchronized private void checkStatus() {
while (status == PAUSE) {
CHAPTER 12. THREADS AND MULTIPROCESSING 638
try {
wait();
}
catch (InterruptedException e) {
}
}
// At this point, status is RUN, STEP, or RESTART.
if (status == RESTART)
throw new IllegalStateException("Restart");
// At this point, status is RUN or STEP, and the solution should proceed.
}
The run() method for the solution thread sets up the initial state of the puzzle and then calls
a solve() method to solve the puzzle. It runs in an infinite loop so that it can solve the puzzle
multiple times. To implement the wait/notify control strategy, run() calls checkStatus() be-
fore starting the solution, and solve() calls checkStatus() after each move. If checkStatus()
throws an IllegalStateException, the call to solve() is terminated early. (We used the method
of throwing an exception to terminate a recursive algorithm before, in Subsection 12.2.2.)
You can check the full source code to see how this all fits into the complete program. If you
want to learn how to use wait() and notify() directly, understanding this example is a good
place to start!
the network. (It’s a little complicated: a socket uses a “buffer” to hold data that is supposed
to be transmitted over the network. A buffer is just a block of memory that is used like a
queue. The output method drops its data into the buffer; lower-level software removes data
from the buffer and transmits it over the network. If the buffer fills up, the output method will
block until space becomes available in the buffer. Note that when the output method returns,
it doesn’t mean that the data has gone out over the network—it just means that the data has
gone into the buffer and is scheduled for later transmission.)
We say that network communication uses blocking I/O, because input and output oper-
ations on the network can block for indefinite periods of time. Programs that use the network
must be prepared to deal with this blocking. In some cases, it’s acceptable for a program to
simply shut down all other processing and wait for input. (This is what happens when a com-
mand line program reads input typed by the user. User input is another type of blocking I/O.)
However, threads make it possible for some parts of a program to continue doing useful work
while other parts are blocked. A network client program that sends requests to a server might
get by with a single thread, if it has nothing else to do while waiting for the server’s responses.
A network server program, on the other hand, can typically be connected to several clients at
the same time. While waiting for data to arrive from a client, the server certainly has other
things that it can do, namely communicate with other clients. When a server uses different
threads to handle the communication with different clients, the fact that I/O with one client is
blocked won’t stop the server from communicating with other clients.
It’s important to understand that using threads to deal with blocking I/O differs in a
fundamental way from using threads to speed up computation. When using threads for speed-
up in Subsection 12.3.2, it made sense to use one thread for each available processor. If only
one processor is available, using more than one thread will yield no speed-up at all; in fact, it
would slow things down because of the extra overhead involved in creating and managing the
threads.
In the case of blocking I/O, on the other hand, it can make sense to have many more threads
than there are processors, since at any given time many of the threads can be blocked. Only the
active, unblocked threads are competing for processing time. In the ideal case, to keep all the
processors busy, you would want to have one active thread per processor (actually somewhat
less than that, on average, to allow for variations over time in the number of active threads).
On a network server program, for example, threads generally spend most of their time blocked
waiting for I/O operations to complete. If threads are blocked, say, about 90% of the time,
you’d like to have about ten times as many threads as there are processors. So even on a
computer that has just a single processor, server programs can make good use of large numbers
of threads.
messages that arrive from the other side of the connection. As soon as the message arrives,
it is displayed to the user; then, the message-reading thread blocks until the next incoming
message arrives. While it is blocked, however, other threads can continue to run. In particular,
the event-handling thread that responds to user actions keeps running; that thread can send
outgoing messages as soon as the user generates them.
The GUIChat program can act as either the client end or the server end of a connection.
(Refer back to Subsection 11.4.3 for information about how clients and servers work.) The
program has a “Listen” button that the user can click to create a server socket that will listen
for an incoming connection request; this makes the program act as a server. It also has a
“Connect” button that the user can click to send a connection request; this makes the program
act as a client. As usual, the server listens on a specified port number. The client needs to
know the computer on which the server is running and the port on which the server is listening.
There are input boxes in the GUIChat window where the user can enter this information.
Once a connection has been established between two GUIChat windows, each user can send
messages to the other. The window has an input box where the user types a message. Pressing
return sends the message. This means that the sending of the message is handled by the usual
event-handling thread, in response to an event generated by a user action. Messages are received
by a separate thread that just sits around waiting for incoming messages. This thread blocks
while waiting for a message to arrive; when a message does arrive, it displays that message to
the user. The window contains a large transcript area that displays both incoming and outgoing
messages, along with other information about the network connection.
I urge you to compile the source code, GUIChat.java, and try the program. To try it
on single computer, you can run two copies of the program on that computer, and make a
connection between one program window and the other program window, using “localhost” or
“127.0.0.1” as the name of the computer. I also urge you to read the source code. I will discuss
only parts of it here.
The program uses a nested class, ConnectionHandler, to handle most network-related tasks.
ConnectionHandler is a subclass of Thread. The ConnectionHandler thread is responsible for
opening the network connection and then for reading incoming messages once the connection
has been opened. By putting the connection-opening code in a separate thread, we make sure
that the GUI is not blocked while the connection is being opened. (Like reading incoming
messages, opening a connection is a blocking operation that can take some time to complete.)
The ConnectionHandler handles opening the connection both when the program acts as a server
and when it acts as a client. The thread is created when the user clicks either the “Listen”
button or the “Connect” button. The “Listen” button makes the thread act as a server, while
“Connect” makes it act as a client. To distinguish these two cases, the ConnectionHandler class
has the two constructors that are shown below. Note that the postMessage() method posts a
message to the transcript area of the window, where it will be visible to the user:
/**
* Listen for a connection on a specified port. The constructor
* does not perform any network operations; it just sets some
* instance variables and starts the thread. Note that the
* thread will only listen for one connection, and then will
* close its server socket.
*/
ConnectionHandler(int port) { // For acting as the "server."
state = ConnectionState.LISTENING;
this.port = port;
CHAPTER 12. THREADS AND MULTIPROCESSING 641
*/
synchronized void send(String message) {
if (state == ConnectionState.CONNECTED) {
postMessage("SEND: " + message);
out.println(message);
out.flush();
if (out.checkError()) {
postMessage("\nERROR OCCURRED WHILE TRYING TO SEND DATA.");
close(); // Closes the connection.
}
}
}
The received() method is called by the connection-handling thread after a message has been
read from the remote user. Its only job is to display the message to the user, but again it is
synchronized to avoid the race condition that could occur if the connection state were changed
by another thread while this method is being executed:
/**
* This is called by the run() method when a message is received from
* the other side of the connection. The message is posted to the
* transcript, but only if the connection state is CONNECTED. (This
* is because a message might be received after the user has clicked
* the "Disconnect" button; that message should not be seen by the
* user.)
*/
synchronized private void received(String message) {
if (state == ConnectionState.CONNECTED)
postMessage("RECEIVE: " + message);
}
connection to the thread. This takes very little time, and in particular will not block. The
run() method of the thread handles the connection in exactly the same way that it would be
handled by the original program. This is not at all difficult to program. Here’s the new version
of the program, with significant changes shown in italic. Note again that the constructor for
the connection thread does very little and in particular cannot block; this is very important
since the constructor runs in the main thread:
import java.net.*;
import java.io.*;
import java.util.Date;
/**
* This program is a server that takes connection requests on
* the port specified by the constant LISTENING PORT. When a
* connection is opened, the program sends the current time to
* the connected socket. The program will continue to receive
* and process connections until it is killed (by a CONTROL-C,
* for example).
*
* This version of the program creates a new thread for
* every connection request.
*/
public class DateServerWithThreads {
public static final int LISTENING PORT = 32007;
public static void main(String[] args) {
ServerSocket listener; // Listens for incoming connections.
Socket connection; // For communication with the connecting program.
/* Accept and process connections forever, or until some error occurs. */
try {
listener = new ServerSocket(LISTENING PORT);
System.out.println("Listening on port " + LISTENING PORT);
while (true) {
// Accept next connection request and create a thread to handle it.
connection = listener.accept();
ConnectionHandler handler = new ConnectionHandler(connection);
handler.start();
}
}
catch (Exception e) {
System.out.println("Sorry, the server has shut down.");
System.out.println("Error: " + e);
return;
}
} // end main()
/**
* Defines a thread that handles the connection with one
* client.
*/
private static class ConnectionHandler extends Thread {
CHAPTER 12. THREADS AND MULTIPROCESSING 645
catch (InterruptedException e) {
continue; // (If interrupted, just go back to start of while loop.)
}
String clientAddress = client.getInetAddress().toString();
try {
System.out.println("Connection from " + clientAddress );
System.out.println("Handled by thread " + this);
Date now = new Date(); // The current date and time.
PrintWriter outgoing; // Stream for sending data.
outgoing = new PrintWriter( client.getOutputStream() );
outgoing.println( now.toString() );
outgoing.flush(); // Make sure the data is actually sent!
client.close();
}
catch (Exception e){
System.out.println("Error on connection with: "
+ clientAddress + ": " + e);
}
}
}
The main program, in the meantime, runs in an infinite loop in which connections are accepted
and added to the queue:
while (true) {
// Accept next connection request and put it in the queue.
connection = listener.accept();
try {
connectionQueue.put(connection); // Blocks if queue is full.
}
catch (InterruptedException e) {
}
}
The queue in this program is of type ArrayBlockingQueue<Socket>. As such, it has a limited
capacity, and the put() operation on the queue will block if the queue is full. But wait—didn’t
we want to avoid blocking the main program? When the main program is blocked, the server is
no longer accepting connections, and clients who are trying to connect are kept waiting. Would
it be better to use a LinkedBlockingQueue, with an unlimited capacity?
In fact, connections in the blocking queue are waiting anyway; they are not being serviced.
If the queue grows unreasonably long, connections in the queue will have to wait for an un-
reasonable amount of time. If the queue keeps growing indefinitely, that just means that the
server is receiving connection requests faster than it can process them. That could happen
for several reasons: Your server might simply not be powerful enough to handle the volume of
traffic that you are getting; you need to buy a new server. Or perhaps the thread pool doesn’t
have enough threads to fully utilize your server; you should increase the size of the thread pool
to match the server’s capabilities. Or maybe your server is under a “Denial Of Service” attack,
in which some bad guy is deliberately sending your server more requests than it can handle in
an attempt to keep other, legitimate clients from getting service.
In any case, ArrayBlockingQueue with limited capacity is the correct choice. The queue
should be short enough so that connections in the queue will not have to wait too long for
service. In a real server, the size of the queue and the number of threads in the thread pool
CHAPTER 12. THREADS AND MULTIPROCESSING 647
should be adjusted to “tune” the server to account for the particular hardware and network on
which the server is running and for the nature of the client requests that it typically processes.
Optimal tuning is, in general, a difficult problem.
There is, by the way, another way that things can go wrong: Suppose that the server needs
to read some data from the client, but the client doesn’t send any data. The thread that is
trying to read the data can then block indefinitely, waiting for the input. If a thread pool is
being used, this could happen to every thread in the pool. In that case, no further processing
can ever take place! The solution to this problem is to have connections “time out” if they are
inactive for an excessive period of time. Typically, each connection thread will keep track of
the time when it last received data from the client. The server runs another thread (sometimes
called a “reaper thread”, after the Grim Reaper) that wakes up periodically and checks each
connection thread to see how long it has been inactive. A connection thread that has been
waiting too long for input is terminated, and a new thread is started in its place. The question
of how long the timeout period should be is another difficult tuning issue.
programming, in which the problem can be broken down into tasks that can be performed
independently, with no communication between the tasks. To apply distributed computing to
this type of problem, we can use one “master” program that divides the problem into tasks and
sends those tasks over the network to “worker” programs that do the actual work. The worker
programs send their results back to the master program, which combines the results from all
the tasks into a solution of the overall problem. In this context, the worker programs are of-
ten called “slaves,” and the program uses the so-called master/slave approach to distributed
computing.
The demonstration program is defined by three source code files: CLMandelbrotMas-
ter.java defines the master program; CLMandelbrotWorker.java defines the worker programs;
and CLMandelbrotTask.java defines the class, CLMandelbrotTask, that represents an individual
task that is performed by the workers. The master divides the overall problem into a collec-
tion of tasks; it distributes those tasks to the workers that will execute the tasks and send the
results back to the master; and the master applies the results from all the individual tasks to
the overall problem.
To run the demonstration, you must first start the CLMandelbrotWorker program on several
computers (probably by running it on the command line). This program uses CLMandelbrot-
Task, so both class files, CLMandelbrotWorker.class and CLMandelbrotTask.class, must be
present on the worker computers. You can then run CLMandelbrotMaster on the master com-
puter. Note that the master program also requires the class CLMandelbrotTask. You must
specify the host name or IP address of each of the worker computers as command line ar-
guments for CLMandelbrotMaster. The worker programs listen for connection requests from
the master program, and the master program must be told where to send those requests. For
example, if the worker program is running on three computers with IP addresses 172.21.7.101,
172.21.7.102, and 172.21.7.103, then you can run CLMandelbrotMaster with the command
java CLMandelbrotMaster 172.21.7.101 172.21.7.102 172.21.7.103
The master will make a network connection to the worker at each IP address; these connections
will be used for communication between the master program and the workers.
It is possible to run several copies of CLMandelbrotWorker on the same computer, but
they must listen for network connections on different ports. It is also possible to run
CLMandelbrotWorker on the same computer as CLMandelbrotMaster. You might even see
some speed-up when you do this, if your computer has several processors. See the comments in
the program source code files for more information, but here are some commands that you can
use to run the master program and two copies of the worker program on the same computer.
Give these commands in separate command windows:
java CLMandelbrotWorker (Listens on default port)
java CLMandelbrotWorker 2501 (Listens on port 2501)
java CLMandelbrotMaster localhost localhost:2501
Every time CLMandelbrotMaster is run, it solves exactly the same problem. (For this
demonstration, the nature of the problem is not important, but the problem is to compute the
data needed for a picture of a small piece of the famous “Mandelbrot Set.” If you are interested
in seeing the picture that is produced, uncomment the call to the saveImage() method at the
end of the main() routine in CLMandelbrotMaster.java.)
You can run CLMandelbrotMaster with different numbers of worker programs to see how
the time required to solve the problem depends on the number of workers. (Note that the
CHAPTER 12. THREADS AND MULTIPROCESSING 649
worker programs continue to run after the master program exits, so you can run the mas-
ter program several times without having to restart the workers.) In addition, if you run
CLMandelbrotMaster with no command line arguments, it will solve the entire problem on its
own, so you can see how long it takes to do so without using distributed computing. In a
trial that I ran on some very old, slow computers, it took 40 seconds for CLMandelbrotMaster
to solve the problem on its own. Using just one worker, it took 43 seconds. The extra time
represents extra work involved in using the network; it takes time to set up a network connec-
tion and to send messages over the network. Using two workers (on different computers), the
problem was solved in 22 seconds. In this case, each worker did about half of the work, and
their computations were performed in parallel, so that the job was done in about half the time.
With larger numbers of workers, the time continued to decrease, but only up to a point. The
master program itself has a certain amount of work to do, no matter how many workers there
are, and the total time to solve the problem can never be less than the time it takes for the
master program to do its part. In this case, the minimum time seemed to be about five seconds.
∗ ∗ ∗
Let’s take a look at how this distributed application is programmed. The master program
divides the overall problem into a set of tasks. Each task is represented by an object of type
CLMandelbrotTask. These tasks have to be communicated to the worker programs, and the
worker programs must send back their results. Some protocol is needed for this communication.
I decided to use character streams. The master encodes a task as a line of text, which is sent
to a worker. The worker decodes the text (into an object of type CLMandelbrotTask) to find
out what task it is supposed to perform. It performs the assigned task. It encodes the results
as another line of text, which it sends back to the master program. Finally, the master decodes
the results and combines them with the results from other tasks. After all the tasks have been
completed and their results have been combined, the problem has been solved.
A worker receives not just one task, but a sequence of tasks. Each time it finishes a task
and sends back the result, it is assigned a new task. After all tasks are completed, the worker
receives a “close” command that tells it to close the connection. In CLMandelbrotWorker.java,
all this is done in a method named handleConnection() that is called to handle a connection
that has already been opened to the master program. It uses a method readTask() to decode
a task that it receives from the master and a method writeResults() to encode the results of
the task for transmission back to the master. It must also handle any errors that occur:
private static void handleConnection(Socket connection) {
try {
BufferedReader in = new BufferedReader(
new InputStreamReader( connection.getInputStream()) );
PrintWriter out = new PrintWriter(connection.getOutputStream());
while (true) {
String line = in.readLine(); // Message from the master.
if (line == null) {
// End-of-stream encountered -- should not happen.
throw new Exception("Connection closed unexpectedly.");
}
if (line.startsWith(CLOSE CONNECTION COMMAND)) {
// Represents the normal termination of the connection.
System.out.println("Received close command.");
break;
}
CHAPTER 12. THREADS AND MULTIPROCESSING 650
The job of a thread is to send a sequence of tasks to a worker thread and to receive the
results that the worker sends back. The thread is also responsible for opening the connection
in the first place. A pseudocode outline for the process executed by the thread might look like:
Create a socket connected to the worker program.
Create input and output streams for communicating with the worker.
while (true) {
Let task = tasks.poll().
If task == null
break; // All tasks have been assigned.
Encode the task into a message and transmit it to the worker.
Read the response from the worker.
Decode and process the response.
}
Send a "close" command to the worker.
Close the socket.
This would work OK. However, there are a few subtle points. First of all, the thread must be
ready to deal with a network error. For example, a worker might shut down unexpectedly. But if
that happens, the master program can continue, provided other workers are still available. (You
can try this when you run the program: Stop one of the worker programs, with CONTROL-C,
and observe that the master program still completes successfully.) A difficulty arises if an
error occurs while the thread is working on a task: If the problem as a whole is going to
be completed, that task will have to be reassigned to another worker. I take care of this by
putting the uncompleted task back into the task list. (Unfortunately, my program does not
handle all possible errors. If the last worker thread fails, there will be no one left to take
over the uncompleted task. Also, if a network connection “hangs” indefinitely without actually
generating an error, my program will also hang, waiting for a response from a worker that
will never arrive. A more robust program would have some way of detecting the problem and
reassigning the task.)
Another defect in the procedure outlined above is that it leaves the worker program idle
while the thread in the master program is processing the worker’s response. It would be nice
to get a new task to the worker before processing the response from the previous task. This
would keep the worker busy and allow two operations to proceed simultaneously instead of
sequentially. (In this example, the time it takes to process a response is so short that keeping
the worker waiting while it is done probably makes no significant difference. But as a general
principle, it’s desirable to have as much parallelism as possible in the algorithm.) We can
modify the procedure to take this into account:
try {
Create a socket connected to the worker program.
Create input and output streams for communicating with the worker.
Let currentTask = tasks.poll().
Encode currentTask into a message and send it to the worker.
while (true) {
Read the response from the worker.
Let nextTask = tasks.poll().
If nextTask != null {
// Send nextTask to the worker before processing the
// response to currentTask.
Encode nextTask into a message and send it to the worker.
}
CHAPTER 12. THREADS AND MULTIPROCESSING 652
sense to define a reusable framework in a package that can be included as a unit in a variety of
projects.
Integrated development environments such as Eclipse make it very easy to use packages:
To use the netgame package in a project in an IDE, simply copy-and-paste the entire netgame
directory into the project. Of course, since netgames use JavaFX, you need to use an Eclipse
project configured to support JavaFX, as discussed in Section 2.6.
If you work on the command line, you should be in a working directory that includes the
netgame directory as a subdirectory. Unless you are using an older JDK that has JavaFX built
in, you need to add JavaFX options to the javac and java commands. Let’s say that you’ve
defined jfxc and jfx commands that are equivalent to the javac and java with JavaFX options
included, as discussed in Subsection 2.6.3. Then, to compile all the java files in the package
netgame.common, for example, you can use the following command in Mac OS or Linux:
jfxc netgame/common/*.java
For Windows, you should use backslashes instead of forward slashes:
jfxc netgame\common\*.java
Of course, for a JDK with built-in JavaFX, you would use javac, not jfxc. You will need similar
commands to compile the source code for the examples in this section, which are defined in
other subpackages of netgame.
To run a main program that is defined in a package, you should again be in a directory
that contains the package as a subdirectory, and you should use the full name of the class that
you want to run. For example, the ChatRoomWindow class, discussed later in this section, is
defined in the package netgame.chat, so you would run it with the command
jfx netgame.chat.ChatRoomWindow
∗ ∗ ∗
The applications discussed in this section are examples of distributed computing, since they
involve several computers communicating over a network. Like the example in Subsection 12.4.5,
they use a central “server,” or “master,” to which a number of “clients” will connect. All
communication goes through the server; a client cannot send messages directly to another
client. In this section, I will refer to the server as a hub, in the sense of “communications hub”:
Client
Client
HUB
Client Client
The main things that you need to understand are that: The hub must be running before
any clients are started. Clients connect to the hub and can send messages to the hub. The
hub processes all messages from clients sequentially, in the order in which they are received.
The processing can result in the hub sending messages out to one or more clients. Each client
CHAPTER 12. THREADS AND MULTIPROCESSING 654
send thread
Client
receive thread
HUB
receive
Incoming
Message Queue
send
Message
Processing
Thread
There is one more thread in the hub, not shown in the illustration. This final thread creates
a ServerSocket and uses it to listen for connection requests from clients. Each time it accepts
a connection request, it hands off the client socket to another object, defined by the nested
class ConnectionToClient, which will handle communication with that client. Each connected
client is identified by an ID number. ID numbers 1, 2, 3, . . . are assigned to clients as they
connect. Since clients can also disconnect, the clients connected at any give time might not
have consecutive IDs. A variable of type TreeMap<Integer,ConnectionToClient> associates the
ID numbers of connected clients with the objects that handle their connections.
The messages that are sent and received are objects. The I/O streams that are used for
reading and writing objects are of type ObjectInputStream and ObjectOutputStream. (See Sub-
section 11.1.6.) The output stream of a socket is wrapped in an ObjectOutputStream to make
it possible to transmit objects through that socket. The socket’s input stream is wrapped in
an ObjectInputStream to make it possible to receive objects. Remember that the objects that
are used with such streams must implement the interface java.io.Serializable.
The netgame Hub class is defined in the file Hub.java, in the package netgame.common.
The port on which the server socket will listen must be specified as a parameter to the Hub
constructor. The Hub class defines a method
CHAPTER 12. THREADS AND MULTIPROCESSING 655
For more information—and to see how all this is implemented—you should read the source
code file Hub.java. With some effort and study, you should be able to understand everything
in that file. (However, you only need to understand the public and protected interface of Hub
and other classes in the netgame framework to write applications based on it.)
∗ ∗ ∗
Turning to the client side, the basic netgame client class is defined in the file Client.java, in
the package netgame.common. The Client class has a constructor that specifies the host name
(or IP address) and port number of the hub to which the client will connect. This constructor
blocks until the connection has been established.
Client is an abstract class. Every netgame application must define a subclass of Client
and provide a definition for the abstract method:
abstract protected void messageReceived(Object message);
This method is called each time a message is received from the netgame hub. A subclass of
client might also override the protected methods playerConnected, playerDisconnected,
serverShutdown, and connectionClosedByError. See the source code for more information.
I should also note that Client contains the protected instance variable connectedPlayerIDs,
of type int[], an array containing the ID numbers of all the clients that are currently connected
to the hub. The most important public methods that are provided by the Client class are
• send(message) — transmits a message to the hub. The message can be any non-null
object that implements the Serializable interface.
• getID() — gets the ID number that was assigned to this client by the hub.
• disconnect() — closes the client’s connection to the hub. It is not possible to send
messages after disconnecting. The send() method will throw an IllegalStateException if
an attempt is made to do so.
The Hub and Client classes are meant to define a general framework that can be used as the
basis for a variety of networked games—and, indeed, of other distributed programs. The low
level details of network communication and multithreading are hidden in the private sections
of these classes. Applications that build on these classes can work in terms of higher-level
concepts such as players and messages. The design of these classes was developed though
several iterations, based on experience with several actual applications. I urge you to look
at the source code to see how Hub and Client use threads, sockets, and I/O streams. In the
remainder of this section, I will discuss three applications built on the netgame framework. I
will not discuss these applications in great detail. You can find the complete source code for
all three in the netgame package.
*/
private class ChatClient extends Client {
/**
* Opens a connection to the chat room server on a specified computer.
*/
ChatClient(String host) throws IOException {
super(host, PORT);
}
/**
* Responds when a message is received from the server. It should be
* a ForwardedMessage representing something that one of the participants
* in the chat room is saying. The message is simply added to the
* transcript, along with the ID number of the sender.
*/
protected void messageReceived(Object message) {
if (message instanceof ForwardedMessage) {
// (no other message types are expected)
ForwardedMessage bm = (ForwardedMessage)message;
addToTranscript("#" + bm.senderID + " SAYS: " + bm.message);
}
}
/**
* Called when the connection to the client is shut down because of some
* error message. (This will happen if the server program is terminated.)
*/
protected void connectionClosedByError(String message) {
addToTranscript(
"Sorry, communication has shut down due to an error:\n "
+ message );
Platform.runLater( () -> {
sendButton.setDisable(true);
messageInput.setEditable(false);
messageInput.setDisable(true);
messageInput.setText("");
});
connected = false;
connection = null;
}
/**
* Posts a message to the transcript when someone joins the chat room.
*/
protected void playerConnected(int newPlayerID) {
addToTranscript(
"Someone new has joined the chat room, with ID number "
+ newPlayerID );
}
/**
* Posts a message to the transcript when someone leaves the chat room.
*/
protected void playerDisconnected(int departingPlayerID) {
addToTranscript( "The person with ID number "
CHAPTER 12. THREADS AND MULTIPROCESSING 659
package netgame.tictactoe;
import java.io.IOException;
import netgame.common.Hub;
/**
* A "Hub" for the network TicTacToe game. There is only one Hub
* for a game, and both network players connect to the same Hub.
* Official information about the state of the game is maintained
* on the Hub. When the state changes, the Hub sends the new
* state to both players, ensuring that both players see the
* same state.
*/
public class TicTacToeGameHub extends Hub {
private TicTacToeGameState state; // Records the state of the game.
/**
* Create a hub, listening on the specified port. Note that this
* method calls setAutoreset(true), which will cause the output stream
* to each client to be reset before sending each message. This is
* essential since the same state object will be transmitted over and
* over, with changes between each transmission.
* @param port the port number on which the hub will listen.
* @throws IOException if a listener cannot be opened on the specified port.
*/
public TicTacToeGameHub(int port) throws IOException {
super(port);
state = new TicTacToeGameState();
setAutoreset(true);
}
/**
* Responds when a message is received from a client. In this case,
* the message is applied to the game state, by calling state.applyMessage().
* Then the possibly changed state is transmitted to all connected players.
*/
protected void messageReceived(int playerID, Object message) {
state.applyMessage(playerID, message);
sendToAll(state);
}
/**
* This method is called when a player connects. If that player
* is the second player, then the server’s listening socket is
* shut down (because only two players are allowed), the
* first game is started, and the new state -- with the game
* now in progress -- is transmitted to both players.
*/
protected void playerConnected(int playerID) {
if (getPlayerList().length == 2) {
shutdownServerSocket();
state.startFirstGame();
sendToAll(state);
}
}
CHAPTER 12. THREADS AND MULTIPROCESSING 661
/**
* This method is called when a player disconnects. This will
* end the game and cause the other player to shut down as
* well. This is accomplished by setting state.playerDisconnected
* to true and sending the new state to the remaining player, if
* there is one, to notify that player that the game is over.
*/
protected void playerDisconnected(int playerID) {
state.playerDisconnected = true;
sendToAll(state);
}
}
A player’s interface to the game is represented by the class TicTacToeWindow. As in the
chat room application, this class defines a nested subclass of Client to represent the client’s
connection to the hub. When the state of the game changes, a message is sent to each client,
and the client’s messageReceived() method is called to process that message. That method,
in turn, calls a newState() method in the TicTacToeWindow class to update the window. That
method is called on the JavaFX application thread using Platform.runLater():
protected void messageReceived(Object message) {
if (message instanceof TicTacToeGameState) {
Platform.runLater( () -> newState( (TicTacToeGameState)message ) );
}
}
To run the TicTacToe netgame, the two players should each run the program Main.java in
the package netgame.tictactoe. This program presents the user with a window where the user
can choose to start a new game or to join an existing game. If the user starts a new game, then a
TicTacToeHub is created to manage the game, and a second window of type TicTacToeWindow
is opened that immediately connects to the hub. The game will start as soon as a second
player connects to the hub. On the other hand, if the user running Main chooses to connect
to an existing game, then no hub is created. A TicTacToeWindow is created, and that window
connects to the hub that was created by the first player. The second player has to know the
name of the computer where the first player’s program is running. As usual, for testing, you
can run everything on one computer and use “localhost” as the computer name.
(This is the first program that we have seen that uses two different windows. Note that
TicTacToeWindow is defined as a subclass of Stage, the JavaFX class that represents windows.
A JavaFX program starts with a “primary stage” that is created by the system and passed
as a parameter to the start() method. But an application can certainly create additional
windows.)
In general outline, the Poker game is similar to the TicTacToe game. There is a Main class
that is run by both players. The first player starts a new game; the second must join that
existing game. There is a class PokerGameState to represent the state of a game. And there is
a subclass, PokerHub, of Hub to manage the game.
But Poker is a much more complicated game than TicTacToe, and the game state is cor-
respondingly more complicated. It’s not clear that we want to broadcast a new copy of the
complete game state to the players every time some minor change is made in the state. Fur-
thermore, it doesn’t really make sense for both players to know the full game state—that would
include the opponent’s hand and full knowledge of the deck from which the cards are dealt. (Of
course, our client programs wouldn’t have to show the full state to the players, but it would
be easy enough for a player to substitute their own client program to enable cheating.) So in
the Poker application, the full game state is known only to the PokerHub. A PokerGameState
object represents a view of the game from the point of view of one player only. When the state
of the game changes, the PokerHub creates two different PokerGameState objects, representing
the state of the game from each player’s point of view, and it sends the appropriate game state
object to each player. You can see the source code for details.
(One of the hard parts in poker is to implement some way to compare two hands, to see
which is higher. In my game, this is handled by the class PokerRank. You might find this class
useful in other poker games.)
Exercises 663
1. Subsection 12.1.3 discusses the need for synchronization in multithreaded programs, and (solution)
it defines a ThreadSafeCounter class with the necessary synchronization. Is this really
important? Can you really get errors by using an unsynchronized counter with multiple
threads? Write a program to find out. Use the following unsynchronized counter class,
which you can include as a nested class in your program:
static class Counter {
int count;
void inc() {
count = count+1;
}
int getCount() {
return count;
}
}
Write a thread class that will repeatedly call the inc() method in an object of type
Counter. The object should be a shared global variable. Create several threads, start
them all, and wait for all the threads to terminate. Print the final value of the counter,
and see whether it is correct.
Let the user enter the number of threads and the number of times that each thread
will increment the counter. You might need a fairly large number of increments to see
an error. And of course there can never be any error if you use just one thread. Your
program can use join() to wait for a thread to terminate (see Subsection 12.1.2).
2. Exercise 3.2 asked you to find the integer in the range 1 to 10000 that has the largest (solution)
number of divisors. Now write a program that uses multiple threads to solve the same
problem, but for the range 1 to 100000. By using threads, your program will take less
time to do the computation when it is run on a multiprocessor computer. At the end of
the program, output the elapsed time, the integer that has the largest number of divisors,
and the number of divisors that it has. The program can be modeled on the sample
prime-counting program ThreadTest2.java from Subsection 12.1.3. For this exercise, you
should simply divide up the problem into parts and create one thread to do each part.
3. In the previous exercise, you divided up a large task into a small number of large pieces (solution)
and created a thread to execute each task. Because of the nature of the problem, this
meant that some threads had much more work to do than others—it is much easier to
find the number of divisors of a small number than it is of a big number. As discussed in
Subsection 12.3.1, a better approach is to break up the problem into a fairly large number
of smaller problems. Subsection 12.3.2 shows how to use a thread pool to execute the
tasks: Each thread in the pool runs in a loop in which it repeatedly takes a task from a
queue and carries out that task. Implement a thread pool strategy for solving the same
maximum-number-of-divisors problem as in the previous exercise.
To make things even more interesting, you should try a new technique for combining
the results from all the tasks: Use two queues in your program. Use a queue of tasks, as
usual, to hold the tasks that will be executed by the thread pool (Subsection 12.3.2). But
also use a queue of results produced by the threads. When a task completes, the result
Exercises 664
from that task should be placed into the result queue. The main program can read results
from the second queue as they become available, and combine all the results to get the
final answer. The result queue will have to be a blocking queue (Subsection 12.3.3), since
the main program will have to wait for results to become available. Note that the main
program knows the exact number of results that it expects to read from the queue, so it
can do so in a for loop; when the for loop completes, the main program knows that all
the tasks have been executed.
4. In previous exercise, you used a thread pool and a queue of tasks to find the integer in the (solution)
range 1 to 100000 that has the largest number of divisors. Subsection 12.3.4 discusses a
higher-level approach that uses an ExecutorService. Write one more program to solve the
problem, this time using an ExecutorService and Futures. The program should still break
up the computation into a fairly large number of fairly small tasks, and it should still print
out the largest number of divisors and the integer that has that number of divisors.
(There is yet another way to solve the same problem: the stream API from Section 10.6.
My solution using the stream API, however, uses an aspect of the stream API that I did
not cover: the interface Optional<T>. My on-line solution of this exercise also discusses
how to use streams to solve the problem.)
5. In Exercise 11.3, you wrote a network server program that can send text files from a (solution)
specified directory to clients. That program used a single thread, which handled all the
communication with each client. Modify the program to turn it into a multithreaded
server. Use a thread pool of connection-handling threads and use an ArrayBlockingQueue
to get connected sockets from the main() routine to the threads. The sample program
DateServerWithThreads.java from Subsection 12.4.3 is an example of a multithreaded
server that works in this way. Your server program will work with the same client program
as the original server. You wrote the client program as the solution to Exercise 11.4.
6. It is possible to get an estimate of the mathematical constant π by using a random process. (solution)
The idea is based on the fact that the area of a circle of radius 1 is equal to π, and the
area of a quarter of that circle is π/4. Here is a picture of a quarter of a circle of radius 1,
inside a 1-by-1 square:
The area of the whole square is one, while the area of the part inside the circle is π/4. If
we choose a point in the square at random, the probability that it is inside the circle is
π/4. If we choose N points in the square at random, and if C of them are inside the circle,
we expect the fraction C/N of points that fall inside the circle to be about π/4. That is,
we expect 4*C/N to be close to π. If N is large, we can expect 4*C/N to be a good estimate
for π, and as N gets larger and larger, the estimate is likely to improve.
We can pick a random point in the square by choosing numbers x and y in the range
0 to 1 (using Math.random()). Since the equation of the circle is x*x+y*y=1, the point
lies inside the circle if x*x+y*y is less than 1. One trial consists of picking x and y and
Exercises 665
testing whether x*x+y*y is less than 1. To get an estimate for π, you have to do many
trials, count the trials, and count the number of trials in which x*x+y*y is less than 1,
For this exercise, you should write a GUI program that does this computation and
displays the result. The computation should be done in a separate thread, and the results
should be displayed periodically. The program can use Labels to the display the results.
It should set the text on the labels after running each batch of, say, one million trials.
(Setting the text after each trial doesn’t make sense, since millions of trials can be done in
one second, and trying to change the display millions of times per second would be silly.
Your program should have a “Run”/”Pause” button that controls the computation.
When the program starts, clicking “Run” will start the computation and change the text
on the button to “Pause”. Clicking “Pause” will cause the computation to pause. The
thread that does the computation should be started at the beginning of the program, but
should immediately go into the paused state until the “Run” button is pressed. Use the
wait() method in the thread to make it wait until “Run” is pressed. Use the notify()
method when the “Run” button is pressed to wake up the thread. Use a boolean signal
variable, running, to control whether the computation thread is paused. (The wait()
and notify() methods are covered in Subsection 12.3.5.)
Here is a picture of the program after it has run many trials:
You might want to start with a version of the program with no control button. In that
version, the computation thread can run continually from the time it is started. Once that
is working, you can add the button and the control feature.
To get you started, here is the code from the thread in my solution that runs one batch
of trials and updates the display labels:
for (int i = 0; i < BATCH SIZE; i++) {
double x = Math.random();
double y = Math.random();
trialCount++;
if (x*x + y*y < 1)
inCircleCount++;
}
double estimateForPi = 4 * ((double)inCircleCount / trialCount);
Platform.runLater( () -> {
countLabel.setText( " Number of Trials: " + trialCount);
piEstimateLabel.setText( " Current Estimate: " + estimateForPi);
} );
The variables trialCount and inCircleCount are of type long in order to allow the
number of trials to be more than the two billion or so that would be possible with a
variable of type int.
(I was going to ask you to use multiple computation threads, one for each available
processor, but I ran into an issue when using the Math.random() method in several threads.
Exercises 666
This method requires synchronization, which causes serious performance problems when
several threads are using it to generate large amounts of random numbers. A solution
to this problem is to have each thread use its own object of type java.util.Random to
generate its random numbers (see Subsection 5.3.1). My on-line solution to this exercise
discusses this problem further.)
7. The chat room example from Subsection 12.5.2 can be improved in several ways. First, it (solution)
would be nice if the participants in the chat room could be identified by name instead of
by number. Second, it would be nice if one person could send a private message to another
person that would be seen just by that person rather than by everyone. Make these two
changes. You can start with a copy of the package netgame.chat. You will also need the
package netgame.common, which defines the netgame framework.
To make the first change, you will have to implement a subclass of Hub that can keep
track of client names as well as numbers. To get the name of a client to the hub, you
can override the extraHandshake() method both in the Hub subclass and in the Client
subclass. The extraHandshake() method is called as part of setting up the connection
between the client and the hub. It is called after the client has been assigned an ID
number but before the connection is considered to be fully established. It should throw an
IOException if some error occurs during the setup process. Note that any messages that
are sent by the hub as part of the handshake must be read by the client and vice versa.
The extraHandshake() method in the Client is defined as:
protected void extraHandshake(ObjectInputStream in, ObjectOutputStream out)
throws IOException
while in the Hub, there is an extra parameter that tells the ID number of the client whose
connection is being set up:
protected void extraHandshake(in playerID, ObjectInputStream in,
ObjectOutputStream out) throws IOException
In the ChatRoomWindow class, the main() routine asks the user for the name of the
computer where the server is running. You can add some code there to ask the user their
name. (Just imitate the code that asks for the host name.) You will have to decide what
to do if two users want to use the same name.
For the second improvement, personal messages, I suggest writing a new PrivateMessage
class. A PrivateMessage object would include both the string that represents the message
and the ID numbers of the player to whom the message is being sent and the player who
sent the message. The hub will have to be programmed to know how to deal with such
messages. A PrivateMessage should only be sent by the hub to the client who is listed as
the recipient of the message. You need to decide how the user will input a private message
and how the user will select the recipient of the message. Don’t forget that PrivateMessage
needs to be declared to implement Serializable.
If you attempt this exercise, you are likely to find it quite challenging.
Quiz 667
Quiz on Chapter 12
(answers)
1. Write a complete subclass of Thread to represent a thread that writes out the numbers
from 1 to 10. Then write some code that would create and start a thread belonging to
that class.
2. Suppose that thrd is an object of type Thread. Explain the difference between calling
thrd.start() and calling thrd.run().
4. How does synchronization prevent race conditions, and what does it mean to say that
synchronization only provides mutual exclusion?
5. Suppose that a program uses a single thread that takes 4 seconds to run. Now suppose
that the program creates two threads and divides the same work between the two threads.
What can be said about the expected execution time of the program that uses two threads?
8. Network server programs are often multithreaded. Explain what this means and why it is
true.
9. Why does a multithreaded network server program often use many times more threads
than the number of available processors?
This appendix contains a list of the examples appearing in the free, on-line textbook
Introduction to Programming Using Java, Eighth Edition.
The web site for the book, http://math.hws.edu/javanotes, has links for downloading the
entire web site. If you do that, you will find the source code files in a directory named sources.
There is also a link for downloading just the source code. The README file from the download
includes some instructions for compiling and running the programs. Note however that some of
these examples depend on other source files, such as TextIO.java, that are not built into Java.
These are classes that I have written. All necessary files are included in the downloads, and
links to the individual files are provided below.
The solutions to end-of-chapter exercises are not listed in this appendix. Each end-of-
chapter exercise has its own Web page, which discusses its solution. The source code of a
sample solution of each exercise is given on the solution page for that exercise. If you want to
compile the solution, you should be able to copy-and-paste the solution out of a Web browser
window and into a text editing program. (You can’t copy-and-paste from the HTML source
of the solution page, since it contains extra HTML markup commands that the Java compiler
won’t understand; the HTML markup does not appear when the page is displayed in a Web
browser.) Exercise solutions are also available as a download from the front page of the web
site. The README file from the download has more information.
725
Source Code Listing 726
• EnumDemo.java, from Section 2.3, a very simple first demonstration of enum types.
The enum types used in this program are defined in the same file as the program. An
alternative version, SeparateEnumDemo.java, uses the same enums, but the enums are
defined in separate files, Day.java and Month.java
• PrintSquare.java, from Section 2.4, reads an integer typed in by the user and prints the
square of that integer. This program depends on TextIO.java. The same is true
for almost all of the programs in the rest of this list.
• Interest2.java, from Section 2.4, calculates interest on an investment for one year, based
on user input. Uses TextIO for user input.
• CreateProfile.java, from Section 2.4, a simple demo of output to a file, using TextIO.
• Interest2WithScanner.java, from Section 2.4, is a version of Interest2.java that uses Scanner
instead of TextIO to read input from the user.
• Interest3.java”, from Section 3.1, the first example that uses control statements.
• ThreeN1.java, from Section 3.2, outputs a 3N+1 sequence for a given stating value.
• ComputeAverage.java, from Section 3.3, computes the average value of some integers
entered by the user.
• CountDivisors.java, from Section 3.4, counts the number of divisors of an integer entered
by the user.
• ListLetters.java, from Section 3.4, lists all the distinct letters in a string entered by the
user.
• LengthConverter.java, from Section 3.5, converts length measurements input by the user
into different units of measure.
• ComputeAverage2.java, from Section 3.7, computes the average value of some real numbers
entered by the user. Demonstrates the use of try..catch for Double.parseDouble.
• AverageNumbersFromFile.java, from Section 3.7, finds the sum and the average of numbers
read from a file. Demonstrates the use of try..catch statements with TextIO.
• BirthdayProblem.java, from Section 3.8, demonstrates random access to array elements
using the “birthday problem” (how many people do you have to choose at random until
two are found whose birthdays are on the same day of the year).
• ReverseInputNumbers.java, from Section 3.8, illustrates the use of a partially full array
by reading some numbers from the user and then printing them in reverse order.
• GuessingGame.java, from Section 4.2, lets the user play guessing games where the com-
puter picks a number and the user tries to guess it. A slight variation of this program,
which reports the number of games won by the user, is GuessingGame2.java.
• RowsOfChars.java, from Section 4.3, a rather useless program in which one subroutine
calls another.
• CopyTextFile.java, from Section 4.3, demonstrates the use of command-line arguments by
using file names from the command line.
• ThreeN2.java, from Subsection 4.4.3, is an improved 3N+1 program that uses subroutines
and prints its output in neat columns.
• RollTwoPairs.java, from Subsection 5.2.2, uses PairOfDice.java to simulate rolling two
pairs of dice until the same total is rolled on both pairs.
Source Code Listing 727
• HighLow.java, from Section 5.4, a simple card game. It uses the classes Card.java and
Deck.java, which are given as examples of object-oriented programming. Also available,
the card-related classes Hand.java and, from Subsection 5.5.1, BlackjackHand.java.
• ReverseWithDynamicArray.java, from Section 7.2, reads numbers from the user then prints
them out in reverse order. It does this using the class DynamicArrayOfInt.java as an exam-
ple of using dynamic arrays. ReverseWithArrayList.java, from Section 7.3, is functionally
identical, but it uses an ArrayList<Integer> instead of a DynamicArrayOfInt.
• SymmetricMatrix.java, from Section 7.5, implements a symmetric 2D array of double.
The program TestSymmetricMatrix.java tests the SymmetricMatrix class.
• LengthConverter2.java, from Section 8.2, converts measurements input by the user to
inches, feet, yards, and miles. This improvement on LengthConverter.java allows inputs
combining several measurements, such as “3 feet 7 inches,” and it detects illegal inputs.
• TryStatementDemo.java, from Section 8.3, a small demo program with a try..catch
statement that includes autoclosing of a resource.
• LengthConverter3.java, from Section 8.3, is a revision of LengthConverter2.java that uses
exceptions to handle errors in the user’s input.
• TowersOfHanoi.java, from Section 9.1, prints out the steps in a solution to the Towers of
Hanoi problem; an example of recursion.
• StringList.java, from Section 9.2, implements a linked list of strings. The program List-
Demo.java tests this class.
• PostfixEval.java, from Section 9.3, evaluates postfix expressions using a stack. Depends
on the StackOfDouble class defined in StackOfDouble.java.
• SortTreeDemo.java, from Section 9.4, demonstrates a binary sort tree of strings.
• SimpleParser1.java, SimpleParser2.java, and SimpleParser3.java, from Section 9.5, are
three programs that parse and evaluate arithmetic expressions input by the user. Sim-
pleParser1 only handles fully parenthesized expressions. SimpleParser2 evaluates ordinary
expressions where some parentheses can be omitted. SimpleParser3 constructs expression
trees to represent input expressions and uses the expression trees to evaluate the expres-
sions.
• WordListWithTreeSet.java, from Section 10.2, makes an alphabetical list of words from a
file. A TreeSet is used to eliminate duplicates and sort the words.
• WordListWithPriorityQueue.java, from Section 10.2, makes an alphabetical list of words
from a file. This is a small modification of the previous example that uses a PriorityQueue
instead of a TreeSet. The result is an alphabetical list of words in which duplicates are
not removed.
• SimpleInterpreter.java, from Section 10.4, demonstrates the use of a HashMap as a symbol
table in a program that interprets simple commands from the user.
• WordCount.java, from Section 10.4, counts the number of occurrences of each word in a
file. The program uses several features from the Java Collection Framework.
• RiemannSumStreamExperiment.java, from Section 10.6, demos Java’s stream API. Runs
an experiment to measure the compute time for a problem when it is solved using a for
loop, using a sequential stream, and using a parallel stream.
• ReverseFileWithScanner.java, from Section 11.2, shows how to read and write files in a
simple command-line application. ReverseFileWithResources.java is a version that uses
the “resource” pattern in try..catch statements.
Source Code Listing 728
• DirectoryList.java, from Section 11.2, lists the contents of a directory specified by the user;
demonstrates the use of the File class.
• CopyFile.java, from Section 11.3, is a program that makes a copy of a file, using file names
that are given as command-line arguments. CopyFileAsResources.java is a version of the
program that also demonstrates uses the “resource” pattern in a try..catch statement.
• PhoneDirectoryFileDemo.java, from Section 11.3, demonstrates the use of a file for storing
data between runs of a program.
• FetchURL.java, from Section 11.4, reads and displays the contents of a specified URL, if
the URL refers to a text file.
• ShowMyNetwork.java, mentioned in Section 11.4, is a short program that prints informa-
tion about each network interface on the computer where it is run, including IP addresses
associated with each interface.
• DateClient.java and DateServer.java, from Section 11.4, are very simple first examples of
network client and server programs.
• CLChatClient.java and CLChatServer.java, from Section 11.4, demonstrate two-way com-
munication over a network by letting users send messages back and forth; however, no
threading is used and the messages must strictly alternate.
• ThreadTest1.java, from section Section 12.1, runs one or more threads that all perform the
same task, to demonstrate that they run simultaneously and finish in an indeterminate
order.
• ThreadTest2.java, from section Section 12.1, divides up a task (counting primes) among
several threads, to demonstrate parallel processing and the use of synchronization. Thread-
Test3.java, from the same section, is a minor modification of ThreadTest2.java that uses
an AtomicInteger instead of synchronization to safely add up values from several threads.
• DateServerWithThreads.java and DateServerWithThreadPool.java, from Section 12.4, are
modifications of chapter11/DateServer.java (Subsection 11.4.4) that use threads to handle
communication with clients. The first program creates a new thread for each connection.
The second uses a thread pool, and it uses a blocking queue to send connections from the
main program to the pool. The threaded servers will work with original client program,
chapter11/DateClient.java.
• CLMandelbrotMaster.java, CLMandelbrotWorker.java, and CLMandelbrotTask.java,
from Section 12.4, are a demonstration of distributed computing in which pieces of a
large computation are sent over a network to be computed by “worker” programs.
higher or lower in value. This program depends on Card.java, Hand.java, Deck.java, and
an image resource file, cards.png.
• MosaicDraw.java, from Section 6.6, demonstrates menus. In this program, the user colors
the squares of a mosaic by clicking-and-dragging the mouse. It uses MosaicCanvas.java to
define the mosaic itself.
• RandomStringsWithArray.java, from Section 7.2, shows multiple copies of a message in
random colors, sizes, and positions. There is an animation in which the copies move
around in the window. This is an improved version of RandomStrings.java that uses an
array to keep track of the data.
• SimplePaint2.java, from Section 7.3, lets the user draw colored curves and stores the data
needed for repainting the drawing surface in a list of type ArrayList<CurveData>.
• Life.java, from Section 7.5, implements John H. Conway’s game of life and is an example
of using 2D arrays. This program depends on MosaicCanvas.java.
• Checkers.java, from Section 7.5, lets two users play a game of checkers against each other.
Illustrates the use of a two-dimensional array and a variety of programming techniques.
(This is the longest program in the book so far, at over 700 lines!)
• Maze.java and LittlePentominos.java are demo programs mentioned in Section 9.1 as ex-
amples of recursion. They use techniques that have not covered until Chapter 12. Note
that LittlePentominos depends on MosaicCanvas.java.
• Blobs.java, from Section 9.1, uses recursion to count groups of colored squares in a grid.
• DepthBreadth.java, from Section 9.3, demonstrates stacks and queues.
• TrivialEdit.java, from Section 11.3, lets the user edit short text files. This program demon-
strates reading and writing files and using file dialogs.
• SimplePaintWithFiles.java, from Section 11.3, demonstrates saving data from a program
to a file in both binary and character form. The program is a simple sketching program
based on SimplePaint2.java.
• SimplePaintWithXML.java, from Section 11.5, demonstrate saving data from a program
to a file in XML format. This program is a modification of SimplePaintWithFiles.java.
• XMLDemo.java, from Section 11.5, is a simple program that demonstrates basic parsing
of an XML document and traversal of the Document Object Model representation of the
document. The user enters the XML to be parsed in a text area.
• RandomArtWithThreads.java, from Section 12.2, uses a thread to drive a very simple
animation.
• QuicksortThreadDemo.java, from Section 12.2, demonstrates using a separate thread to
perform a computation, with simple inter-thread communication.
• BackgroundComputationDemo.java, from Section 12.2, demonstrates using a thread run-
ning at a lower priority to perform a lengthy computation in the background.
• MultiprocessingDemo1.java, from Section 12.2, is a modification of the previous example
that uses several threads to perform the background computation. This speeds up the
computation on multi-processor machines.
• MultiprocessingDemo2.java, from Section 12.3, is a modification of the previous example
that decomposes its task into a large number of fairly small subtasks, in order to achieve
better load balancing. The program uses a thread pool and a queue of tasks.
Source Code Listing 731
• MultiprocessingDemo3.java, from Section 12.3, is yet another version of the previous ex-
amples. This one uses a pool of threads that run forever, taking tasks from a queue and
executing them. To make this possible, a blocking queue is used, defined by the standard
LinkedBlockingQueue class. MyLinkedBlockingQueue.java is a simple example of using
wait() and notify() directly that can be used as a replacement for LinkedBlockingQueue
in MultiprocessingDemo3.
• MultiprocessingDemo4.java, from Section 12.3 has the same functionality as
MultiprocessingDemo3, but uses an ExecutorService, from package java.util.concurrent,
to execute the tasks.
• TowersOfHanoiGUI.java, from Section 12.3, shows an animation of the famous Towers
Of Hanoi problem. The user can control the animation with Run/Pause, Next, and
StartAgain buttons. The program is an example of using wait() and notify() directly
for communication between threads.
• GUIChat.java, from Section 12.4, is a simple GUI program for chatting between two
people over a network. It demonstrates using a thread for reading data from a network
connection.
• netgame.common, from Section 12.5, is a package that defines a framework for networked
games. This framework is used in several examples: A chat room, defined in package
netgame.chat; a tic-tac-toe game, defined in package netgame.tictactoe; and a poker game,
defined in package netgame.fivecarddraw.
• BoundPropertyDemo.java, from Section 13.1, demonstrates bindings and bidirectional
bindings of JavaFX observable properties.
• CanvasResizeDemo.java, from Section 13.1, shows how to use property bindings to resize
a Canvas whenever the Pane that contains the Canvas is resized.
• StrokeDemo.java, from Section 13.2, demonstrates the use of various line properties for
stroking lines and rectangles.
• PaintDemo.java, from Section 13.2, demonstrates using a LinearGradient paint and us-
ing an ImagePattern paint to fill a polygon. Uses the image resource files tile.png and
face-smile.png.
• TransformDemo.java,from Section 13.2, demonstrates applying various transforms, such
as scale and rotate, to the drawing that is done in canvas. Uses the image resource file
face-smile.png.
• ToolPaint.java, from Section 13.2, is a little paint program that illustrates pixel manip-
ulation and the use of a transparent overlay canvas for some drawing operations. This
program requires SimpleDialogs.java.
• SillyStamper.java, from Section 13.3, demonstrates using a ListView whose items are Im-
ageViews. The user can “stamp” images of a selected icon onto a drawing area. This
program uses the icon images in the directory stamper icons as resources.
• SimpleTableDemo.java, from Section 13.3, is a small demo of an uneditable TableView.
• ScatterPlotTableDemo.java, from Section 13.3, demonstrates an editable TableView. The
table holds (x,y) coordinates of points, and the user can edit the coords. A scatter plot
of points is displayed.
• SimpleDialogs.java, from Section 13.4, contains static methods for showing several kinds
of dialog box. The program TestDialogs.java tests the dialog methods by letting the user
click buttons to open the different kinds of dialog.
Source Code Listing 732
• WebBrowser.java, from Section 13.4, is a simple web browser based on JavaFX’s WebView
control. This program shows how to manage multiple windows in a JavaFX application.
It requires BrowserWindow.java, a subclass of Stage that does most of the work, and on
SimpleDialogs.java.
• The Mandelbrot program from Section 13.5, which computes and displays visualizations
of the Mandelbrot set, is defined by several classes in the package edu.hws.eck.mdbfx.
The source code files can be found in the directory edu/hws/eck/mdbfx.
abstract class. A class that cannot be used to create objects, and that exists only for the
purpose of creating subclasses. Abstract classes in Java are defined using the modifier
abstract.
abstract data type (ADT). A data type for which the possible values of the type and the
permissible operations on those values are specified, without specifying how the values
and operations are to be implemented.
access specifier. A modifier used on a method definition or variable specification that deter-
mines what classes can use that method or variable. The access specifiers in Java are
public, protected, and private. A method or variable that has no access specifier is
said to have “package” visibility.
activation record. A data structure that contains all the information necessary to implement
a subroutine call, including the values of parameters and local variables of the subroutine
and the return address to which the computer will return when the subroutine ends.
Activation records are stored on a stack, which makes it possible for several subroutine
calls to be active at the same time. This is particularly important for recursion, where
several calls to the same subroutine can be active at the same time.
actual parameter. A parameter in a subroutine call statement, whose value will be passed
to the subroutine when the call statement is executed. Actual parameters are also called
“arguments”.
address. Each location in the computer’s memory has an address, which is a number that
identifies that location. Locations in memory are numbered sequentially. In modern
computers, each byte of memory has its own address. Addresses are used when informa-
tion is being stored into or retrieved from memory.
algorithm. An unambiguous, step-by-step procedure for performing some task, which is guar-
anteed to terminate after a finite number of steps.
alpha color component. A component of a color that says how transparent or opaque that
color is. The higher the alpha component, the more opaque the color.
ALU. Arithmetic Logic Unit. The ALU is the part of the CPU that performs arithmetic
operations such as addition and subtraction and logical operations such as AND and
OR.
API. Application Programming Interface. A specification of the interface to a software package
or “toolbox.” The API says what classes or subroutines are provided in the toolbox and
how to use them.
applet. A type of Java program that is meant to run on a Web page in a Web browser, as
opposed to a stand-alone application.
735
Glossary 736
blocking queue. A queue in which the dequeue operation will block if the queue is empty,
until an item is added to the queue. If the blocking queue has a limited capacity, the
enqueue operation can also block, if the queue is full.
bottom-up design. An approach to software design in which you start by designing basic
components of the system, then combine them into more complex components, and so
on.
BufferedImage. A class representing “off-screen canvases,” that is, images that are stored in
the computer’s memory and that can be used for drawing images off-screen.
branch. A control structure that allows the computer to choose among two or more different
courses of action. Java has two branch statements: if statements and switch statements.
byte. A unit of memory that consists of eight bits. One byte of memory can hold an eight-bit
binary number.
bytecode. “Java bytecode” is the usual name for the machine language of the Java Virtual
Machine. Java programs are compiled into Java bytecode, which can then be executed
by the JVM.
charset. A particular encoding of character data into binary form. Examples include UTF-8
and ISO-8859-1.
checked exception. An exception in Java that must be handled, either by a try..catch
statement or by a throws clause on the method that can throw he exception. Failure to
handle a checked exception in one way or the other is a syntax error.
class. The basic unit of programming in Java. A class is a collection of static and non-static
methods and variables. Static members of a class are part of the class itself; non-static,
or “instance,” members constitute a blueprint for creating objects, which are then said
to “belong” to the class.
class invariant. A statement about the state of a class, or of an object created from that
class, that is always true. Any method in a class, to be correct, must preserve the truth
of all class invariants.
class variables and class methods. Alternative terms for “static variables” and “static
methods”, which are part of the class itself rather than of objects.
client/server. A model of network communication in which a “server” waits at a known
address on the network for connection requests that are sent to the server by “clients.”
This is the basic model for communication using the TCP/IP protocol.
command-line interface. A way of interacting with the computer in which the user types in
commands to the computer and the computer responds to each command.
comment. In a computer program, text that is ignored by the computer. Comments are for
human readers, to help them understand the program.
compiler. A computer program that translates programs written in some computer language
(generally a high-level language) into programs written in machine language.
component. General term for a visual element of a GUI, such as a window, button, or menu.
constructor. A special kind of subroutine in a class whose purpose is to construct objects
belonging to that class. A constructor is called using the new operator, and is not
considered to be a “method.”
container. A component, such as a BorderPane, that can contain other GUI components.
Glossary 738
contract of a method. The semantic component of the method’s interface. The contract
specifies the responsibilities of the method and of the caller of the method. It says how
to use the method correctly and specifies the task that the method will perform when
it is used correctly. The contract of a method should be fully specified by its Javadoc
comment.
control structure. A program structure such as an if statement or a while loop that affects
the flow of control in a program (that is, the order in which the instructions in the
program are executed).
CPU. Central Processing Unit. The CPU is the part of the computer that actually performs
calculations and carries out programs.
CSS. Cascading Style Sheets, a language that can be used to control the visual appearance of
components in JavaFX or of elements on a web page.
data structure. An organized collection of data, that can be treated as a unit in a program.
deadlock. A situation in which several threads hang indefinitely, for example because each of
them is waiting for some resource that is locked by one of the other threads.
default method. A method in a Java interface that has an implementation. The default
implementation is used in any class that implements the interface but does not override
the method. Default methods are marked with the reserved word default. Not supported
in Java 7 and earlier.
default package. The unnamed package. A class that does not declare itself to be in a named
package is considered to be in the default package.
definite assignment. Occurs at a particular point in a program if it is definitely true that
a given variable must have been assigned a value before that point in the program. It
is only legal to use the value of a local variable if that variable has “definitely” been
assigned a value before it is used. For this to be true, the compiler must be able to verify
that every path through the program from the declaration of the variable to its use must
pass through a statement that assigns a value to that variable.
deprecated. Considered to be obsolete, but still available for backwards compatibility. A
deprecated Java class or method is still part of the Java language, but it is not advisable
to use it in new code. Deprecated items might be removed in future versions of Java.
dialog box. A window that is dependent on another window, called its parent owner. Dialog
boxes are usually popped up to get information from the user or to display a message to
the user.
distributed computing. A kind of parallel processing in which several computers, connected
by a network, work together to solve a problem.
dummy parameter. Identifier that is used in a subroutine definition to stand for the value of
an actual parameter that will be passed to the subroutine when the subroutine is called.
Dummy parameters are also called “formal parameters” (or sometimes just “parameters,”
when the term “argument” is used instead of actual parameter).
enum. Enumerated type. A type that is defined by listing every possible value of that type.
An enum type in Java is a class, and the possible values of the type are objects.
event. In GUI programming, something that happens outside the control of the program, such
as a mouse click, and that the program must respond to when it occurs.
Glossary 739
exception. An error or exceptional condition that is outside the normal flow of control of a
program. In Java, an exception can be represented by an object of type Throwable that
can be caught and handled in a try..catch statement.
factory method. A method, usually a static function, that returns an object. Factory methods
are an alternative to constructors.
fetch-and-execute cycle. The process by which the CPU executes machine language pro-
grams. It fetches (that is, reads) an instruction from memory and carries out (that is,
executes) the instruction, and it repeats this over and over in a continuous cycle.
fill. A drawing operation that applies a color (or other type of fill) to each of the pixels inside
a shape.
flag. A boolean value that is set to true to indicate that some condition or event is true. A
single bit in a binary number can also be used as a flag.
formal parameter. Another term for “dummy parameter.”
frame. One of the images that make up an animation. Also used as another name for activation
record.
function. A subroutine that returns a value.
functional interface. A Java interface that defines only a single subroutine (where the term
“interface” here means an interface that defines a Java type.)
garbage collection. The automatic process of reclaiming memory that is occupied by objects
that can no longer be accessed.
generic programming. Writing code that will work with various types of data, rather than
with just a single type of data. The Java Collection Framework, and classes that use
similar techniques, are examples of generic programming in Java.
getter. An instance method in a class that is used to read the value of some property of that
class. Usually the property is just the value of some instance variable. By convention, a
getter is named getXyz() where xyz is the name of the property.
global variable. Another name for member variable, emphasizing the fact that a member
variable in a class exists outside the methods of that class.
graphics context. The data and methods necessary for drawing to some particular desti-
nation. A graphics context in JavaFX is an object belonging to the GraphicsContext
class.
GUI. Graphical User Interface. The modern way of interacting with a computer, in which the
computer displays interface components such as buttons and menus on a screen and the
user interacts with them—for example by clicking on them with a mouse.
hash table. A data structure optimized for efficient search, insertion, and deletion of objects.
A hash table consists of an array of locations, and the location in which an object is stored
is determined by that object’s “hash code,” an integer that can be efficiently computed
from the contents of the object.
heap. The section of the computer’s memory in which objects are stored.
high level language. A programming language, such as Java, that is convenient for hu-
man programmers but that has to be translated into machine language before it can be
executed.
HSB. A color system in which colors are specified by three numbers (in Java, real numbers in
the range 0.0 to 1.0) giving the hue, saturation, and brightness.
Glossary 740
JRE. Java Runtime Environment. Basic software that supports running standard Java pro-
grams that have already been compiled. A JRE includes a Java Virtual Machine and all
the standard Java classes.
just-in-time compiler. A kind of combination interpreter/compiler that compiles parts of a
program as it interprets them. This allows subsequent executions of the same parts of
the program to be executed more quickly than they were the first time. This can result
is greatly increased speed of execution. Modern JVMs use a just-in-time compiler.
JVM. Java Virtual Machine. The imaginary computer whose machine language is Java byte-
code. Also used to refer to computer programs that act as interpreters for programs
written in bytecode; to run Java programs on your computer, you need a JVM.
lambda expression. A notation that defines an anonymous method. More precisely, a lambda
expression is a kind of literal that represents a value whose type is given by a functional
interface.
linked data structure. A collection of data consisting of a number of objects that are linked
together by pointers which are stored in instance variables of the objects. Examples
include linked lists and binary trees.
linked list. A linked data structure in which nodes are linked together by pointers into a linear
chain.
listener. In GUI programming, an object that can be registered to be notified when events of
some given type occur. The object is said to “listen” for the events.
literal. A sequence of characters that is typed in a program to represent a constant value. For
example, ’A’ is a literal that represents the constant char value, A, when it appears in
a Java program.
location (in memory). The computer’s memory is made up of a sequence of locations. These
locations are sequentially numbered, and the number that identifies a particular location
is called the address of that location.
local class. A class that is defined inside a method definition. Local classes are often anony-
mous, but that is not required.
local variable. A variable declared within a method, for use only inside that method. A
variable declared inside a block is valid from the point where it is declared until the end
of block in which the declaration occurs.
loop. A control structure that allows a sequence of instructions to be executed repeatedly.
Java has three kinds of loops: for loops, while loops, and do loops
loop control variable. A variable in a for loop whose value is modified as the loop is executed
and is checked to determine whether or not to end the loop.
loop invariant. A statement such that, if the statement is true before a loop executes, then
it will remain true after each execution of the loop, and therefore will still be true after
the loop ends. Loop invariants can be a tool for proving correctness of loops.
machine language. A programming language consisting of instructions that can be executed
directed by a computer. Instructions in machine language are encoded as binary num-
bers. Each type of computer has its own machine language. Programs written in other
languages must be translated into a computer’s machine language before they can be
executed by that computer.
main memory. Programs and data can be stored in a computer’s main memory, where
they are available to the CPU. Other forms of memory, such as a disk drive, also store
Glossary 742
information, but only main memory is directly accessible to the CPU. Programs and data
from a disk drive have to be copied into main memory before they can be used by the
CPU.
map (data structure). An associative array; a data structure that associates an object from
some collection to each object in some set. In Java, maps are represented by the generic
interface Map<T,S>
map (stream operator). One of the fundamental operations on streams, defined as part
of Java’s stream API. A map operator applies a function to each element of a stream,
producing a new stream consisting of the values output by the function.
member variable. A variable defined in a class but not inside a method, as opposed to a local
variable, which is defined inside some method.
memory. Memory in a computer is used to hold programs and data.
method. Another term for subroutine, used in the context of object-oriented programming. A
method is a subroutine that is contained in a class or in an object.
method reference. A notation for a lambda expression that represents a method that al-
ready exists in some class or object. A method reference uses the :: operator, such as
Math::sqrt.
module. In general, a component of a larger system that interacts with the rest of the system
in a simple, well-defined, straightforward manner. In Java 9 and later, a module is a
collection of Java packages, allowing explicit control of dependencies between different
modules, and the standard Java packages have been divided into q set of modules.
multitasking. Performing multiple tasks at once, either by switching rapidly back and forth
from one task to another or by literally working on multiple tasks at the same time.
multiprocessing. Multitasking in which more than one processor is used, so that multiple
tasks can literally be worked on at the same time.
mutual exclusion. Prevents two threads from accessing the same resource at the same time.
In Java, this only applies to threads that access the resource in synchronized methods or
synchronized statements. Mutual exclusion can prevent race conditions but introduces
the possibility of deadlock.
MVC pattern. The Model/View/Controller pattern, a strategy for dividing responsibility in
a GUI component. The model is the data for the component. The view is the visual
presentation of the component on the screen. The controller is responsible for reacting
to events by changing the model. According to the MVC pattern, these responsibilities
should be handled by different objects.
NaN. Not a Number. Double.NaN is a special value of type double that represents an undefined
or illegal value.
node. Common term for one of the objects in a linked data structure.
null. A special pointer value that means “not pointing to anything.”
numerical analysis. The field that studies algorithms that use approximations, such as real
numbers, and the errors that can result from such approximation.
off-by-one error. A common type of error in which one too few or one too many items are
processed, often because counting is not being handled correctly or because the processing
stops too soon or continues too long for some other reason.
Glossary 743
object. An entity in a computer program that can have data (variables) and behaviors (meth-
ods). An object in Java must be created using some class as a template. The class of an
object determines what variables and methods it contains.
object type. A type whose values are objects, as opposed to primitive types. Classes and
interfaces are object types.
observable value. A value that generates an event when it is modified, so that observers of
the value can be notified of the change and can react to it.
OOP. Object-Oriented Programming. An approach to the design and implementation of com-
puter programs in which classes and objects are created to represent concepts and entities
and their interactions.
operating system. The basic software that is always running on a computer, without which
it would not be able to function. Examples include Linux, Mac OS, and Windows Vista.
operator. A symbol such as “+”, “<=”, or “++” that represents an operation that can be
applied to one or more values in an expression.
overloading (of operators). The fact that the same operator can be used with different types
of data. For example, the “+” operator can be applied to both numbers and strings.
overloading (of method names). The fact that several methods that are defined in the
same class can have the same name, as long as they have different signatures.
overriding. Redefining in a subclass. When a subclass provides a new definition of a method
that is inherited from a superclass, the new definition is said to override the original
definition.
package. In Java, a named collection of related classes and sub-packages, such as java.io and
javafx.scene.control.
parallel processing. When several tasks are being performed simultaneously, either by mul-
tiple processors or by one processor that switches back and forth among the tasks.
parameter. Used to provide information to a subroutine when that subroutine is called. Values
of “actual parameters” in the subroutine call statement are assigned to the “dummy
parameters” in the subroutine definition before the code in the subroutine is executed.
parameterized type. A type such as ArrayList<String> that includes one or more type pa-
rameters (String in the example).
parsing. Determining the syntactical structure of a string in some language. To parse a string
is to determine whether the string is legal according to the grammar of the language, and
if so, how it can be created using the rules of the grammar.
partially full array. An array that is used to store varying numbers of items. A partially full
array can be represented as a normal array plus a counter to keep track of how many
items are actually stored.
pixel. A “picture element” on the screen or in an image. A picture consists of rows and
columns of pixels. The color of each pixel can be individually set.
polymorphism. The fact that the meaning of a call to an instance method can depend on
the actual type of the object that is used to make the call at run time. That is, if
var is a variable of object type, then the method that is called by a statement such as
var.action() depends on the type of the object to which var refers when the statement
is executed at run time, not on the type of variable var.
Glossary 744
pointer. A value that represents an address in the computer’s memory, and hence can be
thought of as “pointing” to the location that has that address. A variable in Java can
never hold an object; it can only hold a pointer to the location where the object is stored.
A pointer is also called a “reference.”
pragmatics. Rules of thumb that describe what it means to write a good program. For
example, style rules and guidelines about how to structure a program are part of the
pragmatics of a programming language.
precedence. The precedence of operators determines the order in which they are applied,
when several operators occur in an expression, in the absence of parentheses.
precondition. A condition that must be true at some point in the execution of a program,
in order for the program to proceed correctly from that point. A precondition of a
subroutine is something that must be true when the subroutine is called, in order for the
subroutine to function properly. Subroutine preconditions are often restrictions on the
values of the actual parameters that can be passed into the subroutine.
predicate. A function that outputs a boolean value. Predicates in Java can be represented
by the parameterized functional interface Predicate<T>.
priority queue. A data structure representing a collection of items where each item has a
“priority.” A priority queue has operations add and remove. Items can be added in any
order, but the remove operation always removes an item of minimal priority. (Some
version of priority queue use maximum instead of minimum priority.)
postcondition. A condition that is known to be true at some point in the execution of a
program, as a result of the computation that has come before that point. A postcondition
of a subroutine is something that must be true after the subroutine finishes its execution.
A postcondition of a function often describe the return value of the function.
primitive type. One of the eight basic built-in data types in Java, double, float, long, int,
short, byte, boolean, and char. A variable of primitive type holds an actual value, as
opposed to a pointer to that value.
priority of a thread. An integer associated with a thread that can affect the order in which
threads are executed. A thread with greater priority is executed in preference to a thread
with lower priority.
producer/consumer. A classic pattern in parallel programming in which one or more pro-
ducers produce items that are consumed by one or more consumers, and the producers
and consumers are meant to run in parallel. The problem is to get items safely and
efficiently from the producers to the consumers. In Java, the producer/consumer pattern
is implemented by blocking queues.
program. A set of instructions to be carried out by a computer, written in an appropriate
programming language. Used as a verb, it means to create such a set of instructions.
programming language. A language that can be used to write programs for a computer. Pro-
gramming languages range in complexity from machine language to high-level languages
such as Java.
protocol. A specification of what constitutes legal communication in a give context. A protocol
specifies the format of legal messages, when they can be sent, what kind of reply is
expected, and so on.
pseudocode. Informal specification of algorithms, expressed in language that is closer to
English than an actual programming language, and usually without filling in every detail
Glossary 745
of the procedure.
queue. A data structure consisting of a list of items, where items can only be added at one
end and removed at the opposite end of the list.
race condition. A source of possible errors in parallel programming, where one thread can
cause an error in another thread by changing some aspect of the state of the program
that the second thread is depending on (such as the value of variable).
RAM. Random Access Memory. This term is often used as a synonym for the main memory
of a computer. Technically, however, it means memory in which all locations are equally
accessible at any given time. The term also implies that data can be written to the
memory as well as read from it.
recursion. Defining something in terms of itself. In particular, a recursive subroutine is
one that calls itself, either directly, or indirectly through a chain of other subroutines.
Recursive algorithms work by reducing a complex problem into smaller problems which
can be solved either directly or by applying the same algorithm “recursively.”
reduce (stream operator). One of the fundamental operations on a stream. A reduce
operation combines all the elements from a stream in some way, such as by summing
them or finding their maximum, producing a final result.
RGB. A color system in which colors are specified by three numbers (in Java, integers in the
range 0 to 255) giving the red, green, and blue components of the color.
reference. Another term for “pointer.”
return type of a function. The type of value that is returned by that function.
reserved word. A sequence of characters that looks like an identifier but can’t be used as an
identifier because it has a special meaning in the language. For example, class, public,
and if are reserved words in Java.
resource. An image, sound, text, or other data file that is part of a program. Resource files
for Java programs are stored on the same class path where the compiled class files for
the program are stored.
robust program. A program is robust if it is not only correct, but also is capable of handling
errors such as a non-existent file or a failed network connection in a reasonable way.
scene graph. In JavaFX, the hierarchical data structure that contains all the GUI components
that are shown in a window.
set. A collection of objects which contains no duplicates. In Java, sets are represented by the
generic interface Set<T>
scope. The region in a program where the declaration of an identifier is valid.
semantics. Meaning. The semantics rules of a language determine the meaning of strings of
symbols (such as sentences or statements) in that language.
sentinel value. A special value that marks the end of a sequence of data values, to indicate
the end of the data.
setter. An instance method in a class that is used to set the value of some property of that
class. Usually the property is just the value of some instance variable. By convention, a
setter is named setXyz() where xyz is the name of the property.
signature of a method. The name of the method, the number of formal parameters in its
definition, and the type of each formal parameter. Method signatures are the information
Glossary 746
needed by a compiler to tell which method is being called by a given subroutine call
statement.
socket. An abstraction representing one end of a connection between two computers on a
network. A socket represents a logical connection between computer programs, not a
physical connection between computers.
stack. A data structure consisting of a list of items where items can only be added and removed
at one end of the list, which is known as the “top” of the stack. Adding an item to a
stack is called “pushing,” and removing an item is called “popping.” The term stack also
refers to the stack of activation records that is used to implement subroutine calls.
standard input. The standard source from which a program reads input data. It is represented
by the object System.in. Usually, standard input comes from text typed by the user, but
standard input can be “redirected” to read from another source, such as a file, instead.
standard output. The standard destination to which a program writes output text. It is
represented by the object System.out. Usually, standard output is displayed to the user,
but standard output can be “redirected” to write to another destination, such as a file,
instead. There is also an object System.err that is meant for writing error messages.
state machine. A model of computation where an abstract “machine” can be in any of some
finite set of different states. The behavior of the machine depends on its state, and the
state can change in response to inputs or events. The basic logical structure of a GUI
program can often be represented as a state machine.
step-wise refinement. A technique for developing an algorithm by starting with a general
outline of the procedure, often expressed in pseudocode, and then gradually filling in the
details.
stream. In Java 8, an abstraction representing a stream of values that can be processed. A
stream can be created from a Collection, an array, or some other data source. Java’s
stream API includes many predefined operations that can be applied to streams. The
term “stream” also refers to I/O streams, which are used for input and output.
stroke. A drawing operation that applies a color (or other type of paint) to pixels along the
boundary of a shape.
source code. Text written in a high-level programming language, which must be translated
into a machine language such as Java bytecode before it can be executed by a computer.
subclass. A class that extends another class, directly or indirectly, and therefore inherits its
data and behaviors. The first class is said to be a subclass of the second.
subroutine. A sequence of program instructions that have been grouped together and given a
name. The name can then be used to “call” the subroutine. Subroutines are also called
methods in the context of object-oriented programming.
subroutine call statement. A statement in a program that calls a subroutine. When a
subroutine call statement is executed, the computer executes the code that is inside the
subroutine.
super. A special variable, automatically defined in any instance method, that refers to the
object that contains the method, but considered as belonging to the superclass of the class
in which the method definition occurs. super gives access to members of the superclass
that are hidden by members of the same name in the subclass.
syntax. Grammar. The syntax rules of a language determine what strings of symbols are
legal—that is, grammatical—in that language.
Glossary 747
TCP/IP. Protocols that are used for network communication on the Internet.
this. A special variable, automatically defined in any instance method, that refers to the object
that contains the method.
thread. An abstraction representing a sequence of instructions to be executed one after the
other. It is possible for a computer to execute several threads in parallel.
thread pool. A collection of “worker threads” that are available to perform tasks. As tasks
become available, they are assigned to threads in the pool. A thread pool is often used
with a blocking queue that holds the tasks.
top-down design. An approach to software design in which you start with the problems, as a
whole, subdivide it into smaller problems, divide those into even smaller problems, and
so on, until you get to problems that can be solved directly.
type. Specifies some specific kind of data values. For example, the type int specifies integer
data values that can be represented as 32-bit binary numbers. In Java, a type can be
a primitive type, a class names, or an interface name. Type names are used to specify
the types of variables, of dummy parameters in subroutines, and of return values of
subroutines.
type cast. Forces the conversion of a value of one type into another type. For example, in
(int)(6*Math.random()), the (int) is a type-cast operation that converts the double
value (6*Math.random()) into an integer by discarding the fractional part of the real
number.
Unicode. A way of encoding characters as binary numbers. The Unicode character set includes
characters used in many languages, not just English. Unicode is the character set that is
used internally by Java.
URL. Universal Resource Locator; an address for a resource on the Internet, such as a web
page.
variable. A named memory location (or sequence of locations) that can be used to store
data. A variable is created in a program, and a name is assigned to the variable, in a
variable declaration statement. The name can then be used in that program to refer
to the memory location, or to the data stored in that memory location, depending on
context. In Java, a variable has a type, which specifies what kind of data it can hold.
wrapper class. A class such as Double or Integer that makes it possible to “wrap” a primitive
type value in an object belonging to the wrapper class. This allows primitive type
values to be used in contexts were objects are required, such as with the Java Collection
Framework.
XML. eXtensible Markup Language. A very common and well-supported standard syntax for
creating text-based data-representation languages.