The following amendments apply:

Rule 5: Brackets must be brought down to next line, so that opening

and closing brackets are indented by the same number of spaces.
Rule 5: Indentation must be done using four spaces.
Rule 66: Always use opening and closing brackets on an empty loop.

The Elements of Java Style

The Elements of Java Style

Al Vermeulen
Scott W. Ambler
Greg Bumgardner
Eldon Metz
Trevor Misfeldt
Jim Shur
Patrick Thompson

1
PUBLISHED BY CAMBRIDGE UNIVERSITY PRESS (VIRTUAL PUBLISHING)
FOR AND ON BEHALF OF THE PRESS SYNDICATE OF THE UNIVERSITY OF
CAMBRIDGE

PUBLISHED BY THE PRESS SYNDICATE OF THE UNIVERSITY OF

CAMBRIDGE
The Pitt Building, Trumpington Street, Cambridge, United Kingdom

CAMBRIDGE UNIVERSITY PRESS

The Edinburgh Building, Cambridge CB2 2RU, UK http://www.cup.cam.ac.uk
40 West 20th Street, New York, NY 10011-4211, USA http://www.cup.org
10 Stamford Road, Oakleigh, Melbourne 3166, Australia
Ruiz de Alarcn 13, 28014 Madrid, Spain

Published in association with SIGS Books

Cambridge University Press 2000

This edition Cambridge University Press (Virtual Publishing) 2001

All rights reserved.

This book is in copyright. Subject to statutory exception and to the provisions of relevant
collective licensing agreements, no reproduction of any part may take place without the
written permission of Cambridge University Press.

Any product mentioned in this book may be a trademark of its company.

First published in 2000

Reprinted in 2000

Design and composition by David Van Ness

Cover design by Andrea Cammarata

Printed in the United States of America

A catalog record for this book is available from the British Library.

Library of Congress Cataloging in Publication data is on record with the publisher.

ISBN 0 521 77768 2 paperback

e-ISBN 0 511 00339 0 virtual (netLibrary Edition)

2
 The authors would like to thank
our loved ones for enduring us
while we toiled away on this book.

3
Table of Contents
Preface ix
Audience x
Acknowledgments xi
Introduction xiii
1. General Principles 1
2. Formatting Conventions 5
3. Naming Conventions 15
Package Names 18
Type Names 20
Class Names 20
Interface Names 22
Method Names 23
Variable Names 25
Field Names 27
Parameter Names 28
Constant Names 29
4. Documentation Conventions 31
Comment Types 32
Documentation Comments 36
Comment Style 38
Comment Content 49
Internal Comments 52
5. Programming Conventions 57
Type Safety 64
Statements and Expressions 66
Construction 70
Exception Handling 72
Assertions 75
Concurrency 79
Synchronization 80
Efficiency 85
6. Packaging Conventions 89
Summary 95
Glossary 105
Bibliography 119
Index 123

4
Preface

AT ROGUE WAVE, we sell C++ and Java software components. We have always
included source code with our products. Customers often browse through the code to get
a feeling, not just for how it works, but for how to write good software. As a result, we
have always felt pressuremaybe more pressure than most companiesto have good,
consistent style throughout our source code.

As the company grew, making sure programmers were all following the same rules
became difficult. To address this, our founder and first programmer, Tom Keffer, wrote
35 pages that explained how we write C++ code at Rogue Wave. We passed the
document around and made sure new hires got a copy. It worked. When customers asked
how we maintained consistency in our coding, we told them about Toms C++ Design,
Implementation, and Style Guide, and sent them a copy. Word spread and we turned
Toms document into a technical report. We sent out thousands of copies and received
terrific positive feedback.

When Java came along, we decided we needed a document like the C++ Guide. A note
went out to our internal [email protected] mailing list soliciting rules for Java use
that we should be using. The resulting list of rules became the first draft of the Rogue
Wave Java Style Guide.

As the list of rules grew, the style guide began to look more and more like a real book.
This time, we decided to publish our guide instead of simply issuing another Rogue
Wave technical report. To our amazement, the folks at Cambridge University Press
thought this was a great idea, and The Elements of Java Style was born.

One of the first reviewers of that original draft was Scott Ambler, current president of
Ronin International (www.ronin-intl.com). Scott liked the idea of the book and suggested
we check out the coding standards for Java hed been distributing on the Web. We liked
his standards a lot and decided we should work as a team. The result of combining
Scotts standards and the Rogue Wave style document is this book.

Audience

We wrote this book for anyone writing Java code, but especially for programmers who
are writing Java as part of a team. For a team to be effective, everyone must be able to
read and understand everyone elses code. Having consistent style conventions is a good
first step!

We assume you already know the basics of Java and object-oriented programming.

5
Acknowledgments

THIS BOOK was a team effort. The team extends far beyond the seven named authors.
Wed like to thank those who reviewed and contributed to the original Rogue Wave Java
Style Guide and the Ambysoft Inc. Coding Standards for Java. This includes Jeremy
Smith, Tom Keffer, Wayne Gramlich, Pete Handsman, and Cris Perdue.

This book would certainly never have happened without the help and encouragement of
the folks at Cambridge University Press. Our editor, Lothlrien Homet, hooked the
Rogue Wave people up with Scott Ambler and made it all happen with her gentle, yet
persistent, prodding. Thanks Lothlrien!

6
Introduction

style: 1b. the shadow-producing pin of a sundial.

2c. the custom or plan followed in spelling, capitalization, punctuation, and
typographic arrangement and display
Websters New Collegiate Dictionary

THE SYNTAX OF A PROGRAMMING LANGUAGE tells you what code it is possible

to writewhat the machine will understand. Style tells you what you ought to write
what the humans reading the code will understand. Code written with a consistent, simple
style will be maintainable, robust, and contain fewer bugs. Code written with no regard to
style will contain more bugs. It may simply be thrown away and rewritten rather than
maintained.

Our two favorite style guides are classics: Strunk and Whites The Elements of Style and
Kernighan and Plaugers The Elements of Programming Style. These small books work
because they are simplea list of rules, each containing a brief explanation and examples
of correct, and sometimes incorrect, use. We followed the same pattern in this book.

This simple treatmenta series of rulesenabled us to keep this book short and easy to
understand. The idea is to provide a clear standard to follow, so programmers can spend
their time on solving the problems of their customers, instead of worrying about naming
conventions and formatting.

7
1.
General Principles

While it is important to write software that performs well, many other issues should
concern the professional Java developer. All good software performs well. But great
software, written with style, is predictable, robust, maintainable, supportable, and
extensible.

1. Adhere to the style of the original.

When modifying existing software, your changes should follow the style of the original
code.1 Do not introduce a new coding style in a modification, and do not attempt to
rewrite the old software just to make it match the new style. The use of different styles
within a single source file produces code that is more difficult to read and comprehend.
Rewriting old code simply to change its style may result in the introduction of costly yet
avoidable defects.

2. Adhere to the Principle of Least Astonishment.

The Principle of Least Astonishment suggests you should avoid doing things that will
surprise a user of your software. This implies the means of interaction and the behavior
exhibited by your software must be predictable and consistent,2 and, if not, the
documentation must clearly identify and justify any unusual patterns of use or behavior.

To minimize the chances that a user will encounter something surprising in your software,
you should emphasize the following characteristics in the design, implementation, and
documentation of your Java software:

Build simple classes and simple methods. Determine how

Simplicity
much you need to do to meet the expectations of your users.
Ensure each class, interface, method, variable, and object
Clarity has a clear purpose. Explain where, when, why, and how to
use each.
Provide the minimum functionality that any reasonable user
Completeness would expect to find and use. Create complete
documentation; document all features and functionality.
Similar entities should look and behave the same; dissimilar
Consistency entities should look and behave differently. Create and
apply standards whenever possible.
Provide predictable documented behavior in response to
Robustness errors and exceptions. Do not hide errors and do not force
clients to detect errors.

8
3. Do it right the first time.

Apply these rules to any code you write, not just code destined for production. More
often than not, some piece of prototype or experimental code will make its way into a
finished product, so you should anticipate this eventuality. Even if your code never
makes it into production, someone else may still have to read it. Anyone who must look
at your code will appreciate your professionalism and foresight at having consistently
applied these rules from the start.

4. Document any deviations.

No standard is perfect and no standard is universally applicable. Sometimes you will find
yourself in a situation where you need to deviate from an established standard.

Before you decide to ignore a rule, you should first make sure you understand why the
rule exists and what the consequences are if it is not applied. If you decide you must
violate a rule, then document why you have done so.

This is the prime directive.

________________

1 Jim Karabatsos. When does this document apply? In Visual Basic Programming
Standards. (GUI Computing Ltd., 22 March 1996). Accessed online at
http://www.gui.com.au/jkcoding.htm, Aug 1999.

2 George Brackett. Class 6: Designing for Communication: Layout, Structure,

Navigation for Nets and Webs. In Course T525: Designing Educational Experiences
for Networks and Webs. (Harvard Graduate School of Education, 26 August 1999).
Accessed online at http://hgseclass.harvard.edu/t52598/classes/class6/, Aug 1999.

9
 Amendments:
* Brackets must be brought down to next line, so that
opening and closing brackets are indented by the
2. same number of spaces.
Formatting Conventions * Indentation must be done using four spaces.

5. Indent nested code.

One way to improve code readability is to group individual statements into block
statements and uniformly indent the content of each block to set off its contents from the
surrounding code.

If you generate code using a Java development environment, use the indentation style
produced by the environment. If you are generating the code by hand, use two spaces to
ensure readability without taking up too much space: four

class MyClass {
void function(int arg) {
if (arg < 0) {
for (int index = 0; index <= arg; index++) {
//
}
}
}
}

In addition to indenting the contents of block statements, you should also indent the
statements that follow a label to make the label easier to notice:

void function(int arg) {

loop:
for (int index = 0; index <= arg; index++) {
switch (index) {

case 0:
//
break loop; // exit the for statement
default:
//
break; // exit the switch statement
}
}
}

Locate the opening brace { of each block statement in the last character position of the
line that introduced the block. Place the closing brace } of a block on a line of its own,
aligned with the first character of the line that introduced the block. The following
examples illustrate how this rule applies to each of the various Java definition and control
constructs.

Class definitions:

10
public class MyClass {
...
}

Inner class definitions:

public class OuterClass {

...
class InnerClass {
...
}
...
}

Method definitions:

void method(int j) {
...
}

Static blocks:

static {
...
}

For-loop statements:

for (int i = 0; i <= j; i++) {

...
}

If and else statements:

if (j < 0) {
...
}
else if (j > 0) {
...
}
else {
...
}

Try, catch, and finally blocks:

try {
...
}
catch (Exception e) {
...

11
}
finally {
...
}

Switch statements:

switch (value) {
case 0:
...
break;
default:
...
break;
}

Anonymous inner classes:

button.addActionListener(
new ActionEventListener() {

public void actionPerformed() {

...
}
}
)

While statements:

while (++k <= j) {

...
}

Do-while statements:

do {
...
} while (++k <= j);

If you are managing a development team, do not leave it up to individual developers to

choose their own indentation amount and style. Establish a standard indentation policy
for the organization and ensure that everyone complies with this standard.

Our recommendation of two spaces appears to be the most common standard, although
your organization may prefer three or even four spaces. Amendments:
* Indentation must be done using
four spaces.

12
6. Break up long lines.

While a modern window-based editor can easily handle long source code lines by
scrolling horizontally, a printer must truncate, wrap, or print on separate sheets any lines
that exceed its maximum printable line width. To ensure your source code is still readable
when printed, you should limit your source code line lengths to the maximum width your
printing environment supports, typically 80 or 132 characters.

First, do not place multiple statement expressions on a single line if the result is a line
that exceeds your maximum allowable line length. If two statement expressions are
placed on one line:

double x = Math.random(); double y = Math.ran

dom(); // Too Long!

Then introduce a new line to place them on separate lines:

double x = Math.random();
double y = Math.random();

Second, if a line is too long because it contains a complex expression:

double length = Math.sqrt(Math.pow(Math.random(),

2.0) + Math.pow(Math.random(), 2.0)); // Too Long!

Then subdivide that expression into several smaller subexpressions. Use a separate line to
store the result produced by an evaluation of each subexpression into a temporary
variable:

double xSquared = Math.pow(Math.random(),2.0);

double ySquared = Math.pow(Math.random(),2.0);
double length = Math.sqrt(xSquared+ySquared);

Last, if a long line cannot be shortened under the previous two guidelines, then break,
wrap, and indent that line using the following rules:

Step one

If the top-level expression on the line contains one or more commas:

double length = Math.sqrt(Math.pow(x, 2.0),Math

.pow(y,2.0)); // Too Long!

Then introduce a line break after each comma. Align each expression following a comma
with the first character of the expression proceeding the comma:

double length = Math.sqrt(Math.pow(x,2.0),

Math.pow(y,2.0));

13
Step two

If the top-level expression on the line contains no commas:

class MyClass {
private int field;

...
boolean equals(Object obj) {
return this == obj || (obj instanceof MyClass
&& this.field == obj.field); // Too Long!
}
...
}

Then introduce a line break just before the operator with the lowest precedence or, if
more than one operator of equally low precedence exists between each such operator:

class MyClass {
private int field;
...
boolean equals(Object obj) {
return this == obj
|| (this.obj instanceof MyClass
&& this.field == obj.field);
}
...
}

Step three

Reapply steps one and two, as required, until each line created from the original
statement expression is less than the maximum allowable length.

7. Include white space.

White space is the area on a page devoid of visible characters. Code with too little white
space is difficult to read and understand, so use plenty of white space to delineate
methods, comments, code blocks, and expressions clearly.

Use a single space to separate:

A right parenthesis ) or curly brace } and any keyword that immediately

follows, a keyword and any left paren-

thesis ( or curly brace { that immediately follows, or a right parenthesis ) and

any left curly brace { that immediately follows:

14
 for(...){
...
}

while(...){
...
}

do{
...
}while(...);

switch(...){
...
}

if(...){
...
}
elseif(...){
...
}
else{
...
}

try{
...
}
catch(...){
...
}
finally{
...
}

Any binary operator, except for the . qualification operator and the expression
that proceeds and the expression that follows the operator:

double length = Math.sqrt(x * x + y * y);

double xNorm = length > 0.0 ? (x / length) : x;

Use blank lines to separate:

Each logical section of a method implementation:

void handleMessage(Message messsage) {

DataInput content = message.getDataInput();

int messageType = content.readInt();

switch (messageType) {

15
 case WARNING:
... do some stuff here ...
break;

case ERROR:
... do some stuff here ...
break;

default:
... do some stuff here ...
break;

}
}

Each member of a class and/or interface definition:

public class Foo {

/**
* Defines an inner class.
*/
class InnerFoo {
...
}

/**
* The Bar associated with this Foo.
*/
private Bar bar;

/**
* Construct a Foo with the specified Bar.
*/
Foo(Bar bar) {
this.bar = bar;
}
}
Each class and interface definition in a source file:

/*
* ... file description ...
*/

package com.company.xyz;

/**
* ... interface description ...
*/
interface FooInterface {
...
}

16
 /**
* ... class description ...
*/
public class Foo implements FooInterface {
...
}

8. Do not use hard tabs.

Many developers use tab characters to indent and align their source code, without
realizing the interpretation of tab characters varies across environments. Code that
appears to possess the correct formatting when viewed in the original editing
environment can appear unformatted and virtually unreadable when transported to an
environment that interprets tabs differently.

To avoid this problem, always use spaces instead of tabs to indent and align source code.
You may do this simply by using the space bar instead of the tab key or by configuring
your editor to replace tabs with spaces. Some editors also provide a smart indentation
capability. You will need to disable this feature if it uses tab characters.

Your organization should set a common indentation size and apply it consistently to all
its Java code as outlined in Rule #5.

17
3.
Naming Conventions

THE FOLLOWING naming conventions are identical to those used by SUN

MICROSYSTEMS in naming the identifiers that appear in the Java Software
Development Kit.3,4

9. Use meaningful names.

When you name a class, variable, method, or constant, use a name that is, and will remain,
meaningful to those programmers who must eventually read your code. Use meaningful
words to create names. Avoid using a single character or generic names that do little to
define the purpose of the entities they name.

The purpose for the variable a and the constant 65 in the following code is unclear:

if (a < 65) { // What property does 'a' describe?

y = 65 - a; // What is being calculated here?
}
else {
y = 0;
}

The code is much easier to understand when meaningful names are used:

if (age < RETIREMENT_AGE) {

yearsToRetirement = RETIREMENT_AGE - age;
}
else {
yearsToRetirement = 0;
}

The only exception to this rule concerns temporary variables whose context provides
sufficient information to determine their purpose, such as a variable used as a counter or
index within a loop:

for (int i = 0; i < numberOfStudents; ++i) {

enrollStudent(i);
}

Some variable meanings and use scenarios occur frequently enough to be standardized.

(See Rule #28 for more information.)

10. Use familiar names.

18
Use words that exist in the terminology of the target domain. If your users refer to their
clients as customers, then use the name Customer for the class, not Client. Many
developers will make the mistake of creating new or generic terms for concepts when
satisfactory terms already exist in the target industry or domain.

11. Question excessively long names.

The name given an object must adequately describe its purpose. If a class, interface,
variable, or method has an overly long name, then that entity is probably trying to
accomplish too much.

Instead of simply giving the entity a new name that conveys less meaning, first
reconsider its design or purpose. A refactoring of the entity may produce new classes,
interfaces, methods, or variables that are more focused and can be given more meaningful
yet simpler names.

12. Join the vowel generation.

Abbreviations reduce the readability of your code and introduce ambiguity if more than
one meaningful name reduces to the same abbreviation.

Do not attempt to shorten names by removing vowels. This practice reduces the
readability of your code and introduces ambiguity if more than one meaningful name
reduces to the same consonants.

The casual reader can understand the names in this definition:

public Message appendSignature(Message message,

String signature) {
...
}

While the shortened forms are more difficult to read:

public Msg appndSgntr(Msg msg,

String sgntr) {
...
}

If you remove vowels simply to shorten a long name, then you need to question whether
the original name is appropriate. See Rule #11.

13. Capitalize only the first letter in acronyms.

19
This style helps to eliminate confusion in names where uppercase letters act as word
separators, and it is especially important if one acronym immediately follows another:

setDSTOffset() setDstOffset()
loadXMLDocument() loadXmlDocument()

This rule does not apply to:

Acronyms that appear within the name of a constant as these names only contain
capital letters (see Rule #31):

static final String XML_DOCUMENT = "text/XML";

Acronyms that appear at the beginning of a method, variable or parameter name,

as these names should always start with a lowercase letter (see Rules #22 and
#25):

private Document xmlDocument;

14. Do not use names that differ only in case.

The Java compiler can distinguish between names that differ only in case, but a human
reader may fail to notice the difference.

For example, a variable named theSQLInputStream should not appear in the same
scope as a variable named theSqlInputStream. If both names appear in the same scope,
each effectively hides the other when considered from the perspective of a person trying
to read and understand the code.5

Package Names

15. Use the reversed, lowercase form of your organizations Internet domain
name as the root qualifier for your package names.

Any package distributed to other organizations should include the lowercase domain
name of the originating organization, in reverse order.6 For example, if a company
named ROGUE WAVE SOFTWARE, whose Internet domain name is roguewave.com,
decides to distribute an application server package called server, then Rogue Wave would
name that package com.roguewave.server.

SUN MICROSYSTEMS has placed restrictions on the use of the package qualifier names
java and javax. The java package qualifier may only be used by Java vendors to provide
conforming implementations of the standard Java class libraries. SUN

20
MICROSYSTEMS reserves the name javax for use in qualifying its own Java extension
packages.

16. Use a single, lowercase word as the root name of each package.

The qualified portion of a package name should consist of a single, lowercase word that
clearly captures the purpose and utility of the package. A package name may consist of a
meaningful abbreviation. Examples of acceptable abbreviations are the standard Java
packages of java.io and java.net.

17. Use the same name for a new version of a package, but only if that new
version is still binary compatible with the previous version, otherwise, use a new
name.

The intent of this rule is to ensure that two Java classes with identical qualified names
will be binary and behaviorally compatible with each other.

The Java execution model binds the clients of a class to implementation of that class at
run time. This means unless you adopt this convention, you have no way to ensure that
your application is using the same version of the software you had used and tested with
when you built the application.

If you produce a new version of a package that is not binary or behaviorally compatible,
you should change the name of the package. This renaming may be accomplished in a
variety of ways, but the safest and easiest technique is simply to add a version number to
the package name and then increment that version number each time an incompatible
change is made:

com.roguewave.server.v1
com.roguewave.server.v2
:

The one drawback to this approach is the dependency between a client of a package and a
specific implementation of that package is hard-coded into the client code. A package
client can only be bound to a new version of that package by modifying the client code.

Type Names

18. Capitalize the first letter of each word that appears in a class or interface
name.

21
The capitalization provides a visual cue for separating the individual words within each
name. The leading capital letter provides a mechanism for differentiating between class
or interface names and variable names (see Rule #25):

public class PrintStream

extends FilterOutputStream {
...
}
public interface ActionListener
extends EventListener {
...
}

Class Names.
19. Use nouns when naming classes.

Classes define objects, or things, which are identified by nouns:

class CustomerAccount {
...
}
public abstract class KeyAdapter
implements KeyListener {
...
}

20. Pluralize the names of classes that group related attributes, static services, or
constants.

Give classes that group related attributes, static services, or constants a name that
corresponds to the plural form of the attribute, service, or constant type defined by the
class.

The java.awt.font.LineMetrics class is an example of a class that defines an object that

manages a group of related attributes:

/**
* The <code>LineMetrics</code> class gives
* access to the metrics needed to layout
* characters along a line and to layout of
* a set of lines.
*/
public class LineMetrics {
public LineMetrics()
public abstract int getNumChars();
public abstract float getAscent();
public abstract float getDescent();
public abstract float getLeading();
public abstract float getHeight();

22
 ...
}

The java.beans.Beans class is an example of a class that defines a group of related static
services:

/**
* The <code>Beans</code> class provides some
* general purpose beans control methods.
*/

public class Beans {

public static Object instantiate(...) {...}
public static Object getInstanceOf(...) {...}
public static boolean isInstanceOf(...) {...}
public static boolean isDesignTime() {...}
public static boolean isGuiAvailable() {...}
public static void setDesignTime(...) {...}
public static void setGuiAvailable(...) {...}
...
}

The java.sql.Types class is an example of a class that defines a group of related static
constants:

/**
* The <code>Types</code> class defines constants
* that are used to identify SQL types.
*/
public class Types {
public final static int BIT = 7;
public final static int TINYINT = 6;
public final static int SMALLINT = 5;
public final static int INTEGER = 4;
public final static int BIGINT = 5;
public final static int FLOAT = 6;
public final static int REAL = 7;
public final static int DOUBLE = 8;
public final static int NUMERIC = 2;
public final static int DECIMAL = 3;
public final static int CHAR = 1;
...
}

Interface Names
21. Use nouns or adjectives when naming interfaces.

An interface provides a declaration of the services provided by an object, or it provides a

description of the capabilities of an object.

23
Use nouns to name interfaces that act as service declarations:

public interface ActionListener {

public void actionPerformed(ActionEvent e);
}

Use adjectives to name interfaces that act as descriptions

of capabilities. Most interfaces that describe capabilities
use an adjective created by tacking an able or ible
suffix onto the end of a verb:

public interface Runnable {

public void run();
}
public interface Accessible {
public Context getContext();
}

Method Names

22. Use lowercase for the first word and capitalize only the first letter of each
subsequent word that appears in a method name.

The capitalization provides a visual cue for separating the individual words within each
name. The leading lowercase letter provides a mechanism for differentiating between a
method invocation and a constructor invocation:

class MyImage extends Image {

public MyImage() {
...
}
public void flush() {
...
}
public Image getScaledInstance() {
...
}
}

23. Use verbs when naming methods.

Methods and operations commonly define actions, which are verbs.

class Account {
private int balance;
...
public void withdraw(int amount) {

24
 deposit(-1 * amount);
}
public void deposit(int amount) {
this.balance += amount;
}
}

24. Follow the JavaBeansTM conventions for naming property accessor

methods.

The JavaBeans specification7 establishes standard naming conventions8 for methods

that give access to the properties of a JavaBean implementation. You should apply these
conventions when naming methods in any class, regardless of whether it implements a
Bean.

A JavaBean exposes boolean properties using methods that begin with is:

boolean isValid() {
return this.isValid;
}

A JavaBean gives read access to other property types using methods that begin with get:

String getName() {
return this.name;
}

The accessor method for reading an indexed property takes an int index argument:

String getAlias(int index) {

return this.aliases[index];
}

A JavaBean gives write access to boolean and other types of properties using methods
that begin with set:

void setValid(boolean isValid) {

this.isValid = isValid;
}
void setName(String name) {
this.name = name;
}

The accessor method for setting an indexed property takes an int index argument:

25
void setAlias(int index, String alias) {
this.aliases[index] = alias;
}

The Java Development Kit strongly adheres to these conventions. The is/get/set notation
is required for exposing the component properties of a Bean unless you define a BeanInfo
class.9,10

Variable Names

25. Use lowercase for the first word and capitalize only the first letter of each
subsequent word that appears in a variable name.

The capitalization provides a visual cue for separating the individual words within each
name. The leading lowercase letter provides a mechanism for differentiating between
variable names and class names (see Rule #18):

class Customer {
...
private Address address;

private Phone daytimePhone;

...
public Address setAddress(Address address) {
Address oldAddress = this.address;
this.address = address;
return oldAddress;
}
...
public void setDaytimePhone(Phone daytimePhone);
...
}
...
}

26. Use nouns to name variables.

Variables refer to objects, or things, which are identified by nouns:

class Customer {
...
private Address billingAddress;
private Address shippingAddress;
private Phone daytimePhone;
private Vector openOrders;
...
}

27. Pluralize the names of collection references.

26
Give fields and variables that refer to collections of objects a name that corresponds to
the plural form of the object type contained in the collection. This enables a reader of
your code to distinguish between variables representing multiple values from those
representing single values:

Customer[] customers =
newCustomer[MAX_CUSTOMERS];
void addCustomer(int index, Customer customer) {
this.customers[index] = customer;
}

Vector orderItems = new Vector();

void addOrderItem(OrderItem orderItem) {

this.orderItems.addElement(orderItem);
}

28. Establish and use a set of standard names for trivial throwaway variables.

You should use full descriptive names for most variables, but many variable types that
appear frequently within Java code have common shorthand names, which you may
choose to use instead.11, 12 The following table lists a few examples:

Character c, d, e
Coordinate x, y, z
Exception e
Graphics g
Object o
Stream in, out, inOut
String s

Field Names
29. Qualify field variables with this to distinguish them from local variables.

To make distinguishing between local variables and field variables easier, always qualify
field variables using this:

public class AtomicAdder {

private int count;

public AtomicAdder(int count) {

27
 this.count = count;
}

public synchronized int fetchAndAdd(int value) {

int temp = this.count;
this.count += value;
return temp;
}

public synchronized int addAndFetch(int value) {

return this.count += value;
}
}

Parameter Names
30. When a constructor or set method assigns a parameter to a field, give that
parameter the same name as the field.

While hiding the names of instance variables with local variables is generally poor style,
in this case some benefits exist. Using the same name relieves you of the responsibility
for coming up with a name that is different. Using the same name also provides a subtle
clue to the reader that the parameter value is destined for assignment to the field of the
same name.

class Dude {
private String name;

public Dude(String name) {

this.name = name;
}
public setName(String name) {
this.name = name;
}
}

Constant Names

31. Use uppercase letters for each word and separate each pair of words with an
underscore when naming constants.

The capitalization of constant names distinguishes them from other nonfinal variables:

class Byte {
public static final byte MAX_VALUE = 255;
public static final byte MIN_VALUE = 0;
public static final Class TYPE = Byte.class;
}

28
________________

3 Sun Microsystems. Java Code Conventions. (Palo Alto, California: Sun

Microsystems Inc., 20 April 1999). Accessed online at
ftp://ftp.javasoft.com/docs/codeconv/CodeConventions.pdf, Aug 1999.

4 Sun Microsystems. Java 2 Platform, Standard Edition, v1.2.2 API Specification.

(Sun Microsystems Inc., 1999). Accessed online at
http://java.sun.com/products/jdk/1.2/docs/api/index.html, Aug 1999.

5 Jonathan Nagler. Coding Style and Good Computing Practices. The Political
Methodologist, Vol. 6, No. 2 (Spring 1995). Accessed online at
http://wizard.ucr.edu/~nagler/coding_style.html, Aug 1999.

6 James Gosling et al. The Java Language Specification. (Reading, Massachusetts:

Addison-Wesley, 1996), pp. 125126.

7 To be called Java Beans or Beans for the remainder of this book.

8 Sun Microsystems. JavaBeans API Specification., ed. Graham Hamilton, (Mountain

View, California: Sun Microsystems Inc., 1997), pp. 5457. Accessed online at
http://www.javasoft.com/beans/docs/beans.101.pdf, Aug 1999.

9 Ibid., pp. 5657.

10 Patrick Chan, and Rosanna Lee. The Java Class Libraries, Volume 2: java.applet,
java.awt, java.beans., 2nd Edition. (Reading, Massachusetts: AddisonWesley, 1998), pg.
132.

11 James Gosling et al. The Java Language Specification.

12 Sun Microsystems. Java Code Conventions.

29
4.
Documentation Conventions
32. Write documentation for those who must use your code and those who must
maintain it.

Document the public programming interface of your code so others can use it correctly
and effectively. Document the private interface and internal implementation details of
your code so others can maintain and enhance it.

Always assume someone who is completely unfamiliar with your code will eventually
have to read and understand it. In fact, if enough time passes, your own code may
become unfamiliar, so this person may even be you!

33. Keep comments and code in sync.

When the code and the comments disagree, both are probably wrong.Norm Schryer,
Bell Labs

When you modify code, make sure you also update any related comments.13 The code
and documentation together form a software product, so treat each with equal importance.

34. Use the active voice and omit needless words.

Comments are a form of prose. Forceful, clear, and concise language is especially
beneficial for technical documentation. Use it.

Comment Types

Java supports three comment types:

A documentation comment that starts with /**and ends with */:

/**
* A documentation comment.
*/

A standard, or C-style, comment, which starts with /* and ends with */:

/*
* A standard comment.
*/

30
 A one-line, or end-line, comment that begins with // and continues through the
end of the line.

// A one-line comment.
class MyClass {
int myField; // An end-line comment.
...
}

Each comment type serves a particular purpose and you should use each type in a manner
consistent with that purpose.

35. Use documentation comments to describe the programming interface.

You may place documentation comments in front of any class, interface, method,
constructor, or field declaration that appears in your code. These comments provide
information the Javadoc utility uses to generate HTML-formatted class reference or
Application Programming Interface (API) documentation. To create this documentation,
the Javadoc utility reads these comments as it parses all the declarations that

appear in a set of Java source code files. This information is used by Javadoc to produce a
corresponding set of HTML pages that describe, by default, all the public and protected
classes, inner classes, interfaces, constructors, methods, and fields found in those files.14

Javadoc only recognizes documentation comments when they appear immediately before
a class, interface, constructor, method, or field declaration. Javadoc ignores any
documentation comments that appear within the body of a method, so do not use them in
this manner. Javadoc allows only one documentation comment per declaration statement,
so do not attempt to use more than one comment block per declaration.

The primary purpose for documentation comments is to define a programming

contract15 between a client and a supplier of a service. The documentation associated
with a method should describe all aspects of behavior on which a caller of that method
can rely and should not attempt to describe implementation details.

The following code illustrates the use of documentation comments to document a class
that declares an inner-class, a field, a method, and a constructor:

/**
* The <code>Rectangle2D</code> class describes
* a rectangle defined by location (x,y) and
* dimensions (w,h).
* ...
*/
public abstract class Rectangle2D
extends RectangularShape {

31
 /**
* The <code>Double</code> class defines a
* rectangle specified in double coordinates...
*/
static class Double extends Rectangle2D {...}
...

/**
* The bitmask that indicates that a point lies
* below this Rectangle2D...
*/
static int OUT_BOTTOM;
...
/**
* Adds a Rectangle2D to this Rectangle2D...
*/
public void add(Rectangle2D r) {...}
...
/**
* This is an abstract class that cannot be
* instantiated directly...
*/
protected Rectangle2D() {...}
...
}

36. Use standard comments to hide code without removing it.

Use standard C-style comments when you wish to hide code temporarily from the
compiler without actually removing it from the source file. You may only use this
comment type to comment-out a section of code that does
not have another comment block embedded within it.

To avoid any problems with nested comment blocks, because they look very much like
documentation comments, you should not use this type of comment for any purpose other
than for temporarily hiding code.

The following code fragment demonstrates how to use this comment type to hide a
member function definition:

/**
* ...
* @deprecated
*/

/*
I have temporarily removed this method because

32
 it has been deprecated for some time, and I
want to determine whether any other packages
are still using it! - J. Kirk on 9 Dec 1997

public void thisOldFunction() {

// There has got to be a better way!
...
}
*/

37. Use one-line comments to explain implementation details.

Use one or more one-line comments to document:

The purpose of specific variables or expressions.

Implementation-level design decisions.
The source material for complex algorithms.
Defect fixes or workarounds.
Code that may benefit from further optimization or elaboration.
Known problems, limitations, or deficiencies.

Strive to minimize the need for embedded comments by writing code that documents
itself. Do not add comments that simply repeat what the code does.16 Add comments
only if they add useful information:

double totalCost; // Used to total invoice.

...

// Apply the discount to all invoices over $1000.

if (totalCost > 1000.0) { // :TODO: Use constant?

// The discount is hard-coded because current

// customers all use the same discount rate.

// We will need to replace this constant with a

// variable if we ever get a customer who needs
// a different rate, or one that wants to apply
// multiple discount rates!

totalCost = totalCost * DISCOUNT;

}

Documentation Comments

38. Describe the programming interface before you write the code.

The best time to create API reference documentation is early in the development process.
Use documentation or doc comments to define the purpose, use, and behavior of each
class or interface that forms part of a potential design solution. Write these comments

33
while the purpose and rationale for introducing the new type is still fresh in your mind.
Do not think you must wait to complete the implementation of every method before
generating documentationthe Javadoc utility can run on Java source files containing
classes whose methods are simple stubs with no method bodies. This means you can
write documentation comments and run Javadoc in the earliest stages of implementation,
before writing any method bodies.17

The initial description of a type and its methods and fields should not only provide
guidance to the developers who must implement that type, but also form the basis for the
final API reference documentation for that type. A developer tasked with implementing a
class may choose to elaborate on the original documentation when the implementation
details of the public interface become better defined and more apparent.

39. Document public, protected, package, and private members.

Supply documentation comments for all members, including those with package,
protected, and private access. This allows for the generation of detailed, implementation-
level documentation.18 The developer who must learn and understand your code before
implementing an enhancement or bug fix will appreciate your foresight in providing
quality documentation for all class members, not just for the public ones.

40. Provide a summary description and overview for each package.

The Javadoc utility provides a mechanism for including package descriptions in the
documentation it generates. Use this capability to provide a summary description and
overview for each package you create.

To create a description for a package, you must create a package comment file, named
package.html, and place that file in the package directory along with the other package
source files. Javadoc will automatically look for a filename in this location.

The package comment file contains HTML, not Java source code. The package
description must appear within an HTML <body> element. Javadoc treats the first
sentence or phrase that appears within the <body>. . .</body> tags as the summary
description for the package, just as it does when processing normal documentation
comments.

You may use any Javadoc tag within the package description, except for the {@link} tag.
Any @see tags that appear in a package description must use fully qualified names.

41. Provide a summary description and overview for each application or group
of packages.

34
The Javadoc utility provides a mechanism for including a package-independent overview
description in the documentation it generates. Use this capability to provide an overview
description for each application or group of related packages you create.

To create an overview description, you must create an overview comment file, that may
be given any name that ends in .html, such as overview.html. To include this file in
your documentation, you must tell Javadoc where to find the file by using the -overview
option.

The overview comment file contains HTML, not Java source code. The overview must
appear within an HTML <body> element. Javadoc treats the first sentence or phrase that
appears between the <body></body> tags as the summary description for the
application or package group, just as it does when processing normal documentation
comments.

You may use any Javadoc tag within the description, except for the {@link} tag. Any
@see tags that appear in an overview description must use fully qualified names.

Comment Style

The doc comment formatting conventions listed in this section closely follow the
conventions adopted and published by SUN MICROSYSTEMS.19

42. Use a single consistent format and organization for all documentation
comments.

A properly formatted documentation comment contains a description followed by one or

more Javadoc tags.

Format each documentation comment as follows:

Indent the first line of the comment to align the slash character of the start-
comment symbol /** with the first character in the line containing the
associated definition.
Begin each subsequent line within an asterisk *. Align this asterisk with the first
asterisk in the start-comment symbol.
Use a single space to separate each asterisk from any descriptive text or tags that
appear on the same line.
Insert a blank comment line between the descriptive text and any Javadoc tags
that appear in the comment block.
End each documentation comment block with the asterisk in the end-comment
symbol */ aligned with the other asterisks in the comment block:

35
/**
* Descriptive text for this entity.
*
* @tag Descriptive text for this tag.
*/

The following rules specify additional guidelines for creating high-quality, maintainable
documentation comments.

43. Wrap keywords, identifiers, and constants with <code></code> tags.

Nest keywords, package names, class names, interface names, method names, field
names, parameter names, constant names, and constant values that appear in a
documentation comment within HTML <code></code> mark-up tags:

/**
* Allocates a <code>Flag</code> object
* representing the <code>value</code> argument.
* ...
*/
public Flag(boolean value) {...}

The <code></code> tags tell HTML browsers to render the content in a different
style20 than that of normal text, so these elements will stand out.

44. Wrap code with <pre></pre> tags.

Nest code that appears in a documentation comment within HTML <pre></pre> mark-
up tags:

/**
* ...
* The following example uses a
* <code>Class</code> object to print the class
* name of an object:
*
* <pre>
* void printClassName(Object o) {
* System.out.println("The class of "
* + o
* + " is "
* + o.getClass().getName());
* }
* </pre>
* ...
*/
public final class Class {...}

The <pre></pre> tags are used to tell HTML browsers to retain the original formatting,
including indentation and line-ends, of the preformatted element.

36
45. Consider marking the first occurrence of an identifier with a {@link} tag.

Each package, class, interface, method, and field name that appears within a
documentation comment may be converted into a hypertext link by replacing that name
with an appropriately coded21 {@link} tag:

/**
* Allocates a <code>Flag</code> object
* representing the <code>value</code> argument.
* Use this form of constructor as an alternative
* to the {@link #Flag(String)} form.
* ...
*/
public Flag(boolean value) {...}

/**
* Allocates a <code>Flag</code> object
* representing the value <code>true</code> if
* the string argument is not <code>null</code>
* and is equal to the string "true".
* Use this form of constructor as an alternative
* to the {@link #Flag(boolean)} form.
* ...
*/
public Flag(String s) {...}

Do not feel you must create links to every identifier that appears within a comment block.
Creating a significant number of these links can quickly clutter a comment block with
{@link} tags, making it hard to read and maintain the original source.

Create links only when the documentation associated with the referenced element would
truly be of interest or value to the reader. Do not create links for every occurrence of an
identifier. If an identifier appears more than once, simply create a link for the first
occurrence and mark any subsequent instances using the mark-up discussed in Rule #43.
Also, some classes and methods are so commonly used by a proficient Java programmer,
they do not warrant their own linksagain, these identifiers should still be marked as
indicated in Rule #43.

46. Establish and use a fixed ordering for Javadoc tags.

SUN MICROSYSTEMS recommends the following Javadoc tag ordering22:

In classes and interface descriptions:

37
 /**
* Description.
*
* @author
* @version
*
* @see
* @since
* @deprecated
*/

Consider including an @author and @version tag in every class or interface

description.

List multiple @author tags in chronological order, with the class or interface
creator listed first.

In method descriptions:

/**
* Description.
*
* @param
* @return
* @exception
*
* @see
* @since
* @deprecated
*/

Include a @param tag for every parameter. List multiple @param tags in
parameter declaration order.

Include a @return tag if the method returns any type other than void.

Include an @exception tag for every checked exception listed in a throws clause.
Include an @exception tag for every unchecked exception that a user may
reasonably expect to catch. List multiple @exception tags in alphabetical order of
the exception class names.

In field descriptions:

/**
* Description.
*
* @see
* @since
* @deprecated
*/

38
Sort multiple @see tags according to their distance from the current location, in terms
of document navigation and name qualification. Order each group of overloaded methods
according to the number of parameters each accepts, starting with the method that has the
least number of parameters:

/**
* ...
* @see #field
* @see #Constructor()
* @see #Constructor(Type...)
* @see #method()
* @see #method(Type...)
* @see Class
* @see Class#field
* @see Class#Constructor()
* @see Class#Constructor(Type...)
* @see Class#method()
* @see Class#method(Type...)
* @see package.Class
* @see package.Class#field
* @see package.Class#Constructor()
* @see package.Class#Constructor(Type...)
* @see package.Class#method()
* @see package.Class#method(Type...)

* @see package
* @see <a href="URL#label">label</a>
* @see "String"
* ...
*/

47. Write in the third-person narrative form.

When describing the purpose and behavior of classes, interfaces, and methods, use third-
person pronounssuch as they and itand third-person verb formssuch as
sets and getsinstead of second-person formssuch as set and get.

Some of the third-person verb forms that commonly appear in API documentation
include

adds deallocates removes

allocates destroys returns
computes gets sets
constructs provides tests
converts reads writes

48. Write summary descriptions that stand alone.

39
The Javadoc utility uses the first sentence or phrase in a documentation comment as a
summary description of the class, interface, method, or field that immediately follows the
comment block. To locate the end of the summary description, Javadoc starts at the
beginning of the comment block and searches for a period that is followed by a space, tab,
or end-of-line, or a Javadoc tag, whichever comes first.

Because this text provides a summary description for some entity, it must present a clear,
simple, and concise description of that entity. Do not rely on other sentences in the
comment block to provide additional context or elaboration.

Consider the following example:

/**
* Use this function sparingly!
* Applies the Foo-Bar algorithm to this node.
*/
public void doFooBar() {...}

Processing this code with Javadoc produces the following summary description for the
doFooBar method:

Use this function sparingly!

Reordering the comment block produces a superior summary sentence:

/**
* Applies the Foo-Bar algorithm to this node.
* Use this function sparingly!
*/
public void doFooBar() {...}

If the entity is an overloaded method, the summary description must differentiate that
method from the other forms of the same method:

/**
* Allocates a <code>Flag</code> object
* representing the <code>value</code> argument.
* ...
*/
public Flag(boolean value) {...}

/**
* Allocates a <code>Flag</code> object
* representing the value true if the
* string argument is not null and is equal
* to the string "true".
* ...
*/
public Flag(String s) {...}

40
49. Omit the subject in summary descriptions of actions or services.

A summary description does not require a subject because the subject can be determined
from the context in which the description appears.

The following descriptions incorrectly provide a redundant identification of the subject:

/**
* This method applies the Foo-Bar
* algorithm to this node.
* ...
*/
public void doFooBar() {...}

/**
* The <code>doFooBar</code> method applies the
* Foo-Bar algorithm to this node.
* ...
*/
public void doFooBar() {...}

The following description correctly omits the subject:

/**
* Applies the Foo-Bar algorithm to this node.
* ...
*/
public void doFooBar() {...}

50. Omit the subject and the verb in summary descriptions of things.

A summary description for a class, interface, or field that represents a thing does not
require an explicit subject or verb, as the description needs only to identify an object. A
subject is unnecessary because it can be determined from the context. A verb is
unnecessary because it simply states the subject is, exists as, or represents some
object.

The following example illustrates a description that contains an unnecessary subject and
verb23:

/**
* A thread group represents a set of threads.
* ...
*/
public class ThreadGroup {...}

Drop the subject and verb to obtain the correct form of summary description:

41
/**
* A set of threads.
* ...
*/
public class ThreadGroup {...}

51. Use this rather than the when referring to instances of the current class.

When describing the purpose or behavior of a method, use this instead of the to refer
to an object that is an instance of the class defining the method:

/**
* Returns a <code>String</code> representing the
* value of the <code>Flag</code> object.
* ...
*/
public String toString() {...}

/**
* Returns a <code>String</code> representing the
* value of this <code>Flag</code> object.
* ...
*/
public String toString() {...}

52. Do not add parentheses to a method or constructor name unless you want to
specify a particular signature.

A method or constructor reference should not include any parentheses unless the
reference identifies an overloaded method or constructor and you wish to refer to a single
form of the overloaded operation.

Do not add an empty pair of parenthesis () to indicate a name refers to a method. This
practice causes confusion if the name is associated with an overloaded method and one of
the overloaded forms of that method takes no arguments.

Consider the following pair of overloaded methods, as defined by the java.lang.String

class:

public class String {

...
public String toLowerCase() {...}
public String toLowerCase(Locale locale) {...}
...
}

42
If you use the identifier toLowerCase() to refer to any or all of the toLowerCase
methods, then you will likely confuse those who read your documentation. Most likely,
your users will think you meant the first form of the method, not either or all forms of it.
Use parentheses only when you want to specify the exact signature of a method or
constructor:

toLowerCase Refers to either or both forms of the method.

toLowerCase() Refers only to the first form of the method.
toLowerCase(Locale) Refers only to the second form of the method.

Comment Content
53. Provide a summary description for each class, interface, field, and method.

Every class, interface, field, and method should be proceeded by a documentation

comment that contains at least one sentence that acts as a summary description of that
entity.

54. Fully describe the signature of each method.

The documentation for each method should always include a description for each
parameter, each checked exception, any relevant unchecked exceptions, and any return
value. See Rule #46 for information about Javadoc tag ordering and use.

55. Include examples.

One of the easiest ways to explain and understand how to use software is by giving
specific examples.

Try to include a simple example in each nontrivial class and method description. Use the
HTML <pre></pre> tags to maintain the formatting of each example:

/**
* ...
* If you are formatting multiple numbers, it is
* more efficient to get the format just once so
* the system does not have to fetch the
* information about the local language and
* country conventions multiple times:
* <pre>
* DateFormat df = DateFormat.getDateInstance();
* for (int i = 0; i < a.length; ++i) {
* output.println(df.format(myDate[i]) + "; ");
* }
* </pre>

43
 * To format a number for a different Locale,
* specify the locale in the call to
* <code>getDateInstance</code>:
* <pre>
* DateFormat df;
* df = DateFormat.getDateInstance(Locale.US);
* </pre>
* ...
*/
public abstract class DateFormat extends Format {
...
}

56. Document preconditions, postconditions, and invariant conditions.

A precondition is a condition that must hold true before a method starts if this method is
to behave properly. A typical precondition may limit the range of acceptable values for a
method argument.

A postcondition is a condition that must hold true following the completion of a method
if this method has behaved properly. A typical postcondition describes the state of an
object that should result from an invocation of the method given an initial state and the
invocation parameters.

An invariant is a condition that must always hold true for an object. A typical invariant
might restrict the integer field vacationDays to a value between 0 and 25.

As preconditions, postconditions, and invariants are the assumptions under which you use
and interact with a class, documenting them is important, especially if these conditions
are too costly to verify using run-time assertions.

57. Document known defects and deficiencies.

Identify and describe any outstanding problems associated with a class or method.
Indicate any replacements or workarounds that exist. If possible, indicate when the
problem might be resolved.

While no one likes to publicize problems in his or her code, your colleagues and
customers will appreciate the information. This information will give them the chance to
implement a workaround or to isolate the problem to minimize the impact of future
changes.

58. Document synchronization semantics.

The presence of the synchronized modifier in the signature of a method normally reveals
whether that method serializes calling threads to protect the state of an object. A user can
determine whether a method is synchronized by looking at the documentation generated

44
by Javadoc because Javadoc will include the modifier as part of the signature of each
method that is declared as synchronized.

Java also provides a second synchronization mechanism that applies to a block of code
instead of an entire method. Methods that use this second mechanism may, in fact, be
thread-safe, but the signature of these methods will not indicate this. In this situation, you
must indicate these are internally synchronized methods within the documentation for
each such method.

Internal Comments

59. Add internal comments only if they will aid others in understanding your
code.

Avoid the temptation to insert comments that provide useless or irrelevant information:

public int occurrencesOf(Object item) {

// This turned out to be much simpler
// than I expected. Let's Go Mets!!
return (find(item) != null) ? 1 : 0;
}

Add comments only when they provide information that will help others understand how
the code works:

public int occurrencesOf(Object item) {

// This works because no duplicates are allowed:
return (find(item) != null) ? 1 : 0;
}

If an internal comment does not add any value, it is best to let the code speak for itself.

60. Describe why the code is doing what it does, not what the code is doing.

Good code is self-documenting. Another developer should be able to look at well-written

code and determine what it does.

For example, a quick examination of the following code reveals that the program appears
to give a 5 percent discount when an invoice totals over a thousand dollars:

if (this.invoiceTotal > 1000.0) {

this.invoiceTotal = this.invoiceTotal * 0.95;
}

The following comment provides little additional information:

// Apply a 5% discount to all invoices

// over a thousand dollars:

45
if (this.invoiceTotal > 1000.0) {
this.invoiceTotal = this.invoiceTotal * 0.95;
}

After reading this code, a reasonable developer may still want to know:

Why is the discount 5 percent?

Who determined the discount and dollar amount?
When or why would these amounts change?

Identify and explain any domain-specific knowledge that is required to understand code:

// This term corrects for the effects of Jupiter,

// Venus, and the flattening of the earth:
sigma += (c1 * Angle.sin(a1)
+ c2 * Angle.sin(Angle.minus(L1,F))
+ c3 * Angle.sin(a2) );

61. Avoid the use of end-line comments.

End-line comments, one-line comments appended to a line of working code, should be

used with care. They can easily interfere with the visual structure of code. Modifications
to a commented line of code may push the comment far enough to the right that line
wrapping or horizontal scrolling must be employed before the comment can be seen
within an editor. Some programmers try to improve the appearance of end-line comments
by aligning them so they are left justified, only to find themselves constantly realigning
the comments each time the code is modified. This is a waste of time.

Place one-line comments on a separate line immediately preceding the code to which
they refer. The only exception to this rule involves local variable declarations whose
descriptions are short enough that an end-line comment can describe them without
producing an unacceptably long line of code.

62. Explain local variable declarations with an end-line comment.

If the description of a local variable is quite short, consider placing the description in a
comment on the same line as the declaration:

int cur = 0; // Index of current pattern element

int prev = 0; // Index of previous pattern element

Do not worry about positioning the comment so it appears aligned with other end-line
comments.

63. Establish and use a set of keywords to flag unresolved issues.

46
Establish a set of keywords for use in creating special comments that you and other
developers can use to signal unresolved issues, which must eventually be dealt with
before the code is considered complete. These comments should include a date and the
initials of the person who raised the issue. The keywords should be chosen to minimize
the chances that the same text may appear elsewhere within the code. In the following
example, a pair of colons decorates the word UNRESOLVED to increase the
likelihood that it is unique:

// :UNRESOLVED: EBW, 11 July 1999

// This still does not handle the case where
// the input overflows the internal buffer!!
while (everMoreInput) {
...
}

64. Label closing braces in highly nested control structures.

While you should generally avoid creating deeply nested control structures, you can
improve the readability of such code by adding end-line comments to the closing braces
of each structure:

for (i...) {
for (j...) {
while (...) {
if (...) {
switch (...) {
...
} // end switch
} // end if
} // end while
} // end for j
} // end for i

65. Add a fall-through comment between two case labels, if no break

statement separates those labels.

When the code following a switch statements case label does not include a break but,
instead, falls through into the code associated with the next label, add a comment to
indicate this was your intent. Other developers will either incorrectly assume a break
occurs, or wonder whether you simply forgot to code one:

switch (command) {
case FAST_FORWARD:
isFastForward = true;
// Fall through!
case PLAY:
case FORWARD:
isForward = true;
break;

47
 case FAST_REWIND:
isFastRewind = true;
// Fall through!
case REWIND:
isRewind = true;
break;
...
}

Note that two adjacent labels do not require an intervening comment.

66. Label empty statements.

When a control structure, such as while or for loop, has an empty statement by design,
add a comment to indicate this was your intent.
Amendments:
// Strip leading spaces
while ((c = reader.read()) == SPACE); * Always use opening
// Empty! and closing brackets
on an empty loop.
________________

13 Brian Kernighan, and P. J. Plauger. The Elements of Programming Style. (New York:
McGrawHill, 1988), pg. 118.

14 Sun Microsystems Javadoc home page

http://java.sun.com/products/jdk/javadoc/index.html.

15 Bertrand Meyer, Object-Oriented Software Construction, 2nd Edition. (Englewood

Cliffs, New Jersey: Prentice Hall, Inc, 1997).

16 Brian Kernighan and P. J. Plauger. The Elements of Programming Style. (New York:
McGrawHill, 1988), pg. 119.

17 Sun Microsystems Javadoc home page:

http://java.sun.com/products/jdk/javadoc/index.html.

18 The javadoc utility acquired the capability to generate documentation for protected
and private members in version 1.1.2.

19 Sun Microsystems. How to Write Doc Comments for Javadoc, ed. Doug Kramer, (Sun
Microsystems Inc., 1999). Accessed online at
http://java.sun.com/products/jdk/javadoc/writingdoccomments.html, Aug 1999.

20 Using a fixed, instead of a proportional-width typeface and using a different color.

21 Suns Javadoc documentation describing @link tag. Accessed online at

http://java.sun.com/products/jdk/1.3/docs/tooldocs/solaris/javadoc.html, Aug 1999.

48
22 Sun Microsystems. How to Write Doc Comments for Javadoc.

23 Note: Sun Microsystems does not currently appear to follow this convention. The
summary descriptions for most classes and interfaces in the Java API still include both a
subject and a verb as of version 1.2.2.

49
5.
Programming Conventions
67. Consider declaring classes representing fundamental data types as final.

Simple classes representing fundamental data types, such as a ComplexNumber class in

an engineering package, find widespread use within their target domain. As such,
efficiency can become an issue of some importance. Declaring a class as final allows its
methods to be invoked more efficiently.

Of course, declaring your class as final will prohibit its use as a superclass. Nevertheless,
there is seldom any reason to extend a class that implements a fundamental data type. In
most such cases, object composition is a more appropriate mechanism for reuse.

68. Build concrete types from native types and other concrete types.

Each nonnative, nonconcrete type that appears in the interface of a concrete type
introduces a new, potentially volatile dependency, effectively exposing every client of
that concrete type to volatility in these other types.

Minimize the number of dependencies that a concrete class has on nonnative,

nonconcrete types. A concrete type that is defined purely in terms of native types
provides better isolation and stability than a concrete type built from other concrete types.
This is especially important for classes that implement fundamental data types, as
dependencies on these low-level classes tend to proliferate throughout an application.

Consider the public interface of the java.util.BitSet class:

public final class BitSet ... {

public BitSet() {...}
public BitSet(int) {...}
public void set(int) {...}
public void clear(int) {...}
public boolean get(int) {...}
public void and(BitSet) {...}
public void or(BitSet) {...}
public void xor(BitSet) {...}
public int hashCode() {...}
public int size() {...}
public boolean equals(Object) {...}
public Object clone() {...}
public String toString() {...}
}

This class uses five data types in its interface: BitSet, which is a self-reference, the
primitive types int and boolean, and the Java native types Object and String. Because the

50
BitSet interface uses stable, native Java types, little possibility exists that changes outside
this class will affect this class or its clients.

69. Define small classes and small methods.

Smaller classes and methods are easier to design, code, test, document, read, understand,
and use. Because smaller classes generally have fewer methods and represent simpler
concepts, their interfaces tend to exhibit better cohesion.

Try to limit the interface of each class to the bare minimum number of methods required
to provide the necessary functionality. Avoid the temptation to add convenience forms
of a method when only one general-purpose form will suffice.

All sorts of informal guidelines exist for establishing the maximum size of a class or
methoduse your best judgment. If

a class or method seems too big, then consider refactoring that class or method into
additional classes or methods.

70. Define subclasses so they may be used anywhere their superclasses may be
used.

A subclass that changes or restricts the behavior of its ancestor class by overriding
something is a specialization of that class, and its instances may have limited
substitutability for the instances of its ancestor class. A specialization may not always be
used anywhere the parent class could be used.

A subclass that is behaviorally compatible with its ancestor class is a subtype and its
instances are fully substitutable for instances of its ancestor class. A subclass that
implements a subtype does not override anything in its ancestor class; it only extends the
services provided by that class. A subtype has the same attributes and associations as its
supertype.

The following design principle addresses the question of substitutability24, 25:

The Liskov Substitution Principle

Methods that use references to base classes must be able to use objects of derived classes
without knowing it.26

According to this principle, the ability to substitute a derived class object for a superclass
object is characteristic of good design. Such designs offer more stability and reliability
when compared with designs that fail to uphold this principle. When a design adheres to

51
this principle, it generally indicates the designer did a good job identifying the base
abstractions and generalizing their interfaces.

Any design that requires code changes to handle the introduction of a newly derived class
is a bad design. Whenever a derived class violates the existing contract between its super-
classes and their clients, it forces changes in the existing code. When a method accepts a
superclass instance, yet uses the derived type of this instance to control its behavior,
changes will be required for the introduction of each new derived class. Changes of this
kind violate the OpenClosed Principle and are something to avoid.

The OpenClosed Principle

Software entities (Classes, Modules, Functions, and so forth) should be open for
extension, but closed for modification.27

Consider the following example:

class Shape {
...
public Shape getNext() { return this.next; }
public int getDepth() { return this.depth; }
...
}

class Circle extends Shape {...}

class Rectangle extends Shape {...}

class Canvas {

public void drawShapes(ShapeList list) {

Shape shape = list.getNextShape();
// Use null to detect end of list
while (shape != null) {
drawShape(shape);
shape = list.getNextShape();
}
}

public void drawShape(Shape shape) {

// Use derived type to call relevant method
if (shape instanceof Circle)
drawCircle((Circle) shape);
else if (shape instanceof Rectangle)

drawRectangle((Rectangle) shape);
}

52
 public void drawCircle(Circle circle) {...}

public void drawRectangle(Rectangle circle) {

...
}

class ShapeList {
...
protected Shape first;
protected int entries;

public int getEntries() { this.entries; }

public Shape getNextShape() {

Shape temp = this.first;
if (null != this.first) {
this.first = temp.getNext();
this.entries--;
}
// Returns null when empty
return temp;
}
...
}

These classes make up a simple drawing package where Shapes are stored in ShapeLists
and passed to a Canvas for rendering. The drawShapes method of a Canvas object reads
one Shape object at a time from a list and dispatches each to the appropriate draw method.
The process continues until the getNextShape method returns a null value, indicating the
list has been fully traversed.

Now, suppose our design calls for a DepthFilteredShapeList class that can filter out any
Shapes whose depth value falls outside a specified target range:

class DepthFilteredShapeList
extends ShapeList {

protected int min;

protected int max;

public Shape getNextShape(){

Shape temp = this.first;

if (null != this.first)
this.first = temp.next;
this.entries--;
int depth = temp.getDepth();

// Is the shape in range?

if (this.min > depth || depth > this.max) {

53
 // No - return null instead
temp = null;
}
}
// Returns null when filtered!
return temp;
}
}

This implementation has a problem, however. The getNextShape method not only returns
a null value once the end of the list is reached, but also returns a null value for each
Shape that is filtered. The Canvas class does not expect this behavior from a ShapeList
and will incorrectly stop rendering when it tries to read a Shape subject to filtering. This
example violates the Liskov Substitution Principle.

In this case, we can satisfy the principle by changing the getNextShape method so it will
continue to traverse the shape list until it finds an unfiltered shape or until it reaches the
end of the list.

The Liskov Substitution Principle also applies to methods. A method designed to

recognize particular derivations of a superclass may not know how to handle a new
derivation. The drawShape method in the Canvas class illustrates this problem. This
method interrogates each incoming shape

to determine its type in to dispatch it to the appropriate drawing routine. Developers

would have to change the Canvas class and the drawShape method each time they wanted
to add a new subclass of the Shape class.

This problem is solved by adding a drawSelf method to the Shape subclasses and
replacing the shape-specific methods on the canvas with a set of primitive drawing
operations that Shapes can use to draw themselves. Each subclass of Shape would
override the drawSelf method to call the canvas drawing operations necessary to produce
that particular shape. The drawShapes method on the canvas would no longer call
drawShape to dispatch shapes to canvas routines, but would, instead, call the drawSelf
method on each Shape subclass:

class Shape {
...
public abstract void drawSelf(Canvas canvas);
...
}

class Circle extends Shape {

...
public void drawSelf(Canvas canvas) {...}
...
}

class Canvas {

54
 ...
public void drawShapes(ShapeList list) {
Shape shape = list.getNextShape();

// Use null to detect end of list

while (shape != null) {
// Tell the shape to draw itself
shape.drawSelf(this);
shape = list.getNextShape();
}
}

// Define the operations the shapes will use

public void drawLine(int x1,
int y1,

int x2,
int y2) {...}

public void drawCircle(int x,

int y,
int radius) {...}
...
}

71. Make all fields private.

Out of sight, out of mind.Anonymous

Doing so ensures the consistency of the member data because only the owning class may
make changes to it. Access all member data through object methods. This minimizes
coupling between objects, which enhances program maintainability.

72. Use polymorphism instead of instanceof.

Do not use instanceof to choose behavior depending upon an objects type. This forces
you to modify the choice selection code every time the set of choice object types changes,
leading to brittle code.

Instead, implement object-specific behavior in methods derived from a base class. This
enables a client to interact with the base class abstraction without knowledge of the
derived classes, allowing new classes to be introduced without the clients knowledge.

Type Safety

73. Wrap general-purpose classes that operate on java.lang.Object to provide

static type checking.

55
Ensure type safety by wrapping a general class that barters in Object types with one that
casts objects to a specific type. The following code shows how to wrap a general-purpose
queue to create a type-specific one:

public class Queue {

public void enqueue(Object object) {...};
public Object dequeue() {...};
}
public class OrderQueue {
private Queue queue;

public OrderQueue()
this.queue = new Queue();
}

public void enqueue(Order order) {

this.queue.enqueue(order);
}

public Order dequeue() {

return (Order)this.queue.dequeue();
}
}

74. Encapsulate enumerations as classes.

Encapsulate enumerations as classes to provide type-safe comparisons of enumerator

values:

public class Color {

private static int count = 0;
public static final Color RED =
new Color(count++);
public static final Color GREEN =
new Color(count++);
public static final Color BLUE =
new Color(count++);
private int value;

private Color(int value)

this.value = value;
}

public boolean equals(Color other) {

return this.value == other.value;
}

public static int Color.count() {

return count;
}
}

56
Color aColor = Color.RED;

if (anotherColor.equals(aColor)) {
...
}

Statements and Expressions

75. Replace repeated nontrivial expressions with equivalent methods.

Write code once and only once. Factor out common functionality and repackage as a
method or a class. This makes code easier to learn and understand. Changes are localized,
so maintenance is easier and testing effort is reduced.

76. Use block statements instead of expression statements in control flow

constructs.

The Java block statement provides a mechanism for treating any number of statements as
a single compound statement. A block statement may be used anywhere a regular
statement may be used, including the expression statement bodies of a Java control
construct.

While the language enables you to use simple, nonblock statements as the body of these
constructs, you should always use a block statement in these situations.

Block statements reduce the ambiguity that often arises when control constructs are
nested, and they provide a mechanism for organizing the code for improved readability.

The following code fragment is confusing because the indentation makes it appear as
though the else clause is associated with the first if statement, while the compiler will
associate it with the second if statement. The Java language specification refers to this as
the dangling else problem. The use of block statements eliminates this problem:

if (x >= 0)
if (x > 0) positiveX();
else // Oops! Actually matches most recent if!
negativeX();

if (x >= 0) {
if (x > 0) positiveX();
}
else {
negativeX(); // This is what we really wanted!
}

57
In the following example, the code on the top is more difficult to modify than the code on
the bottom. This is because you cannot add another statement without changing the
existing code structure. Because the code on the bottom already uses block statements,
modifications are easier to make:

for (int i = n; i >= 0; i--)

for (int j = n; j >= 0; j--)
f(i,j);
// g(i,j) Cannot add here!

for (int i = n;i >= 0;i--) {

for (int j = n;j >= 0;j--) {
f(i,j);
g(i,j); // Can add here!
}
}

If a control statement has a single, trivial statement as its body, you may put the entire
statement on a single line, but only if it improves readability. Treat this case as the
exception rather than the norm.

77. Clarify the order of operations with parentheses.

The order of operations in a mathematical expression is not always obvious. Even if you
are certain as to the order, you can safely assume others will not be so sure.

// Extraneous but useful parentheses.

int width = (( buffer * offset ) / pixelWidth )
+ gap;

78. Always code a break statement in the last case of a switch statement.

The following switch statement was coded with the assumption that no other cases would
follow the Y case, so no break statement was required:

switch (...) {
case X:
...
break;
case Y:
...
}

What if a new case is needed, however, and the person adding this case decides simply to
add it after the last existing case, but fails to notice this case did not have a break
statement? This person may inadvertently introduce a hard-to-detect fall-through error,
as shown here:

58
switch (...) {
case X:
...
break;
case Y:
... // Oops! Unintended fall-through!

case Z:
...
}

To prevent future problems, you should always code a break statement for the last case in
the switch statement, even if it is the default case:

switch (...) {
case X:
...
break;
case Y:
...
break; // OK! No more fall-through!
case Z:
...
break;
default:
... // Complain about value!
break;
}

Do not forget to add a fall-through comment in those cases that really do fall-
through. See Rule #65.

79. Use equals(), not ==, to test for equality of objects.

Many C++ programmers make this mistake when dealing with Java dates and strings:

Date today = new Date();

while (date != today) {
...
}

String name;
...
if (name == "Bob") {
hiBob();
}

59
In Java, the != and == operators compare object identities, not object values. You
must use the equals method to compare the actual strings:

Date today = new Date();

while (!date.equals(today)) {
...
}

String name;
...
if ("Bob".equals(name)) {
hiBob();
}

Note, unlike the expression name.equals("Bob"), the expression "Bob".equals(name)

does not throw an exception if name is null.

Construction

80. Always construct objects in a valid state.

Never allow an invalid object to be constructed. If an object must be constructed in an

invalid state and then must undergo further initialization before it becomes valid, use a
static method that coordinates the multistage construction. The construction method
should construct the object so that when the method completes, the new object is in a
valid state. Hide any constructors that do not construct valid instances by making them
protected or private.

81. Do not call nonfinal methods from within a constructor.

Subclasses may override nonfinal methods and Java will dispatch a call to such a method
according to the actual type of the constructed objectbefore executing the derived class
constructors. This means when the constructor invokes the derived method, the derived
class may be in an invalid state. To prevent this, call only final methods from the
constructor.

82. Use nested constructors to eliminate redundant code.

To avoid writing redundant constructor code, call lower-level constructors from higher-
level constructors.

This code implements the same low-level initialization in two different places:

60
class Account {
String name;
double balance;
final static double DEFAULT_BALANCE = 0.0d;

Account(String name, double balance) {

this.name = name;
this.balance = balance;
}

Account(String name) {
this.name = name;
this.balance = DEFAULT_BALANCE;
}
}

This code implements the low-level initialization in one place only:

class Account {
String name;
double balance;
final static double DEFAULT_BALANCE = 0.0d;

Account(String name, double balance) {

this.name = name;
this.balance = balance;
}

Account(String name) {
this(name, DEFAULT_BALANCE);
}
}

This approach is also helpful if you are using assertions, as it typically reduces the
number of places a given constructor argument appears, thus reducing the number of
places the validity of that argument is checked.

Exception Handling

83. Use unchecked, run-time exceptions to report serious unexpected errors that
may indicate an error in the programs logic.

Catching and handling run-time exceptions is possible, however, they are usually of such
a severe nature that program termination is imminent. Run-time exceptions are usually
thrown because of programming errors, such as failed assertions, using an out-of-bound
index, dividing by zero, or referencing a null pointer.

84. Use checked exceptions to report errors that may occur, however rarely,
under normal program operation.

61
Checked exceptions indicate a serious problem that should not occur under normal
conditions. The caller must catch this exception. Depending upon the application, a
program may be able to recover from a checked exception; that is, it doesnt indicate a
fundamental flaw in the programs logic.

85. Use return codes to report expected state changes.

For expected state changes, use a return code, sentinel, or special method that returns a
status. This makes code more readable and the flow of control straightforward. For
example, in the course of reading from a file, it is expected the end of the file will be
reached at some point.

86. Only convert exceptions to add information.

Retain all exception information; never discard lower-level explanations:

try {
for(int i = v.size();--i >= 0;) {
ostream.println(v.elementAt(i));
}
}
catch (ArrayOutOfBounds e) {
// should never get here
throw new UnexpectedExceptionError(e);
}

87. Do not silently absorb a run-time or error exception.

Breaking this rule makes code hard to debug because information is lost:

try {
for(int i = v.size();--i >= 0;) {
ostream.println(v.elementAt(i));
}
}
catch (ArrayOutOfBounds e) {
// Oops! We should never get here...
// ... but if we do, nobody will ever know!
}

Even if you have coded a catch block simply to catch an exception you do not expect to
occur, at least print a stack trace. You never know when something impossible might
occur within your software:

try {
for (int i=v.size();--i>=0;) {
ostream.println(v.elementAt(i));
}
}

62
catch (ArrayOutOfBounds e) {
// Oops! Should never get here...
// But print a stack trace just in case...
e.printStackTrace();
}

88. Use a finally block to release resources.

Once a try-block is entered, any corresponding finally block is guaranteed to be executed.

This makes the finally block a good place to release any resources acquired prior to
entering or within the try-block.

In this first example, if an exception or return occurs following the creation of the output
stream, the function will exit without closing and flushing the stream:

pubilc void logSomeStuff() {

OutputStream log = new FileOutputStream("log");
...
// could get exception here!
...
log.close() ;
}

In this example, we use a finally block to ensure the stream is always closed when the
thread of execution exits the try-block. This is done even if the thread exits the block
because an exception has been thrown or a return statement was executed:

OutputStream log = null;

try {
log = new FileOutputStream("log");
}
finally {
if (log != null) {
log.close();
}
}

Assertions

89. Program by contract.

Consider each method a contract between the caller and the callee. The contract states28
the caller must abide by the preconditions of a method and the method, in turn, must
return results that satisfy the postconditions associated with that method.

Abiding by the preconditions of a method usually means passing parameters as the

method expects them; it may also mean calling a set of methods in the correct order. To
abide by the postconditions of the method, the method must correctly complete the work
it was called upon to perform and it must leave the object in a consistent state.

63
Check preconditions and postconditions by assertion (see Rule #92) in any appropriate
public methods. Check preconditions at the beginning of a method, before any other code
is executed, and check postconditions at the end of a method before the method returns.

Derived class methods that override base class methods must preserve the pre- and
postconditions of the base class method. To ensure this, use the template method design
pattern.29

Make each public method final and create a parallel nonfinal protected method that
implements the body of the function. The public final method tests preconditions, calls
the associated protected method, and then tests postconditions. A deriving class may
override public behavior in the base class by overriding the nonfinal protected method
associated with each public final method:

class LinkedList {
public final synchronized void
prepend(Object object) {

// Test pre-condition
if (Assert.ENABLED)
Assert.isTrue(object != null);
doPrepend(object);
// Test post-condition
if (Assert.ENABLED)
Assert.isTrue(first() == object);
}

protected void doPrepend(Object object)

Node node = new Node(object);
if (this.head == null)
this.head = node;
else {
node.next = this.head;
this.head = node;
}
}
}

This technique also works to ensure that a method will remain synchronized, even when
overridden. A derived class may violate the synchronization semantics of a base class by
overriding a synchronized method with an unsynchronized version of the method
derived class methods do not inherit the synchronized qualifier. A superclass can
guarantee synchronization by providing a public final synchronized method that calls the
nonfinal method.

90. Use dead code elimination to implement assertions.

64
An assertion is an expression you, the programmer, insist must hold true for a piece of
code to operate correctly. Assertions are used in code to ensure basic coding assumptions
are not violated. If an assertion evaluates to false, the code is flawed.

Use assertions liberally throughout code to test the basic premises upon which the code
was built. Assertions take time to execute, though, and we usually want to remove them
from the released code. For this, we can take advantage of dead code elimination.

Dead code elimination occurs when a Java compiler eliminates unreachable code. For
example, when a compiler sees the following piece of code, it knows the variable FALSE
will always evaluate to false, as it is a static final variable. This allows the compiler to
eliminate the block of code following the if expression, as the compiler knows it can
never evaluate to true:

class DeadCode {
static final boolean FALSE = false;
public void example() {
if (FALSE) {
System.out.println("Never to be seen.");
}
}
}

Using what we know about dead code elimination, we can write an assertion class that
will enable us to choose when to include assertions in the compiler-generated code:

public class Assert {

public static final boolean ENABLED=true;
public static final void
isTrue(boolean assertion) {
if (Assert.ENABLED && !assertion)

throw new
RuntimeException("Assertion Failed");
}
}
...
if (Assert.ENABLED) Assert.isTrue(a > b);
...

To turn assertions off, set the ENABLED variable in the Assert class to false.

Failed assertions indicate an error in program logic, either in the use of a method or in the
implementation of that method. For this reason, report assertion failures by throwing an
unchecked exception such as RuntimeException or some derivation thereof.

65
91. Use assertions to catch logic errors in your code.

An assertion is a Boolean expression that must hold true for a program to operate
correctly. Use assertions to validate the assumptions made by a program.

92. Use assertions to test pre- and postconditions of a method.

A methods preconditions are those conditions required for the methods proper
execution. For example, a precondition may test the validity of the parameters passed to
the method or test that the object is in a valid state.

Postcondition assertions execute at the completion of a method to verify the object is still
in a valid state and the return values of the method are reasonable.

Concurrency

Concurrency exists when two or more threads make progress, executing instructions at
the same time. A single processor system can support concurrency by switching
execution between two or more threads. A multiprocessor system can support parallel
concurrency by executing a separate thread on each processor. A class is multithread-hot
or MT-hot if it creates additional threads to accomplish its task.

Many applications can benefit from the use of concurrency in their implementation. In a
concurrent model of execution, an application is divided into two or more processes or
threads, each executing in its own sequence of statements or instructions. An application
may consist of one or more processes and a process may consist of one or more threads.
Execution may be distributed on two or more machines in a network, two or more
processors in a single machine, or interleaved on a single processor.

The separately executing processes or threads must generally compete for access to
shared resources and data, and must cooperate to accomplish their overall task.

Concurrent application development is a complicated task. Designing a concurrent

application involves determining the necessary number of processes or threads, their
particular responsibilities, and the methods by which they will interact. It also involves
determining the good, legal, or invariant program states and the bad or illegal program
states. The critical problem is to find and implement a solution that maintains or
guarantees good program states while prohibiting bad program states, even in those
situations where two or more threads may be acting on the same resource.

In a concurrent environment, a programmer maintains desirable program states by

limiting or negotiating access to shared resources using synchronization. The principle
role of synchronization is to prevent undesirable or unanticipated interference between
simultaneously executing instruction sequences.

93. Use threads only where appropriate.

66
Threads are not a silver bullet for improving application performance. An application
not suited for multithreading may run slower following the introduction of multiple
threads because of the overhead required to switch between threads.

Before you introduce threads into your application, determine whether it can benefit from
their use. Use threads if your application needs:30

To react to many events simultaneously.

Examples: An Internet browser or server.
To provide a high level of responsiveness.
Example: A user-interface implementation that can continue to respond to user
actions even while the application is performing other computations.
To take advantage of machines with multiple processors.
Example: An application targeted for particular machine architectures.

Synchronization

Synchronization describes the set of mechanisms or processes for preventing undesirable

interleaving of operations or interference between concurrent threads. A programmer
may choose between one of two synchronization techniques: mutual exclusion or
condition synchronization.

Mutual exclusion involves combining fine-grained atomic actions into coarse-grained

actions and arranging to make these composite actions atomic.

Condition synchronization describes a process or mechanism that delays the execution of

a thread until the program satisfies some predicate or condition.

A thread that is no longer executing because it is delayed or waiting on some

synchronization mechanism is blocked. Once unblocked, awakened, or notified, a thread
becomes runnable and eligible for further execution.

Two basic uses exist for thread synchronization: to protect the integrity of shared data
and to communicate changes efficiently in program state between cooperating threads.

Java supports both mutual exclusion and condition synchronization via a mechanism
provided by the Object class.

94. Avoid synchronization.

Synchronization is expensive. It takes time to acquire and release the synchronization

objects necessary to synchronize a section of code. Moreover, synchronization serializes
access to an object, minimizing concurrency. Think before you synchronize and only
synchronize when its truly necessary.

67
Do not arbitrarily synchronize every public method. Before synchronizing a method,
consider whether it accesses shared and nonsynchronized states. If it does notif the
method only operates on its local variables, parameters, or synchronized objectsthen
synchronization is not required.

Do not synchronize classes that provide fundamental data types or structures. Let the
users of the object determine whether synchronization is necessary. Users may always
synchronize the object externally under the jurisdiction of a separate lock object.

95. Use synchronized wrappers to provide synchronized interfaces.

Use synchronized wrappers to provide synchronized versions of classes. Synchronized

wrappers provide the same interface as the original class, but its methods are
synchronized. A static method of the wrapped class provides access to the synchronized
wrapper. The following example demonstrates a stack, which has a default,
nonsynchronized interface and a synchronized interface provided by a wrapper class:

public class Stack {

public void push(Object o) {...};
public Object pop() {...};
public static Stack createSynchronizedStack() {
return new SynchronizedStack();
}
}

class SynchronizedStack extends Stack {

public synchronized void push(Object o) {

super.push(o);
}

public synchronized Object pop() {

return super.pop();
}
}

96. Do not synchronize an entire method if the method contains significant

operations that do not need synchronization.

To maximize concurrency in a program, we must minimize the frequency and duration of

lock acquisition. A thread entering a synchronized method or block attempts to acquire a
lock. Only one thread at a time may acquire ownership of a

68
lock, so a lock may be used to serialize access to code or a program state. When a thread
has finished executing in the synchronized section of code, it releases the lock so others
threads may attempt to acquire ownership.

A method annotated with the synchronized keyword acquires a lock on the associated
object at the beginning of the method and holds that lock until the end of the method. As
is often the case, however, only a few operations within a method may require
synchronization. In these situations, the method-level synchronization can be much too
coarse.

The alternative to method-level synchronization is to use the synchronized block

statement:

protected void processRequest () {

Request request = getNextRequest();
RequestId id = request.getId();
synchronize(this) {
RequestHandler handler =
this.handlerMap.get(id);
}
handler.handle(request);
}

97. Avoid unnecessary synchronization when reading or writing instance

variables.

The Java language guarantees read-and-write actions are atomic for object references and
all primitives, except for type long and type double. We can, therefore, avoid the use of
synchronization when reading or writing atomic data. Be careful, though. If the value of
an atomic variable is dependent on or related to the other variables, then synchronization
is necessary.

In this example, the assignment of x and y must be synchronized together because they
are interdependent values:

public void synchronized setCenter(int x, int y) {

this.x = x;
this.y = y;
}

The following example does not require synchronization because it uses an atomic
assignment of an object reference:

public void setCenter(Point p) {

this.point = (Point)p.clone();
}

98. Consider using notify() instead of notifyAll().

69
The notify() method of java.lang.Object awakens a single thread waiting on a condition,
while notifyAll() awakens all threads waiting on the condition. If possible, use notify()
instead of notifyAll() because notify() is more efficient.

Use notify() when threads are waiting on a single condition and when only a single
waiting thread may proceed at a time. For example, if the notify() signals that an item has
been written to a queue, only one thread will be able to read the item from the queue. In
this case, waking up more than one thread is wasteful.

Use notifyAll() when threads may wait on more than one condition or if it is possible for
more than one thread to proceed in response to a signal.

99. Use the double-check pattern for synchronized initialization.

Use the double-check pattern31 in situations where synchronization is required during

initialization, but not after it.

In the following code, the instance variable log needs to be initialized, but only if it is
null. To prevent two threads from trying to initialize the field simultaneously, the
function getLog() is declared synchronized:

synchronized Log getLog() {

if (this.log==null) {
this.log = new Log();
}
return this.log;
}

This code also protects against simultaneous initialization, but it uses the double-check
pattern to avoid synchronization except during initialization:

Log getLog() {
if (this.log==null) {
synchronized (this) {
if (this.log==null) {
this.log = new Log();
}
}
}
return this.log;
}

Efficiency

100. Use lazy initialization.

Do not build something until you need it. If an object may not be needed during the
normal course of program execution, then do not build the object until it is required.

70
Use an accessor method to gain access to the object. All users of that object, including
within the same class, must use the accessor to get a reference to the object:

class PersonalFinance {
LoanRateCalculator loanCalculator = null;

LoanRateCalculator getLoanCalculator() {
if (this.loanCalculator == null)

this.loanCalculator =
new LoanRateCalculator();
return this.loanCalculator;
}
}

101. Avoid creating unnecessary objects.

This is especially important if the new objects have short life spans or are constructed,
but never referenced. This not only wastes execution time to create the object, but it also
uses time during garbage collection.

Redundant initialization, as illustrated in the following code, is quite common, and

wasteful:

Color getTextColor() {
Color c = new Color(...);
if (this.state < 2) {
c = new Color(...);
}
return c;
}

Avoid creating an object until you know what you want:

Color getTextColor() {
Color c = null;
if (this.state < 2) {
c = new Color(...);
} else {
c = new Color(...);
}
return c;
}

102. Reinitialize and reuse objects to avoid new object construction.

Cache and reuse frequently created objects that have limited life spans.

Use accessor methods instead of constructors to reinitialize the object.

71
Take care to choose an implementation that does not need to create its own objects to
manage the objects being cached. This would defeat the purpose!

Use a factory implementation to encapsulate mechanisms for caching and reusing objects.
To manage these mechanisms properly, you must return objects obtained from an object
factory back to the same factory. This means the association between an object and its
factory must be maintained somewhere:

In the classa single static factory is associated with the class of the object, and
that factory manages all objects of that class.
In the objectthe object maintains a reference to the factory that manages it.
In the owner of the objectan owner of an object maintains a reference to the
factory from which the object obtained.

103. Leave optimization for last.

First Rule of Optimization:

Dont do it.
Second Rule of Optimization (For experts only):
Dont do it yet.
Michael Jackson,
Michael Jackson Systems Ltd.

Do not waste time optimizing code until you are sure you need to do it.

Remember the 8020 rule3220 percent of the code in a system uses 80 percent of the
resources (on average). If you are going to optimize, make sure it falls within the 20
percent portion.

________________

24 Barbara Liskov, and John Guttag. Abstraction and Specification in Program

Development. (New York: McGraw-Hill, 1986).

25 Robert Martin. Engineering Notebook, C++ Report, Vol. 8, No. 3 (Mar 1996).
Accessed online at http://www.objectmentor.com/publications/lsp.pdf, Aug 1999.

26 Barbara Liskov originally described this concept with the following statement: If for
each object O1 of type S there is an object O2 of type T such that for all programs P
defined in terms of T, the behavior of P is unchanged when O1 is substituted for O2 then
S is a subtype of T.

27 Robert Martin. Engineering Notebook: The OpenClosed Principle, C++ Report,

Vol. 8, No. 1 (Jan 1996). Accessed online at
http://www.objectmentor.com/publications/ocp.pdf, Aug 1999.

72
28 Bertrand Meyer. Object-Oriented Software Construction.

29 Erich Gamma et al. Design Patterns: Elements of Reusable Object-Oriented Software.

(Reading, Massachusetts: AddisonWesley, 1995). pp. 325330.

30 Doug Lea. Concurrent Programming in Java: Design Principles and Patterns.

(Reading, Massachusetts: AddisonWesley, 1997). pp. 12.

31 Douglas C. Schmidt and Tim Harrison. Pattern Languages of Program Design.

(Reading, Massachusetts: AddisonWesley, 1997).

32 Steve McConnell. Code Complete. (Redmond, Washington: Microsoft Press, 1993).

pp. 681682.

73
6.
Packaging Conventions

THIS SECTION contains guidelines for creating packages. See Rules #1517 for
conventions related to package naming.

104. Place types that are commonly used, changed, and released together, or
mutually dependent on each other, into the same package.

This rule encompasses several related package design principles33:

The Common Reuse Principle

A package consists of classes you reuse together.
If you use one of the classes in the package, you use all of them.

Place classes and interfaces you usually use together into the same package. Such classes
are so closely coupled you cannot use one class without usually using the other. Some
examples of closely related types include

Containers and iterators.

Database tables, rows, and columns.
Calendars, dates, and times.
Points, lines, and polygons.

The Common Closure Principle

A package consists of classes, all closed against the same kind of changes. A change that
affects the package affects all the classes in that package.

Combine classes that are likely to change at the same time, for the same reasons, into a
single package. If two classes are so closely related that changing one of them usually
involves changing the other, then place them in the same package.

The Reuse/Release Equivalence Principle

The unit of reuse is the unit of release. Effective reuse requires tracking of releases from
a change control system. The package is the effective unit of reuse and release.

Treating individual classes as a unit of release is not very practical. A typical application
may consist of tens of hundreds of classes, so releasing code on a class-by-class basis will
dramatically complicate the integration and testing process, and dramatically increase the
overall rate of change within the software.

A package provides a much more convenient mechanism for releasing several classes and
interfaces. Each class or interface within a package may undergo several independent
revisions between releases, but a package release captures only the latest version of each
class and interface. Use packages as the primary unit of release and distribution.

74
The Acyclic Dependencies Principle
The dependency structure between packages must be a directed acyclic graph; there must
be no cycles in the dependency structure.

If two packages directly or indirectly depend on each other, you cannot independently
release one without releasing the other because changes in one package will often force
changes in the other. Such cyclic dependencies dramatically increase the fragility of a
system and can eliminate any reduction in schedule realized by assigning the
development of each package to separate developers or teams.

Take steps to eliminate cyclic dependencies, either by combining the mutually dependent
packages or by introducing a new package of abstractions that both packages can depend
on instead of each other.

105. Isolate volatile classes and interfaces in separate packages.

Avoid placing volatile classes and interfaces in the same package with stable classes and
interfaces. If you use packages as your principle unit of release and distribution, users can
gain access to the latest changes in the volatile classes and interfaces only if you re-
release the entire package. Each time you release the package, your users must absorb the
cost of reintegrating and retesting against all the classes in the package, although many
may not have changed.

Separate volatile classes from stable classes to reduce the code footprint affected by new
releases of code, thereby reducing the impact on users of that code.

106. Avoid making packages that are difficult to change dependent on packages
that are easy to change.

This rule derives from the following design principle34:

The Stable Dependencies Principle

The dependencies between packages should be oriented in the direction of increasing
stability. A package should only depend on packages more stable than it is.

If a package containing difficult-to-change types is dependent on a package that contains

easy, or likely to change, types, then the dependent package effectively acts to impede
change in the volatile package.

In a software system, especially one that is incrementally developed, some packages will
always remain somewhat volatile. The developers of such a system must feel free to
modify and extend these volatile packages to complete the implementation of the system
and must be able to do so without worrying too much about downstream effects.

75
Do not create a package that depends on less stable packages. If necessary, create new
abstractions that can be used to invert the relationship between the stable code and the
unstable code.

107. Maximize abstraction to maximize stability.

This rule derives from the following design principle35:

The Stable Abstractions Principle

The stability exhibited by a package is directly proportional to its level of abstraction.
The more abstract a package is, the more stable it tends to be. The more concrete a
package is, the more unstable it tends to be.

Use stable abstractions to create stable packages. Capture high-level, stable concepts in
abstract classes and interfaces, and provide implementations using concrete classes.
Separate abstract classes and interfaces from the concrete classes to form stable and
unstable packages. This ensures the derived classes in the unstable packages depend on
the abstract base classes and interfaces in the stable packages.

108. Capture high-level design and architecture as stable abstractions organized

into stable packages.

To plan and manage a software development effort successfully, the top-level design
must stabilize quickly and remain that way. No development manager can hope to
accurately plan, estimate, schedule, and allocate resources if the architecture of the
system continues to change.

Once the design of the high-level architecture is complete, use packages to separate the
stable parts of the design from the volatile implementation. Create packages to capture
the high-level abstractions of the design. Place the detailed implementation of those
abstractions into separate packages that depend on the high-level abstract packages.

________________

33 Robert Martin. Engineering Notebook: Granularity, C++ Report, Vol. 8, No. 10

(Nov 1996), pp. 5762. Accessed online at
http://www.objectmentor.com/publications/granularity.pdf, Aug 1999.

34 Robert Martin. Engineering Notebook: Stability, C++ Report, Vol. 9, No. 2 (Feb
1997). Accessed online at http://www.objectmentor.com/publications/stability.pdf, Aug
1999.

35 Ibid.

76
Summary

1. Adhere to the style of the original.

2. Adhere to the Principle of Least Astonishment.
3. Do it right the first time.
4. Document any deviations.
5. Indent nested code.
6. Break up long lines.
7. Include white space.
8. Do not use hard tabs.
9. Use meaningful names.
10. Use familiar names.
11. Question excessively long names.
12. Join the vowel generation.
13. Capitalize only the first letter in acronyms.
14. Do not use names that differ only in case.
15. Use the reversed, lowercase form of your organizations Internet domain name as
the root qualifier for your package names.
16. Use a single, lowercase word as the root name of each package.
17. Use the same name for a new version of a package, but only if that new version is
still binary compatible with the previous version, otherwise, use a new name.
18. Capitalize the first letter of each word that appears in a class or interface name.
19. Use nouns when naming classes.
20. Pluralize the names of classes that group related attributes, static services, or
constants.
21. Use nouns or adjectives when naming interfaces.
22. Use lowercase for the first word and capitalize only the first letter of each
subsequent word that appears in a method name.
23. Use verbs when naming methods.
24. Follow the JavaBeans conventions for naming property accessor methods.
25. Use lowercase for the first word and capitalize only the first letter of each
subsequent word that appears in a variable name.
26. Use nouns to name fields.
27. Pluralize the names of collection references.
28. Establish and use a set of standard names for trivial throwaway variables.
29. Qualify field variables with this to distinguish them from local variables.
30. When a constructor or set method assigns a parameter to a field, give that
parameter the same name as the field.
31. Use uppercase letters for each word and separate each pair of words with an
underscore when naming constants.
32. Write documentation for those who must use your code and those who must
maintain it.
33. Keep comments and code in sync.
34. Use the active voice and omit needless words.
35. Use documentation comments to describe the programming interface.

77
36. Use standard comments to hide code without removing it.
37. Use one-line comments to explain implementation details.
38. Describe the programming interface before you write the code.
39. Document public, protected, package, and private members.
40. Provide a summary description and overview for each package.
41. Provide a summary description and overview for each application or group of
packages.
42. Use a single consistent format and organization for all documentation comments.
43. Wrap keywords, identifiers, and constants with <code></code> tags.
44. Wrap code with <pre></pre> tags.
45. Consider marking the first occurrence of an identifier with a {@link} tag.
46. Establish and use a fixed ordering for Javadoc tags.
47. Write in the third-person narrative form.
48. Write summary descriptions that stand alone.
49. Omit the subject in summary descriptions of actions or services.
50. Omit the subject and the verb in summary descriptions of things.
51. Use this rather than the when referring to instances of the current class.
52. Do not add parentheses to a method or constructor name unless you want to
specify a particular signature.
53. Provide a summary description for each class, interface, field, and method.
54. Fully describe the signature of each method.
55. Include examples.
56. Document preconditions, postconditions, and invariant conditions.
57. Document known defects and deficiencies.
58. Document synchronization semantics.
59. Add internal comments only if they will aid others in understanding your code.
60. Describe why the code is doing what it does, not what the code is doing.
61. Avoid the use of end-line comments.
62. Explain local variable declarations with an end-line comment.
63. Establish and use a set of keywords to flag unresolved issues.
64. Label closing braces in highly nested control structures.
65. Add a fall-through comment between two case labels, if no break statement
separates those labels.
66. Label empty statements.
67. Consider declaring classes representing fundamental data types as final.
68. Build concrete types from native types and other concrete types.
69. Define small classes and small methods.
70. Define subclasses so they may be used anywhere their superclasses may be used.
71. Make all fields private.
72. Use polymorphism instead of instanceof.
73. Wrap general-purpose classes that operate on java.lang.Object to provide static
type checking.
74. Encapsulate enumerations as classes.
75. Replace repeated nontrivial expressions with equivalent methods.
76. Use block statements instead of expression statements in control flow constructs.
77. Clarify the order of operations with parentheses.

78
78. Always code a break statement in the last case of a switch statement.
79. Use equals(), not==, to test for equality of objects.
80. Always construct objects in a valid state.
81. Do not call nonfinal methods from within a constructor.
82. Use nested constructors to eliminate redundant code.
83. Use unchecked, run-time exceptions to report serious unexpected errors that may
indicate an error in the programs logic.
84. Use checked exceptions to report errors that may occur, however rarely, under
normal program operation.
85. Use return codes to report expected state changes.
86. Only convert exceptions to add information.
87. Do not silently absorb a run-time or error exception.
88. Use a finally block to release resources.
89. Program by contract.
90. Use dead code elimination to implement assertions.
91. Use assertions to catch logic errors in your code.
92. Use assertions to test pre- and postconditions of a method.
93. Use threads only where appropriate.
94. Avoid synchronization.
95. Use synchronized wrappers to provide synchronized interfaces.
96. Do not synchronize an entire method if the method contains significant operations
that do not need synchronization
97. Avoid unnecessary synchronization when reading or writing instance variables.
98. Consider using notify() instead of notifyAll().
99. Use the double-check pattern for synchronized initialization.
100. Use lazy initialization.
101. Avoid creating unnecessary objects.
102. Reinitialize and reuse objects to avoid new object construction.
103. Leave optimization for last.
104. Place types that are commonly used, changed, and released together, or
mutually dependent on each other, into the same package.
105. Isolate volatile classes and interfaces in separate packages.
106. Avoid making packages that are difficult to change dependent on packages
that are easy to change.
107. Maximize abstraction to maximize stability.
108. Capture high-level design and architecture as stable abstractions
organized into stable packages.

79
Glossary

abstract class
A class that exists only as a superclass of another class and can never be directly
instantiated. In Java, an abstract class contains or inherits one or more abstract methods
or includes the abstract keyword in its definition.

abstract method
A method that has no implementation.

abstract data type

Defines a type that may have many implementations. Abstract data types include things
like stacks, queues, and trees.

abstract type
Defines the type for a set of objects, where each object must also belong to a set of
objects that conform to a known subtype of the abstract type. An abstract type may have
one or more implementations.

abstraction
The process and result of extracting the common or general characteristics from a set of
similar entities.

accessor
A method that sets or gets the value of an object property or attribute.

active object
An object that possesses its own thread of control.

acyclic dependency
A dependency relationship where one entity has a direct or indirect dependency on a
second entity, but the second entity has no direct or indirect dependency on the first.

aggregation
An association representing a wholepart containment relationship.

architecture
A description of the organization and structure of a software system.

argument
Data item specified as a parameter in a method call.

assertion
A statement about the truth of a logical expression.

80
attribute
A named characteristic or property of a type, class, or object.

behavior
The activities and effects produced by an object in response to an event.

binary compatible
A situation where one version of a software component may be directly and transparently
substituted for another version of that component without recompiling the components
clients.

block statement
The Java language construct that combines one or more statement expressions into a
single compound statement, by enclosing them in curly braces {...}.

Boolean
An enumerated type whose values are true and false.

built-in type
A data type defined as part of the language. The built-in or native types defined by Java
include the primitive types boolean, byte, char, double, float, int, long, short, and void,
and the various classes and interfaces defined in the standard Java API, such as Object,
String, Thread, and so forth.

checked exception
Any exception that is not derived from java.lang.RuntimeException or java.lang.Error, or
that appears in the throws clause of a method. A method that throws, or is a recipient of, a
checked exception must handle the exception internally or otherwise declare the
exception in its own throws clause.

class
A set of objects that share the same attributes and behavior.

class hierarchy
A set of classes associated by inheritance relationships.

client
An entity that requests a service from another entity.

cohesion
The degree to which two or more entities belong together or relate to each other.

component
A physical and discrete software entity that conforms to a set of interfaces.

81
composition
A form of aggregation where an object is composed of other objects.

concrete class
A completely specified class that may be directly instantiated. A concrete class defines a
specific implementation for an abstract class or type.

concrete type
A type that may be directly instantiated. A concrete type may refine or extend an abstract
type.

concurrency
The degree by which two or more activities occur or make progress at the same time.

constraint
A restriction on the value or behavior of an entity.

constructor
A special method that initializes a new instance of a class.

container
An object whose purpose is to contain and manipulate other objects.

contract
A clear description of the responsibilities and constraints that apply between a client and
a type, class, or method.

coupling
The degree to which two or more entities are dependent on each other.

critical section
A block of code that allows only one thread at a time to enter and execute the instructions
within that block. Any threads attempting to enter a critical section while another thread
is already executing within that section are blocked until the original thread exits.

cyclic dependency
A dependency relationship where one entity has a direct or indirect dependency on a
second entity and the second entity also has a direct or indirect dependency on the first.

data type
A primitive or built-in type that represents pure data and has no distinct identity as an
object.

delegation
The act of passing a message, and responsibility, from one object to a second object to
elicit a desired response.

82
dependency
A relationship where the semantic characteristics of one entity rely upon and constrain
the semantic characteristics of another entity.

derivation
The process of defining a new type or class by specializing or extending the behavior and
attributes of an existing type or class.

documentation comment
A comment that begins with a /** and ends with */, and contains a description and
special tags that are parsed by the Javadoc utility to produce documentation.

domain
An area of expertise, knowledge, or activity.

encapsulation
The degree to which an appropriate mechanism is used to hide the internal data, structure,
and implementation of an object or other entity.

enumeration
A type that defines a list of named values that make up the allowable range for values of
that type.

factor
The act of reorganizing one or more types or classes by extracting responsibilities from
existing classes and synthesizing new classes to handle these responsibilities.

field
An instance variable or data member of an object.

fundamental data type

A type that typically requires only one implementation and is commonly used to
construct other, more useful types. Dates, complex numbers, linked-lists, and vectors are
examples of common fundamental data types.

generalization
The process of extracting the common or general characteristics from a set of similar
entities to create a new entity that possesses these common characteristics.

implementation
The concrete realization of a contract defined by a type, abstract class, or interface. The
actual code.

implementation class
A concrete class that provides an implementation for a type, abstract class, or interface.

83
implementation inheritance
The action or mechanism by which a subclass inherits the implementation and interface
from one or more parent classes.

inheritance
The mechanism by which more specialized entities acquire or incorporate the
responsibilities or implementation of more generalized entities.

inner-class
A class defined within the scope of another class.

instance
The result of instantiating a classthe concrete representation of an object.

instantiation
The action or mechanism by which a type or class is reified to create an actual object.
The act of allocating and initializing an object from a class.

interface
The methods exposed by a type, class, or object. Also a set of operations that define an
abstract service.

interface inheritance
The action or mechanism by which a subtype or subinterface inherits the interface from
one or more parent types or interfaces.

invariant
An expression that describes the well-defined, legal states of an object.

keyword
A language construct. The keywords of the Java language include:

abstract finally public

boolean float return
break for short
byte [goto] static
case if super
catch implements switch
char import synchronized
class instanceof this
[const] int throw

84
 continue interface throws
default long transient
do native try
double new void
else package volatile
extends private while
final protected

Bracketed keywords are reserved but not used.

lazy initialization
When an implementation delays the initialization of a data value until the first use or
access of the data value.

local variable
A variable that is automatically allocated and initialized on the call stack. Includes
variables bound as function arguments.

method
The implementation of an operation. An operation defined by an interface or class.

multiple inheritance
Inheritance relationship where a subtype inherits from two or more supertypes. Java
supports multiple inheritance by allowing an interface to extend multiple interfaces.

mutex
A synchronization mechanism used to provide mutually exclusive access to a resource.

native type
A data type defined as part of the language. The native or built-in types defined by Java
include the primitive types boolean, byte, char, double, float, int, long, short, and void,
and the various classes and interfaces defined in the standard Java API, such as Object,
String, and Thread.

object
The result of instantiating a class. An entity with state, behavior, and identity.

operation
A service that may be requested from an object to effect behavior. Alternatively viewed
as a message sent from a client to an object.

package
A mechanism organizing and naming a collection of related classes.

85
package access
The default access-control characteristic applied to interfaces, classes, and class members.
Class members with package access are accessible only to code within the same package
and are heritable by subclasses in the same package. Classes and interfaces with package
access are not visible to code outside the package. They are only accessible and
extendable by classes and interfaces in the same package.

parameter
A variable that is bound to an argument value passed into a method.

polymorphic
A trait or characteristic of an object whereby that object can appear as several different
types at the same time.

polymorphism
The concept or mechanism by which objects of different types inherit the responsibility
for implementing the same operation, but respond differently to the invocation of that
operation.

postcondition
A constraint or assertion that must hold true following the completion of an operation.

precondition
A constraint or assertion that must hold true at the start of an operation.

primitive type
A basic language type that represents a pure value and has no distinct identity as an
object. The primitives provided by Java include boolean, byte, char, double, float, int,
long, and short.

private access
An access-control characteristic applied to class members. Class members declared with
the private access modifier are only accessible to code in the same class and are not
inherited by subclasses.

property
A named characteristic or attribute of a type, class, or object.

protected access
An access-control characteristic applied to class members. Class members declared with
the protected access modifier are accessible to code in the same class and package, and
from code in subclasses, and they are inherited by subclasses.

public access
An access-control characteristic applied to interfaces, classes, and class members. Class
members declared with the public access modifier are accessible anywhere the class is

86
accessible and are inherited by subclasses. Classes and interfaces declared with the public
access modifier are visible, accessible, and heritable outside of a package.

qualifier
A name or value used to locate or identify a particular entity within a set of similar
entities.

realization
A relationship where one entity agrees to abide by or to carry out the contract specified
by another entity.

responsibility
A purpose or obligation assigned to a type.

role
The set of responsibilities associated with an entity that participates in a specific
relationship. A Java interface often defines a role for an object.

service
One or more operations provided by a type, class, or object to accomplish useful work on
behalf of one or more clients.

signature
The name, parameter types, return type, and possible exceptions associated with an
operation.

state
The condition or value of an object between events.

static type checking

Compile-time verification of the assumptions made about the use of object reference and
data value types.

subclass
A class that inherits attributes and methods from another class.

subtype
The more specific type in a specializationgeneralization relationship.

superclass
A class whose attributes and methods are inherited by another class.

supertype
The more general type in a specializationgeneralization relationship.

87
synchronization
The process or mechanism used to preserve the invariant states of a program or object in
the presence of multiple threads.

synchronized
A characteristic of a method or a block of code. A synchronized method or block allows
only one thread at a time to execute within the critical section defined by that method or
block.

thread
A single flow of control flow within a process that executes a sequence of instructions in
an independent execution context.

type
Defines the common responsibilities, behavior, and operations associated with a set of
similar objects. A type does not define an implementation.

unchecked exception
Any exception that is derived from java.lang. RuntimeException or java.lang.Error. A
method that throws, or is a recipient of, an unchecked exception is not required to handle
the exception or declare the exception in its throws clause.

variable
A typed, named container for holding object references or a data values.

visibility
The degree to which an entity may be accessed from outside of a particular scope.

88
Bibliography

Ambler, S. W. The Object Primer: The Application Developers Guide to Object

Orientation, second edition. New York: SIGS Books/Cambridge University Press, 1995.
ISBN 0-521-78519-7.

Ambler, S. W. Building Object Applications That Work: Your Step-By-Step Handbook for
Developing Robust Systems with Object Technology. New York. SIGS Books/Cambridge
University Press, 1998. ISBN 0-521-64826-2.

Ambler, S. W. Process PatternsBuilding Large-Scale Systems Using Object

Technology. New York: SIGS Books/Cambridge University Press, 1998. ISBN 0-521-
64568-9.

Ambler, S. W. More Process PatternsDelivering Large-Scale Systems Using Object

Technology. New York: SIGS Books/Cambridge University Press, 1999. ISBN 0-521-
65262-6.

Arnold, K. and J. Gosling. The Java Programming Language. Reading, Massachusetts:

AddisonWesley, 1996. ISBN 0-201-63455-4.

Chan, P. et al. The Jave Class Libraries, Volume 1: java.io, java.lang, java.math,
java.net, java.text, java.util, 2nd Edition. Reading, Massachusetts: AddisonWesley,
1998. ISBN 0-201-31002-3.

Chan P. et al. The Java Class Libraries, Volume 1: Supplement for the Java 2
Platform, Standard Edition, v1.2, 2nd Edition. Reading, Massachusetts: AddisonWesley,
1999. ISBN 0-201-48552-4.

Chan, P. and R. Lee. The Java Class Libraries, Volume 2: java.applet, java.awt,
java.beans, 2nd Edition. Reading, Massachusetts: AddisonWesley, 1998. ISBN 0-201-
31003-1.

Gamma, E. et al. Design Patterns: Elements of Reusable Object-Oriented Software.

Reading, Massachusetts: Addison-Wesley, 1995. ISBN 0-201-63361-2.

Gosling, J. et al. The Java Language Specification. Reading, Massachusetts: Addison

Wesley, 1996. ISBN 0-201-63451-1.

Kernighan, B. and P. J. Plauger. The Elements of Programming Style, 2nd Edition. New
York: McGrawHill, June 1988. ISBN 0-07-034-207-5.

Lea, D. Concurrent Programming in Java Design Principles and Patterns. Reading,

Massachusetts: Addison-Wesley, 1997. ISBN 0-201-69581-2.

89
Liskov, B. and J. Guttag. Abstraction and Specification in Program Development. New
York: McGraw-Hill, 1986.

Maguire, S. Writing Solid Code. Redmond, Washington: Microsoft Press, 1993. ISBN 1-
55615-4.

Martin, R. Granularity, C++ Report, Vol. 8, No. 10 (Nov 1996), pp. 5762. Accessed
online at http://www.objectmentor.com/publications/granularity.pdf, Aug 1999.

Martin, R. Stability, C++ Report, Vol. 9, No. 2 (Feb 1997). Accessed online at
http://www.objectmentor.com/publications/stability.pdf, Aug 1999.

McConnell, S. Code Complete. Redmond, Washington: Microsoft Press, 1993. ISBN 1-

55615-484-4.

McConnell, S. Software Project Survival Guide. Redmond, Washington: Microsoft Press,

1998. ISBN 1-57231-621-7.

Meyer, B. Object-Oriented Software Construction, 2nd Edition. Englewood Cliffs, New

Jersey: Prentice-Hall, Inc., 1997. ISBN 0-13-629155-4.

Nagler, J. Coding Style and Good Computing Practices, The Political Methodologist,
Spring 1995, Volume 6, No. 2.

Strunk, W., Jr., and E. B. White. The Elements of Style. New York: Macmillan, 1979.
ISBN 0-02-418200-0.

Stroustrup, B. The C++ Programming Language, 3rd Edition. Reading, Massachusetts:

AddisonWesley, 1997. ISBN 0-201-88954-4.

Sun Microsystems, Package Naming Conventions In Clarifications and Amendments to

The Java Language Specification. Palo Alto, California: Sun Microsystems, 26 Aug
1999. Accessed online at http://java.sun.com/docs/books/jls/clarify.html, Aug 1999.

Sun Microsystems, JavaBeans API Specification., ed. G. Hamilton, Mountain View,

California: Sun Microsystems Inc., 1997. Accessed online at
http://www.javasoft.com/beans/docs/beans.101.pdf, Aug 1999.

Sun Microsystems, Java Code Conventions. Palo Alto, California: Sun Microsystems,
20 April 1999. Accessed online at ftp://ftp.javasoft.com/docs/codeconv/
CodeConventions.pdf, Aug 1999.

Sun Microsystems, JDK 1.2 API Documentation. Palo Alto, California: Sun
Microsystems, 1999. Accessed online at
http://java.sun.com/products/jdk/1.2/docs/index.html, Aug 1999.

90
Taligent Press, Taligents Guide to Designing Programs. Reading, Massachusetts:
AddisonWesley, 1994. ISBN 0-201-40888-0

91