A Simplified Introduction to L

A
T
E
X
Harvey J. Greenberg
University of Colorado at Denver
Mathematics Department
PO Box 173364
Denver, CO 80217-3364
Harvey.Greenberg@cudenver.edu
http://www.cudenver.edu/~hgreenbe/
February 19, 2006
i
Table of Contents
List of Figures iii
List of Tables vi
Preface viii
Acknowledgments ix
Sources of L
A
T
E
X Software ix
1 Overview 1
2 Text 5
2.1 Fonts and Paragraphs . . . . . . . . . . . . . . . . . . . . . . 6
2.2 Lists . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.3 Making Tables . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.4 Special Characters . . . . . . . . . . . . . . . . . . . . . . . . 20
2.5 Tabbing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.6 Line, Page, and Word Breaks . . . . . . . . . . . . . . . . . . 23
2.7 Spacing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3 Bibliography with BibT
E
X 28
3.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.2 The bib File . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.2.1 Main body . . . . . . . . . . . . . . . . . . . . . . . . 28
3.2.2 Web citations . . . . . . . . . . . . . . . . . . . . . . . 32
3.2.3 Additional features . . . . . . . . . . . . . . . . . . . . 34
3.3 Declaration and Citation . . . . . . . . . . . . . . . . . . . . . 36
3.4 Some Controls . . . . . . . . . . . . . . . . . . . . . . . . . . 37
Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
4 Counters, Labels, and References 39
4.1 Basic Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . 39
4.2 Intrinsic Counters . . . . . . . . . . . . . . . . . . . . . . . . . 39
4.3 Figures and Tables . . . . . . . . . . . . . . . . . . . . . . . . 40
4.4 Defining Your Own . . . . . . . . . . . . . . . . . . . . . . . . 42
Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
ii
5 Math Mode 45
5.1 Mathematical Symbols . . . . . . . . . . . . . . . . . . . . . . 46
5.2 Fractions and Variable Size Functionality . . . . . . . . . . . 48
5.3 Arrays and Equations . . . . . . . . . . . . . . . . . . . . . . 51
5.4 Special Functions and Alphabets . . . . . . . . . . . . . . . . 58
5.5 Derivatives and Integrals . . . . . . . . . . . . . . . . . . . . . 60
5.6 Theorems and Definitions . . . . . . . . . . . . . . . . . . . . 61
5.7 Refinements . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
5.8 Grammar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
6 Graphics 71
6.1 Picture Environment . . . . . . . . . . . . . . . . . . . . . . . 71
6.2 PSTricks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
6.3 Importing pictures . . . . . . . . . . . . . . . . . . . . . . . . 86
Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
7 Making Special Parts 93
7.1 Cover Page . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
7.2 Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
7.3 Other Front Matter . . . . . . . . . . . . . . . . . . . . . . . . 97
7.4 Back Matter . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
7.5 Footnotes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
8 Taking Control 100
8.1 Your Own Abbreviations and Commands . . . . . . . . . . . 100
8.2 Your Own Names, Titles and Numbers . . . . . . . . . . . . . 102
8.3 Your Own Environments . . . . . . . . . . . . . . . . . . . . . 103
8.4 Your Own Margins and Spacing . . . . . . . . . . . . . . . . . 103
8.5 Your Own Output Control . . . . . . . . . . . . . . . . . . . . 108
8.6 Your Own Bibliography . . . . . . . . . . . . . . . . . . . . . 109
Closing Remarks 110
Appendix 111
Some Tips 117
References 119
Index 121
iii
List of Figures
1 The Structure of a L
A
T
E
X Document. . . . . . . . . . . . . . 1
2 Your First L
A
T
E
X Source File . . . . . . . . . . . . . . . . . . 2
3 Command Sequence from Source to Postscript or PDF . . . 4
4 An Introductory Document Source (Result in Figure 5) . . . 5
5 An Introductory Document Result (Source in Figure 4) . . . 6
6 Positioning Paragraphs Source (Result in Figure 7) . . . . . 6
7 Positioning Paragraphs Result (Source in Figure 6) . . . . . 7
8 Centering Source (Result in Figure 9) . . . . . . . . . . . . . 7
9 Centering Result (Source in Figure 8) . . . . . . . . . . . . . 8
10 Some Font Sizes Source (Result in Figure 11) . . . . . . . . . 9
11 Some Font Sizes Result (Source in Figure 10) . . . . . . . . . 9
12 Skipping Line Spaces Source (Result in Figure 13) . . . . . . 11
13 Skipping Line Spaces Result (Source in Figure 12) . . . . . . 11
14 Description List Environment . . . . . . . . . . . . . . . . . . 13
16 Itemize List Environment Result (Source in Figure 15) . . . 13
15 Itemize List Environment Source (Result in Figure 16) . . . 14
17 Enumerate List Environment Source (Result in Figure 18) . 14
18 Enumerate List Environment Result (Source in Figure 17) . 15
19 A 2 3 Table . . . . . . . . . . . . . . . . . . . . . . . . . . 15
20 A 2 3 Table with Horizontal and Vertical Lines . . . . . . . 16
21 A Table with Partially Spanning Horizontal and Vertical Lines 16
22 Nested Tables Source (Result in Figure 23) . . . . . . . . . . 17
23 Nested Tables Result (Source in Figure 22) . . . . . . . . . . 17
24 \parbox Source (Result in Figure 25) . . . . . . . . . . . . . 19
25 \parbox Result (Source in Figure 24) . . . . . . . . . . . . . 19
26 Multicolumn Source (Result in Figure 27) . . . . . . . . . . . 20
27 Multicolumn Result (Source in Figure 26) . . . . . . . . . . . 20
28 Obtaining Brackets in a Description List Environment . . . . 21
29 Tabbing Source (Result in Figure 30) . . . . . . . . . . . . . 23
30 Tabbing Result (Source in Figure 29) . . . . . . . . . . . . . 23
31 Adding bibtex to the Command Sequence . . . . . . . . . . 28
32 A Document to Print the Bibliographic Database . . . . . . 37
33 Framed Figure 34 Source . . . . . . . . . . . . . . . . . . . . 42
34 Framed Figure with Caption at Bottom . . . . . . . . . . . . 42
35 Framed Figure with Caption at Top . . . . . . . . . . . . . . 42
36 Alternative enumerate Symbols Source (Result in Figure 37) 44
37 Alternative enumerate Symbols Result (Source in Figure 36) 44
38 Variable Sizes Source (Result in Figure 39) . . . . . . . . . . 48
39 Variable Sizes Result (Source in Figure 38) . . . . . . . . . . 48
40 \displaystyle Source (Result in Figure 41) . . . . . . . . . 49
41 \displaystyle Result (Source in Figure 40) . . . . . . . . . 50
iv
42 Examples to Compare Text and Display Modes . . . . . . . . 50
43 eqnarray Environment Source (Result in Figure 44) . . . . . . 53
44 eqnarray Environment Result (Source in Figure 43) . . . . . . 53
45 Matrix Equation Source (Result in Figure 46) . . . . . . . . 54
46 Matrix Equation Result (Source in Figure 45) . . . . . . . . 54
47 Nested Arrays Source (Result in Figure 48) . . . . . . . . . . 55
48 Nested Arrays Result (Source in Figure 47) . . . . . . . . . . 55
49 Horizontal Braces Source (Result in Figure 50) . . . . . . . . 56
50 Horizontal Braces Result (Source in Figure 49) . . . . . . . . 56
51 \flushleft in parbox Source (Result in Figure 52) . . . . . 57
52 \flushleft in parbox Result (Source in Figure 51) . . . . . 57
53 \raggedright in parbox Source (Result in Figure 54) . . . . 58
54 \raggedright in parbox Result (Source in Figure 53) . . . . 58
55 gather* Environment Source (Result in Figure 56) . . . . . . 64
56 gather* Environment Result (Source in Figure 55) . . . . . . 64
57 Commutative Diagram Source (Result in Figure 58) . . . . . 66
58 Commutative Diagram Result (Source in Figure 57) . . . . . 66
59 Vertical Diagram Source (Result in Figure 60) . . . . . . . . 72
60 Vertical Diagram Result (Source in Figure 59) . . . . . . . . 72
61 Variety of Objects in Picture Environment . . . . . . . . . . 72
62 Source for Figure 61 . . . . . . . . . . . . . . . . . . . . . . . 73
63 Line Parameters . . . . . . . . . . . . . . . . . . . . . . . . . 75
64 PSTricks Source for Connecting Nodes . . . . . . . . . . . . . 82
65 Graph Source (Result in Figure 66) . . . . . . . . . . . . . . 82
66 Graph Result (Source in Figure 65) . . . . . . . . . . . . . . 83
67 Source Code for Drawing Histogram of Test Scores . . . . . . 84
68 Sequence of PSTricks Commands to Draw Histogram . . . . 87
69 Applying \includegraphics to Import an eps File . . . . . . 88
70 Specifying Dimensions in \includegraphics . . . . . . . . . 88
71 Title Page Source (Result in Figure 72) . . . . . . . . . . . . 93
72 Title Page Result (Source in Figure 71) . . . . . . . . . . . . 93
73 Adding Addresses to Authors . . . . . . . . . . . . . . . . . . 94
74 Footnotes in the Cover Page Source (Result in Figure 75) . . 95
75 Footnotes in the Cover Page Result (Source in Figure 74) . . 95
76 Authors with same footnote (Result in Figure 77) . . . . . . 95
77 Authors with same footnote (Source in Figure 76) . . . . . . 96
78 Making an Abstract Source (Result in Figure 79) . . . . . . . 96
79 Making an Abstract Result (Source in Figure 78) . . . . . . . 97
80 Some Front Matter Specifications for This Document . . . . 97
81 Adding makeindex to the Command Sequence . . . . . . . . 99
82 Setting a Footnote Source (Result in Figure 83) . . . . . . . 99
83 Setting a Footnote Result (Source in Figure 82) . . . . . . . 99
84 Setting a Footnote Source (Result in Figure 85) . . . . . . . 100
v
85 Setting a Footnote Result (Source in Figure 84) . . . . . . . 100
86 Document Margins . . . . . . . . . . . . . . . . . . . . . . . . 105
87 Varying \itemsep to control item spacing in a list . . . . . . 106
88 Changing the Left Margin of a List (Result in Figure 89) . . 107
89 Changing the Left Margin of a List (Source in Figure 88) . . 107
90 Array with Fixed Width Column Source (Result in Figure 91) 107
91 Array with Fixed Width Column Result (Source in Figure 90) 107
92 \ifthenelse Source (Results in Figure 93) . . . . . . . . . . 108
93 \ifthenelse Results (Source in Figure 92) . . . . . . . . . . 108
94 Most of the Preamble for this Book . . . . . . . . . . . . . . 111
vi
List of Tables
1 Intrinsic Font Styles . . . . . . . . . . . . . . . . . . . . . . . 8
2 Writing Special Characters . . . . . . . . . . . . . . . . . . . 21
3 Some Accents for Letters . . . . . . . . . . . . . . . . . . . . 22
4 The Tabbing Environment . . . . . . . . . . . . . . . . . . . 22
5 The \kill Tabbing Command . . . . . . . . . . . . . . . . . 23
6 Figure and Table Location Options . . . . . . . . . . . . . . 41
7 Numerals to Print Counters . . . . . . . . . . . . . . . . . . . 43
8 Default Settings for enumerate Counters . . . . . . . . . . . . 44
9 Some Mathematical Operations . . . . . . . . . . . . . . . . 46
10 Set Notation . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
11 The `mathfont Commands . . . . . . . . . . . . . . . . . . . 47
12 Variable Size Mathematical Operation Symbols . . . . . . . . 49
13 Some Symbols in Logic . . . . . . . . . . . . . . . . . . . . . 51
14 Order Relations . . . . . . . . . . . . . . . . . . . . . . . . . 51
15 Transpose of a Vector . . . . . . . . . . . . . . . . . . . . . . 54
16 Some Common Mathematical Functions . . . . . . . . . . . . 59
17 Examples of Mathematical Functions . . . . . . . . . . . . . 59
18 Notation Using mathbb Fonts from amssymb Package . . . . 59
19 Some Basic Drawing Commands in PSTricks . . . . . . . . . 79
20 Boxes in PSTricks . . . . . . . . . . . . . . . . . . . . . . . . 80
21 Parameters for \psaxes . . . . . . . . . . . . . . . . . . . . . 86
22 Intrinsic Name Parameters . . . . . . . . . . . . . . . . . . . 102
23 Margin Parameters . . . . . . . . . . . . . . . . . . . . . . . 104
24 Spacing Parameters . . . . . . . . . . . . . . . . . . . . . . . 106
25 Conversions of Common Units of Measurement . . . . . . . . 111
26 Reference Tables . . . . . . . . . . . . . . . . . . . . . . . . . 112
27 Commands/Environments for Text Font Appearance . . . . . 112
28 Commands/Environments for Controlling Text Position . . . 112
29 Text Accents and Special Symbols . . . . . . . . . . . . . . . 113
30 Commands for Counters . . . . . . . . . . . . . . . . . . . . . 113
31 Commands/Environments to Organize Document . . . . . . 113
32 Commands to Control Document Style . . . . . . . . . . . . 113
33 Commands to Control Fonts in Math Mode . . . . . . . . . . 113
34 Accents in Math Mode . . . . . . . . . . . . . . . . . . . . . 114
35 Spacing Commands in Math Mode . . . . . . . . . . . . . . . 114
36 Greek and Special Letters . . . . . . . . . . . . . . . . . . . . 114
37 Frequently Used Mathematical Symbols . . . . . . . . . . . . 115
38 Binary Operations . . . . . . . . . . . . . . . . . . . . . . . . 115
39 Operators and Quantifiers . . . . . . . . . . . . . . . . . . . . 115
40 Special Functions . . . . . . . . . . . . . . . . . . . . . . . . 115
41 Relation Symbols . . . . . . . . . . . . . . . . . . . . . . . . 115
vii
42 Arrows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
43 Dots, Circles, Triangles and Lines . . . . . . . . . . . . . . . 116
44 Variable Size Symbols . . . . . . . . . . . . . . . . . . . . . . 117
45 Special Symbols in Both Text and Math Modes . . . . . . . 117
46 Commands and Parameters in Picture Environment . . . . . 117
viii
Preface
The majority of this book is about using L
A
T
E
X2
ε
[2, 10], a descendant of
L
A
T
E
X, designed by Leslie Lamport [9], based on T
E
X, originated by Donald
E. Knuth [8]. This is a typesetting program, not a word processor. You
enter some editor that saves plain text files. Then, you type text freely until
you need something special, such as italic font or a complex mathematical
expression, like
lim
ε→0
+

a
i

a
i

1 + (r −j)
2
dr
Φ(ε)
.
It was the desire to have high quality, low cost publications in mathe-
matics and related disciplines that caused Knuth (pronounced Kah-nooth) to
invent T
E
X (pronounced Tek) in the late 1970’s. Originally believing that he
could write a program in less than a year that could typeset documents, he
actually ended up defining an entire branch of research in computer science.
It was 10 years later that he published his seminal book [8], but he published
articles along the way, and he permanently changed the way mathematical
documents are prepared. L
A
T
E
X (pronounced Lah-tek or Lay-tek) is a col-
lection of macros built on top of T
E
X that “represents a balance between
functionality and ease of use” [9, p. xiii]. L
A
T
E
X2
ε
is the current version,
developed by a team of volunteers: Johannes L. Braams, David P. Carlisle,
Alan Jeffrey, Frank Mittelbach, Chris Rowley, and Rainer Schöpf [2].
A comprehensive coverage of L
A
T
E
X and the many enhancements to it is
given by the The L
A
T
E
X Companion [5]. By contrast, this book is designed to
be a succinct introduction, omitting many of the things L
A
T
E
X2
ε
can do. My
goal is to offer enough of an introduction that someone not acquainted with
L
A
T
E
X (or with T
E
X) can write a term paper, thesis, or article, using L
A
T
E
X2
ε
to produce high quality results. Exercises are provided for guided instruction,
which should be just a few classes. For one who is well acquainted with
computers, particularly unix, the basics that are covered should take less
than 10 hours, and one could do all of the exercises. For one who is just
learning how to use a computer, it will take longer, especially getting used
to functioning at the command line. In any case, the finer points require
more study.
Happy L
A
T
E
Xing.
— Harvey J. Greenberg
Denver, Colorado
ix
Acknowledgments
The author thanks the many contributors in the comp.text.tex newsgroup,
particularly Donald Arseneau, Herman Bruyninckx, David Carlisle, Robin
Fairbairns, Jonathan Fine, Denis Girou, David Haller, Dan Luecking, Timo-
thy Murphy, Sebastian Rahtz, Axel Reichert, Thomas Ruedas, Bernd Schandl,
Anton Schwaighofer, Mårten Svantesson, and Matt Swift, who were very gen-
erous with taking time to answer so many questions on a regular basis. I also
received useful comments from people who read an earlier draft that I made
available on the web, notably William Briggs. I especially thank Kasper B.
Graversen whose in-depth review has made this version much better than my
original. One student, Andrea Dean, provided feedback that led to several
points of clarification. Last, but not least, I thank Allen G. Holder, who
taught me L
A
T
E
X in the first place.
Sources of L
A
T
E
X Software
The basic L
A
T
E
X software system is available free of charge for unix systems,
and MiKTeX [13] is available free of charge for DOS systems. The best
source of these, and additional packages that extend the L
A
T
E
X capabilities
(to which I refer in this book), is at the Comprehensive TeX Archive Network
(CTAN) [4], at three host sites (and many mirrors):
1. http://ctan.tug.org/ in Boston, MA, USA,
2. http://www.tex.ac.uk/ in Cambridge, UK, and
3. http://www.dante.de/ in Mainz, Germany (in German).
These all describe how to search and browse the FTP sites for software and
documents.
1
1 Overview
You will create a file, called the L
A
T
E
X source, which is plain text. To keep
things simple, its suffix is .tex. For example, I refer to myfile.tex as a
plain text source file that you create. Figure 1 shows the structure of this
file, which I shall describe in greater detail throughout this book.
% This is myfile.tex
% notes to yourself can go here

Anything following % is ignored
(used for comments).
\documentclass[ojtion:]{:tn|c}
optional specifications
— e.g., declaring use of packages

Preamble
(blank lines do not matter)
\begin{document}
.
.
.
\end{document}
¸
Body
This is the document environment.
All that follows is ignored
(could be used for comments).
Figure 1: The Structure of a L
A
T
E
X Document.
In the preamble, there are many options, depending upon the style; the
intrinsic document styles are: article, book, letter, report, and slides.
Most publishers have their own style, which they provide for you to down-
load. Among these are professional societies, notably the American Mathe-
matical Society (amsmath style) and the Society for Industrial and Applied
Mathematics (siam style).
The focus throughout this book is on the article style. Further, I shall
be using defaults for almost everything, concentrating on getting started
with using L
A
T
E
X as quickly as possible. Later, some of the options, like
margin settings and other preamble specifications are covered, as well as
more advanced topics for customizing your document.
Are you ready to write your first L
A
T
E
X document? Copy the source file
shown in Figure 2 and name it myfile.tex. Then, at the command line,
enter:
latex myfile
(In an MS Windows system, the command line is the DOS command line,
which you enter by running Start−→Programs−→MS-DOS Prompt, or by
Start−→Run and entering cmd or command into the window.)
This is called compiling your source, which creates several output files.
The only one you need to be concerned with now is the dvi file, which the
2 1 OVERVIEW
\documentclass{article}
\begin{document}
Hello world.
\end{document}
Figure 2: Your First L
A
T
E
X Source File
latex program (called a “compiler”) names myfile.dvi. One of three things
will have occurred:
Case 1. You got messages, but they were not fatal errors.
Among the non-fatal messages you will generally see are warnings like:
Overfull \hbox ...
Overfull \vbox ...
Underfull \hbox ...
Underfull \vbox ...
Do not worry about these.
Case 2. You got a fatal error message.
You must find and correct it. This is called debugging your source.
Sometimes the error message tells you what went wrong, such as miss-
ing a brace (characters { and }, which you will come to know and
love), or some command was not recognized due to being misspelled.
Many times the message is not very informative, so you are advised to
compile often. That way you will know that what you did in the last
few minutes contains the error.
Case 3. You got no messages.
Something went wrong and you need to ask for help.
If all went well, the first thing to do is save a backup by copying your
source file to another subdirectory, or to a different name. In unix do this
by entering:
cp -p myfile.tex myfile-1.tex
(The -p is to keep the date and time of the source file.) Change -1 to another
qualifier each time (e.g., -2 . . . ), so you have a collection of backups. If
you are running under DOS, use copy myfile.tex destination, where the
destination is either a: or some backup file name. (If you are familiar with
3
DOS, nothing more need be said; if not, you need to learn how to create,
edit, and save plain text files.)
Next you want to view the result. If you are in a unix environment, you
can view the result with the dvi viewer, xdvi. At the command line enter:
xdvi myfile
and it will come on your screen. (There is more to do if you are working
remotely, in which case you might ask someone for help.)
If you are using DOS, the viewer that comes with MiKTeX [13], a free
software system by Christian Schenk, is called YAP. At the DOS command
line you enter:
YAP myfile
You will see various options for viewing and printing.
Under unix, xdvi does not have a print option, so you first need to
convert the dvi file to postscript. This is done with the program, dvips. At
the command line enter:
dvips myfile -o
(The -o tells the system you want the output to go to a file, rather than
just print; your installation might already have file output as the default,
in which case the -o is not needed.) This will result in the creation of the
postscript file myfile.ps. You can print it by any number of ways, including
the unix command, lpr myfile.ps.
The same conversion program can be run under DOS (and comes with
MiKTeX), and you might want to obtain myfile.ps for a variety of reasons,
including posting it on the web. To view or print a postscript file under
DOS, you can run a program called ghostview. You will need to find out
more about viewing and printing postscript files that suit your particular
needs.
Summarizing, you begin by entering a plain text editor. In unix this could
be pico, emacs, vi or vim. In DOS you can use EDIT at the command line,
or you can use Notepad, Wordpad or MS Word. If you use a word processor,
however, you must take absolutely no advantage of its formatting. You
should even put in hard return characters (i.e., press Enter at the end of
a line instead of letting the word processor do it for you), and never use
tabs. In MS Word, when saving the file, be sure to specify plain text, and
you must continue to specify the suffix as .tex (otherwise, it will use .doc
as its suffix). If you want to check that the file is really plain text, enter
EDIT myfile.tex
4 1 OVERVIEW
at the DOS command line and see how the file appears. (There are some
free text editors on the web; use your favorite search engine to find them.)
Once you have your source ready to compile, enter latex myfile, and if
all is well, enter your dvi viewer. Under unix or DOS use dvips to convert the
dvi file to a postscript (ps) file, which can be printed. You can alternatively
(or also) convert the dvi file to a pdf file by the command
dvipdf myfile
These steps are given in Figure 3. Execute these commands for the source
file shown in Figure 2. The result should be one line of output: Hello World.
Congratulations!
myfile.tex myfile.dvi myfile.ps
create/edit view/print print/post
compile
latex
convert
dvips
myfile.pdf dvipdf
Figure 3: Command Sequence from Source to Postscript or PDF
Now change your document to specify a font size of 12pt (default is 10pt)
by changing your first line to the following:
\documentclass[12pt]{article}
The “pt” (abbreviation for “point”) is one of the units of measurement, about
1
72
in; other units used in many parts of L
A
T
E
X are in (inches), cm (centime-
ters), and em (like the letter m, which is a printer measure equal to the width
of M in the current font).
This book is designed for quick entry into using L
A
T
E
X, but do not be
reluctant to read the last chapter. It tells you how to define your own com-
mands and how to separate them into an input file that simplifies changing
things, like notation. I also cannot elaborate just yet on “using packages,”
indicated in Figure 1, except to say that they are used to fulfill some func-
tion, and I shall introduce specific packages throughout this book. (One of
the strengths of L
A
T
E
X is the community of people who provide packages for
everyone to use at no cost.) The orientation here is by function, beginning
with how to write text.
5
2 Text
We begin by illustrating the most common text formatting, much like you
would want in a word processor. (The power of L
A
T
E
X will be evident when
we get to mathematical expressions, but even some text, especially tables,
will demonstrate the superior quality of the L
A
T
E
X results.) First, consider
how to make sections and subsections in article style. Figure 4 is the source
that produces the result in Figure 5, showing how sections and subsections
are defined. Note the automatic numbering, and how extra spaces and blank
lines have no effect.
\documentclass[12pt]{article}
% We have defined the document to be an article using 12 point font.
% Blank lines mean nothing here, in the preamble.
\begin{document} % Begin document "environment".
\section{This is a Section}
\subsection{This is a subsection}
This is the body of the subsection.
I can move to a new line anytime, and I can put in lots
of blanks with no effect.
Skipping four lines is the same as skipping one line
--- it starts a new paragraph.
\subsection{Here is another subsection}
\section{Here is another section}
\end{document}
Figure 4: An Introductory Document Source (Result in Figure 5)
6 2 TEXT
1 This is a Section
1.1 This is a subsection
This is the body of the subsection. I can move to a new line
anytime, and I can put in lots of blanks with no effect.
Skipping four lines is the same as skipping one line — it starts
a new paragraph.
1.2 Here is another subsection
2 Here is another section
Figure 5: An Introductory Document Result (Source in Figure 4)
2.1 Fonts and Paragraphs
Figure 6 shows the source to produce different paragraph positions: cen-
tered, flush left, flush right, and justified (the default). Note that these are
environments, a concept you need to understand about L
A
T
E
X. The general
form of an environment uses the following syntax:
\begin{environment}
.
.
.
\end{environment}
\begin{center}
The text is centered because I have entered the center environment.
Text remains centered as long as we remain in this environment.
\end{center}
\begin{flushleft}
Now we are out of the centering environment, and have begun the
flushleft environment.
\end{flushleft}
\begin{flushright}
This is another paragraph, but in the flushright environment.
You will have occasion to use all four paragraph positions.
\end{flushright}
I am back to normal justification. The added space you see between
the above paragraphs is due to entering those environments.
Figure 6: Positioning Paragraphs Source (Result in Figure 7)
2.1 Fonts and Paragraphs 7
The text is centered because I have entered the center
environment. Text remains centered as long as we remain in
this environment.
Now we are out of the centering environment, and have begun
the flushleft environment.
This is another paragraph, but in the flushright environment.
You will have occasion to use all four paragraph positions.
I am back to normal justification. The added space you see
between the above paragraphs is due to entering those environ-
ments.
Figure 7: Positioning Paragraphs Result (Source in Figure 6)
Instead of the center environment, you can use the \centerline com-
mand; they differ in that the environment skips a line before and after the
paragraph, shown in Figures 8 and 9.
This precedes center environment.
\begin{center} This line is centered. \end{center}
This continues after centering.
This precedes centerline.
\centerline{This line is centered.}
This continues after centering.
Figure 8: Centering Source (Result in Figure 9)
8 2 TEXT
This precedes center environment.
This line is centered.
This continues after centering.
This precedes centerline.
This line is centered.
This continues after centering.
Figure 9: Centering Result (Source in Figure 8)
You can also suppress indentation of the first line of a paragraph with
the \noindent command. Here is an example:
\noindent This paragraph is not indented. produces:
This paragraph is not indented.
Table 1 lists the fonts that are intrinsic in a basic latex installation. (More
fonts are available in packages, usually free of charge.) In technical writing,
you will have particular use for the italic font, as it is used when introducing
a new term. For example,
A \textit{group} is defined on a set of elements \dots
⇒A group is defined on a set of elements . . .
(The ⇒ symbol can be read as “produces.”) Note the use of the \dots
command, which produces the ellipsis.
Table 1: Intrinsic Font Styles
What you write How it appears
This is \textbf{boldface}. ⇒ This is boldface.
This is \textit{italic}. ⇒ This is italic.
This is \textrm{roman}. ⇒ This is roman.
This is \textsc{small caps}. ⇒ This is small caps.
This is \textsf{sans serif}. ⇒ This is sans serif.
This is \textsl{slanted}. ⇒ This is slanted.
This is \texttt{typewriter}. ⇒ This is typewriter.
Some combinations of font styles can be produced. For example,
\textbf{\textit{bolditalic}} ⇒ bolditalic.
The argument of \textbf is \textit{bolditalic}. The general form is
`textfont{text}, where font is one of {bf, it, rm, sc, sf, sl, tt}, as
seen in Table 1.
2.1 Fonts and Paragraphs 9
Not all combinations are in the basic L
A
T
E
X2
ε
installation. In particular,
you must put \usepackage[T1]{fontenc} in your preamble to obtain:
\textbf{\textsc{bold small caps}} ⇒ bold small caps.
Font size can also be varied at will. Figures 10 and 11 give the source and
result for common variations. Notice how the paragraph spacing changes
to accommodate the variation in font size. These size variations can be
combined with font styles, such as using {\Large\textbf{heading}} for
some heading.
You can make the text {\large large} or {\Large larger} or
even {\LARGE larger still}. You can also make it {\huge huge}.
You might want to make something {\small small} or
{\footnotesize smaller} or even {\scriptsize smaller still}.
You can make it really {\tiny tiny}.
Figure 10: Some Font Sizes Source (Result in Figure 11)
You can make the text large or larger or even larger still.
You can also make it huge. You might want to make some-
thing small or smaller or even smaller still. You can make it really
tiny.
Figure 11: Some Font Sizes Result (Source in Figure 10)
The use of braces to enclose a font size specification is like an environ-
ment. Optionally, we can explicitly use the environment syntax: \begin{size}
. . . \end{size}. For example, \begin{large} This is large.\end{large}
produces the same result as {\large This is large.}. The environment
syntax is useful when you want to keep the size for a large block of text,
and the braces format is useful for short phrases. (There is no intrinsic
environment for font styles.)
It is straightforward to underline text — just write \underline{text}.
We can also frame text just by writing \frame{text}. We can give frame
some room around the edges by using \fbox instead. (More on framing in
§6, p. 71.) To overline is as straightforward, but learning it must wait until
we enter math mode.
Now consider ways to indent a block of text. Here is an example using the
quote environment, which was generated by putting \begin{quote} before
the text and \end{quote} after it:
10 2 TEXT
The construction of the real number system, notably by Dedekind
cuts, was motivated by the need to fix calculus, which ran into
trouble due to insufficient rigor in dealing with limits.
The quote environment is intended for short quotes, generally one short
paragraph (as above), or a sequence of one line quotes, separated by blank
lines. The quotation environment is used for long quotations, having more
than one paragraph (separated by blank lines). The indentation is the same
as the quote, except the first line of each new paragraph is indented. (Just
as in the regular text, this can be overridden by the \noindent command.)
Here is an example that was created by putting \begin{quotation} before
the text and \end{quotation} after it.
“Computers do not dream, any more than they play. We are
far from certain what dreams are good for, but we know what
they indicate: a great deal of information processing goes on far
beneath the surface of man’s purposive behavior, in ways and
for reasons that are only very indirectly reflected in his overt
activity.”
— Alan M. Turing
“There are reports that many executives make their decisions
by flipping a coin or by throwing darts, etc. It is also rumored
that some college professors prepare their grades on such a basis.
Sometimes it is important to make a completely ‘unbiased’ de-
cision; this ability is occasionally useful in computer algorithms,
for example in situations where a fixed decision made each time
would cause the algorithm to run more slowly.”
— Donald E. Knuth
The quotes are by two pioneers of algorithms, Alan M. Turing and Donald
E. Knuth. Their names appear on the right, after their quote, by skipping
a line and entering \hfill (which means horizontal fill), to make the line
flush right. Here are some other things to notice about this example:
• There are left and right quotation marks. I used ‘‘ ’’, not " ", to
create this more stylistic quotation punctuation.
• The dash that appears before each name is created by three minus
signs, ---. The more minus signs you use, the longer the dash. The
convention is that one dash is for hyphenation, two are for ranges,
such as page numbers, and three are for punctuation — i.e., use ---
preceding “i.e.”
2.1 Fonts and Paragraphs 11
• There is extra space between the two quotations. This was done with
the \bigskip command.
Figures 12 and 13 illustrate three levels of skipping: small, medium and
big. Later, we shall look at a way to have a much finer range of vertical
spacing.
This is a first line. \bigskip
The space you see above is a big skip. \medskip
The space you see just above is a medium skip. \smallskip
The space you now see just above is a small skip.
This is just an ordinary line space.
Figure 12: Skipping Line Spaces Source (Result in Figure 13)
This is a first line.
The space you see above is a big skip.
The space you see just above is a medium skip.
The space you now see just above is a small skip.
This is just an ordinary line space.
Figure 13: Skipping Line Spaces Result (Source in Figure 12)
The verse environment indents oppositely: lines after the first. The fol-
lowing was generated by putting \begin{verse} before the text and \end{verse}
after it:
Neglect of mathematics works injury to all knowledge, since he
who is ignorant of it cannot know the other sciences or the
things of this world. And what is worse, men who are thus
ignorant are unable to perceive their own ignorance and so do
not seek a remedy. — Roger Bacon
The italics were specified in the usual way, by enclosing Bacon’s verse
with \textit{. . . }. (Designed for poetry, each line is a stanza in the verse,
12 2 TEXT
and if a stanza runs long, this form of indentation makes it clear.) Bacon’s
name appears flush right, again from the \hfill command, but this time it
is on the last line of the verse, rather than a new line. This is achieved by
not skipping a line after the verse:
\begin{verse}
\textit{Neglect of mathematics ...
} \hfill --- Roger Bacon
\end{verse}
2.2 Lists
There are three intrinsic list environments, distinguished by what appears
at the beginning of each item: number, bullet, or your description (perhaps
nothing). To illustrate, here is the use of a description list environment
to itemize steps involved in learning L
A
T
E
X, whose source is indicated by
Figure 14.
Basic Document Preparation. Knowing how to setup the latex source
file, make paragraphs, vary fonts, and list items are enough to prepare
a basic document without mathematics or tables (like a resume).
Making Tables. L
A
T
E
X provides a means to make tables with the tabular
environment, and its versatility puts it far ahead of word processors.
Bibliography. Knowing how to create a bibliography, in particular with
BibT
E
X.
Mathematics. This is a power of L
A
T
E
X and one reason why it has become
standard in writing mathematical papers. I will show you how to do
virtually any mathematical expression in line with the text, or in math
display mode.
Graphics. This has progressed a great deal in the past few years thanks to
many people who have provided packages free of charge.
Other. There are a great many things to learn beyond the simple introduc-
tion when using L
A
T
E
X to prepare a thesis, report or article.
Two new things appear in the example: the use of \LaTeX to produce
L
A
T
E
X, and the use of ~ (called “tilde”) to enter a space. Without the
tilde, the result would be L
A
T
E
Xprovides, even with a space after \LaTeX in
the source. (The reason is that a space (or some delimiter) is needed after
\LaTeX (or any keyword) in order to distinguish it completely, and one might
want a punctuation mark, like a comma, following \LaTeX, which requires
no space.)
2.2 Lists 13
\begin{description}
\item [Basic Document Preparation.] Knowing how to setup ...
\item [Making Tables.] \LaTeX~ provides a means ...
\item [Bibliography.] Knowing how to create a bibliography ...
\item [Mathematics.] This is the power of \LaTeX~ and one ...
\item [Graphics.] This has progressed a great deal in the ...
\item [Other.] There are a great many things to learn ...
\end{description}
Figure 14: Description List Environment
The text within the square brackets is an option. If present, as in this
example, it is printed in boldface. With no option, the description list is one
way to have text indented the opposite of a normal paragraph: the first line
is at the left and subsequent lines are indented. For example,
\begin{description}
\item \textsf{This is how one item in a description list
environment looks with no optional text at the beginning.}
\end{description}
produces the following result:
This is how one item in a description list environment looks with no optional
text at the beginning.
Unlike the verse environment, the first line goes almost to the left margin,
and the lines extend all the way to the right margin.
Next, Figures 15 and 16 illustrate the itemize list environment, which
prints bullets. Note the indentation of each item and the spacing between
items. You see the nesting of two itemize lists, but any type of list can be
nested within any other type.
• This is item 1 and our task has just begun. Blank lines
before an item have no effect.
• This is item 2 and we shall limit to just this few.
A blank line within an item does create a new paragraph,
using the indentation of the itemize environment.
– A second (nested) itemized list changes the bullet and
indents another level.
Figure 16: Itemize List Environment Result (Source in Figure 15)
14 2 TEXT
\begin{itemize}
\item This is item 1 and our task has just begun. Blank lines
before an item have no effect.
\item This is item 2 and we shall limit to just this few.
A blank line within an item does create a new paragraph,
using the indentation of the itemize environment.
\begin{itemize}
\item A second (nested) itemized list changes the bullet
and indents another level.
\end{itemize}
\end{itemize}
Figure 15: Itemize List Environment Source (Result in Figure 16)
Finally, I describe the enumerate list environment, where the default
numbering is with Arabic numerals. With nested enumeration, the number-
ing changes at each level. Figures 17 and 18 illustrate with three levels of
nesting.
\begin{enumerate}
\item This is item 1, and we are having fun.
\item This is item 2, and it’s time to number anew.
\begin{enumerate}
\item Back to item 1, but we are not yet done.
\item Two is new.
\begin{enumerate}
\item One again!
\item Two (b) or knot 2b?
\end{enumerate}
\end{enumerate}
\end{enumerate}
Figure 17: Enumerate List Environment Source (Result in Figure 18)
2.3 Making Tables 15
1. This is item 1, and we are having fun.
2. This is item 2, and it’s time to number anew.
(a) Back to item 1, but we are not yet done.
(b) Two is new.
i. One again!
ii. Two (b) or knot 2b?
Figure 18: Enumerate List Environment Result (Source in Figure 17)
2.3 Making Tables
A table is made with the tabular environment, which has the following
syntax:
\begin{tabular}{column specs} options
first row spec \\
.
.
.
last row spec [\\ options]
\end{tabular}
As indicated, each row ends with two backslashes, \\. Each column spec-
ification can be left, center or right, abbreviated by just one character:
l, c or r, respectively. In the body of the table, each column is separated by
&. Figure 19 shows an example of a 2 3 table.
How it appears What you write
left center right
1 2 3
\begin{tabular}{lcr}
left & center & right \\
1 & 2 & 3
\end{tabular}
Figure 19: A 2 3 Table
We can draw a horizontal line before any new row by specifying \hline.
To draw a line after the last row, enter \\ \hline (the \\ is simply part
of the syntax and does not add an extra row to the table). The column
specifications can have | on either side to indicate a vertical line. Figure 20
illustrates a combined use of these options.
We could draw lines that span some rows and/or columns. The way to
vary vertical line drawing is with the column specifications: put | only where
you want a vertical line. The way to vary horizontal line drawing is by using
16 2 TEXT
How it appears What you write
-110 -120.12 -130
210 220. 230
\begin{tabular}{|l|c|r|} \hline
-110 & 120 & -130 \\ \hline
210 & -220 & 230 \\ \hline
\end{tabular}
Figure 20: A 2 3 Table with Horizontal and Vertical Lines
\cline{first col -last col } instead of \hline. This is illustrated in Figure 21.
How it appears What you write
Name Test 1 Test 2
Bob 67 72
Sue 72 67
\begin{tabular}{l|cc|}
Name & Test 1 & Test 2 \\ \cline{1-1}
Bob & 67 & 72 \\
Sue & 72 & 67 \\ \cline{2-3}
\end{tabular}
Figure 21: A Table with Partially Spanning Horizontal and Vertical Lines
We can have tables nested within tables. Figures 22 and 23 illustrate
this, while showing more variation with lines and using various fonts. Here
are some things to note:
• The entire table uses sans serif font style. This is done by specifying
\textsf{ before entering the tabular environment (and closing with }
just after it).
• Within the tables, fonts are varied: Roman is in the Roman font, speci-
fied by \textrm{Roman}, Greek is in italic, specified by \textit{Greek},
and upper case is in small caps, specified by \textsc{upper case}.
• A new column specification is introduced: p{length}, where any unit
of measure can be used as the length of the spacing. In this example
.3 inches is specified. Note that this counts as a column, so you see &&
to separate the two tables, each being a column of the main table.
• The \underline command is used to underline Table 1, which is col-
umn 1 of the main table, whereas \cline{3-3} is used to underline all
of column 3 of the main table, headed by Table 2.
There are times when we want to put a good bit of text into some columns
2.3 Making Tables 17
\textsf{
\begin{tabular}{lp{.3in}l} \\
\underline{Table 1} && Table 2 \\ \cline{3-3}
\\
\begin{tabular}{|lc|} \hline
Object & Symbols used \\ \hline
variable & lower case \textrm{Roman} \\
parameter & \textit{Greek} \\
constant & \textsc{upper case} \textrm{Roman} \\
\end{tabular}
&& % Begin Table 2
\begin{tabular}{|rcc|} \hline
* & 1 & 2 \\ \cline{2-2}
& 3 & 4 \\ \cline{1-1}\cline{3-3}
\end{tabular}
\end{tabular}
} % end sf
Figure 22: Nested Tables Source (Result in Figure 23)
Table 1 Table 2
Object Symbols used
variable lower case Roman
parameter Greek
constant upper case Roman
* 1 2
3 4
Figure 23: Nested Tables Result (Source in Figure 22)
of a table. Suppose, for example, we write the following:
\begin{tabular}{|l|l|} \hline
This amount of text is too long to fit on one line of
the page. & This is column 2. \\ \hline
\end{tabular}
The result will be to run off the edge of the paper:
This amount of text is too long to fit on one line of the page. This is column 2.
One solution is to insert new rows and break up the text manually:
18 2 TEXT
\begin{tabular}{|l|l|} \hline
This amount of text is too long to fit on one
& This is column 2. \\
line of the page. & \\ \hline
\end{tabular}

This amount of text is too long to fit on one This is column 2.
line of the page.
Instead, one can assign a fixed width to a column by specifying p{length}.
For example,
\begin{tabular}{|p{2in}|l|} \hline
This amount of text is too long to fit on one line of the page.
& This is column 2. \\ \hline
\end{tabular}

This amount of text is too long
to fit on one line of the page.
This is column 2.
Another solution is to use the \parbox command (short for “paragraph
box”). This has the form \parbox[option]{width}{text}, where the option
is the placement: t = top and b = bottom (default is center). Here are two
examples:
\begin{tabular}{|l|l|} \hline
\parbox{2in}{This amount of text is too long to fit on
one line of the page.} & This is column 2. \\ \hline
\end{tabular}

This amount of text is too long
to fit on one line of the page.
This is column 2.
\begin{tabular}{|l|l|} \hline
\parbox[t]{2in}{This amount of text is too long to fit on
one line of the page.} & This is column 2. \\ \hline
\end{tabular}

This amount of text is too long
to fit on one line of the page.
This is column 2.
They differ only in the placement of the paragraph box, the latter being at
the top to align it with column 2 in the manner shown.
When making a column or parbox small, the spacing can become un-
sightly due to being justified. This is overcome with the flushleft environ-
ment. Figures 24 and 25 illustrate this, and note that it contains other
commands that can be in any paragraph.
2.3 Making Tables 19
\begin{center}
\begin{tabular}{ll}
\parbox[t]{3in}{\begin{flushleft}
This is column 1, and I might want to display something:
\medskip\centerline{\fbox{How sweet it is.}}\medskip
This is not the same as
\medskip\fbox{\centerline{How sweet it is.}}
\end{flushleft} }
& \parbox[t]{1in}{\begin{flushleft}\textsf{This is column 2,
which I have put in sans serif font.}
\end{flushleft} }
\end{tabular}
\end{center}
Figure 24: \parbox Source (Result in Figure 25)
This is column 1, and I might want to
display something:
How sweet it is.
This is not the same as
How sweet it is.
This is column
2, which I have
put in sans serif
font.
Figure 25: \parbox Result (Source in Figure 24)
Any measurement, such as the width of a paragraph box, can be deter-
mined by some length parameter, rather than a fixed constant. For example,
see exercise 9 at the end of this chapter and consider \parbox{.3\linewidth}...
If we want some heading to span several columns, this is done by the
command, \multicolumn{number}{col spec}{entry}. The first argument is
the number of columns to span (starting where \multicolumn is specified).
This must be in the range 1 to however many columns remain from the
current position. The second argument is any valid column specification,
such as l, c, r, with, or without, a vertical line specification, |, on either
side. Finally, the third argument is the text.
The \multicolumn command can also serve to override some column
specification. Suppose, for example, we want the columns to be left justified,
but we want the headers to be centered. Figures 26 and 27 illustrate these
uses of \multicolumn. The first is used to center ‘Test number’ over columns
2 and 3. The line in the source begins with & to skip column 1, then the
20 2 TEXT
\multicolumn specifies 2 columns, centered with vertical lines before and
after. The second use simply centers the ‘Student’ header. The last use of
\multicolumn centers ‘Taken in class’ over columns 2 and 3. Unlike the first
use, the vertical line at the end is missing because c was specified instead of
c|.
\begin{center}
\begin{tabular}{l|cc|c}
& \multicolumn{2}{|c|}{Test number} \\
\multicolumn{1}{c|}{Student} & 1 & 2 & Average \\ \hline
Bill & 67 & 72 & 70.5 \\
Charleetah & 72 & 67 & 70.5 \\ \hline
& \multicolumn{2}{c}{Taken in class} \\ \cline{2-3}
\end{tabular}
\end{center}
Figure 26: Multicolumn Source (Result in Figure 27)
Test number
Student 1 2 Average
Bill 67 72 70.5
Charleetah 72 67 70.5
Taken in class
Figure 27: Multicolumn Result (Source in Figure 26)
Tables that are too long to fit on one page could be broken manually,
but the longtable package enables automatic page breaks by the L
A
T
E
X com-
piler. (You obtain the package from CTAN [4].) In the preamble, specify
\usepackage{longtable}. Then, instead of the tabular environment, specify
the longtable environment, which has most of the same options.
2.4 Special Characters
We have already seen that some characters are special, in that they delimit
something. In particular, \ delimits every L
A
T
E
X command, and % ends a
line, enabling comments. How do we print such characters? One way is with
the symbol, itself, like \%. Other times a keyword, like \textbackslash, is
needed. The Appendix contains complete tables of these special characters
(including those I do not cover explicitly in the chapters). Of particular
importance are the braces, written as \{ \} to obtain { }. (Recall that the
braces by themselves create a local environment, like {\large. . . }.)
When using a keyword to specify a special character, it appears with
whatever font is active. Table 2 (next page) illustrates this with commonly
2.4 Special Characters 21
used special characters. The brackets, [ ], are different because they can
be entered directly, except when they are used to delimit an option in the
syntax, in which case they can be obtained by enclosing them in braces.
One example is the description list environment, illustrated in Figure 28.
(Another is in the tabular environment (page 15), but I omitted a discussion
of position options that are specified by brackets.)
How it appears What you write
This is option for item.
\begin{description}
\item [This is option] for item.
\end{description}
[This is not option] for item.
\begin{description}
\item {[This is not option]} for item.
\end{description}
Figure 28: Obtaining Brackets in a Description List Environment
Table 2: Writing Special Characters
Character How you
Other fonts
(Roman) write it italic large
{ } \{ \} { } { }
% $ # & _ \% \$ \# \& \_ % $ # & _ % $ # & _
\ \textbackslash \ \
^ \textasciicircum ^ ^
~ \textasciitilde ~ ~
r ( \textregistered r ( r (
[ ] {[ ]} [ ] [ ]
Another way to print the unprintable is with the verbatim environment
or the \verb command. Unlike all other commands, \verb does not use
braces to delimit its argument. It uses any other character to delimit a
string, which can contain any character except itself. For example, we can
write \verb@{}%$#\@ to generate the string {}%$#\ (delimited by @), which
is printed in typewriter font style. The verbatim environment uses the usual
syntax:
\begin{verbatim}
.
.
.
\end{verbatim}
22 2 TEXT
This is how the source code was created for the figures, like Figure 26 (p. 20).
Another class of special characters are letters with accents. Table 3 shows
some common examples; a complete table is in the Appendix. For example,
write Poincar\’{e} to have Poincaré and G\"{o}del to have Gödel. An
accent could be applied to any letter, even if it does not relate to some
language — for example, \"{b}\~{c}\^{d} ⇒
¨
b˜c
ˆ
d.
L
A
T
E
X has a basic library of accents and special characters for writing in
languages other than English, some of which are shown in Appendix Table 29
— e.g., ?‘⇒¿ and \aa⇒å. In some cases, these are not sufficient, particularly
if the entire document is to be in a non-English language. For that purpose
there are some packages, such as Babel [1] (also see [5, Chapter 9]).
Table 3: Some Accents for Letters
What you write What you see
\"{a} ⇒ ä
\‘{e} ⇒ è
\’{i} ⇒ í
\~{o} ⇒ õ
\^{u} ⇒ û
2.5 Tabbing
The tabbing environment provides an alternative to the tabular environment
by letting you set your own column tabs. Table 4 shows a simple case with
two basic tabbing commands, \= to define a tab setting, and \> to move
to a tab setting. In addition, \\ ends each row, but unlike the tabular
environment, the first sentence continues normally, without extra spaces, so
that the position of the tab is not equivalent to that of a table’s column.
Table 4: The Tabbing Environment
What you see What you write
Begin: set tab 1. . . set tab 2
skip to 1 then to 2
skip to 2
\begin{tabbing}
Begin: \=set tab 1\dots \=set tab 2\\
\>skip to 1 \>then to 2\\
\> \>skip to 2
\end{tabbing}
Sometimes we do not want to have the longest portion of text first, yet
it is needed to define the tab. Table 5 illustrates how this is solved with the
2.6 Line, Page, and Word Breaks 23
\kill command. In the first tabbing, the lines are in the order we want,
but the tab is set by the shorter string ‘1-3’, making ‘4-6-8’ extend past
the tab. The second tabbing puts the longer field first, in order to set the
tab correctly, then specifies \kill instead of \\ to suppress (or “kill”) the
printing of the line.
Table 5: The \kill Tabbing Command
What you see What you write
1-3 sting like a bee
4-6-8 don’t be late
\begin{tabbing}
1-3 \= sting like a bee \\
4-6-8 \> don’t be late \\
\end{tabbing}
1-3 sting like a bee
4-6-8 don’t be late
\begin{tabbing}
4-6-8 \= don’t be late \kill
1-3 \> sting like a bee \\
4-6-8 \> don’t be late \\
\end{tabbing}
Figures 29 and 30 illustrate the tabbing environment with fixed field
widths. It first uses the \hspace* command for horizontal spacing, then it
uses the name of the last field to set what follows.
\begin{tabbing}
\= \hspace*{.5in} \= \hspace*{2in} \= Last field: \= \kill
\> Field 1 (following tab 1)
\\ \> \> Field 2 on new line \> Last field
\\ \> \> \> Last field on new line
\end{tabbing}
Figure 29: Tabbing Source (Result in Figure 30)
Field 1 (following tab 1)
Field 2 on new line Last field
Last field on new line
Figure 30: Tabbing Result (Source in Figure 29)
2.6 Line, Page, and Word Breaks
You can cause a new line by entering \linebreak. When text is justified (the
default), this could result in an undesirable appearance, like the following:
24 2 TEXT
\textsf{This example is \linebreak extreme.}
⇒This example is
extreme.
The \newline command forces a new line without justifying it.
\textsf{Here is the extreme \newline example.}
⇒Here is the extreme
example.
The \nolinebreak command works analogously, preventing a line break,
even if it means extending into the right margin. Also, it is better style to
keep certain ‘words’ together, such as ‘figure 1’ or ‘p. 10’. To prevent a line
break where you want a blank, use the space character ~. We would thus
write figure~1 or p.~10.
There are two commands to force a page break: \pagebreak and \newpage.
The \newpage command follows the analogy with \newline in forcing a page
break precisely at the point it is specified, rather than completing the line as
\pagebreak does. The \nopagebreak command disallows a page break im-
mediately following the next blank line. The \samepage command prevents
a page break within its scope. Here is an example that keeps line 1 on the
same page as line 2.
{\samepage
line 1
\nopagebreak
line 2
}
Word breaks are hyphenations that L
A
T
E
X does for you. Sometimes,
however, you want to suppress hyphenation. This can be done by specifying
\usepackage{hyphenat} in the preamble. (You might have to download
the package from CTAN [4], and follow the simple installation instructions.)
Then, to suppress hyphenation, you specify the \nohyphens command. For
example, the first sentence of this paragraph has a hyphen; to suppress it,
enter:
\nohyphens{Word breaks are hyphenations that \LaTeX\ does for you.
Sometimes, however, you want to suppress hyphenation.}
Then, you obtain the following:
Word breaks are hyphenations that L
A
T
E
X does for you. Sometimes,
however, you want to suppress hyphenation.
2.7 Spacing 25
2.7 Spacing
We have already seen the use of ~ to insert one space and \hfill to put the
remaining text flush right. The most versatile method to insert horizontal
spaces is with \hspace and \hspace*. These have one argument: the
amount of space to be inserted. For example,
I insert .3~in \hspace{.3in} here. ⇒ I insert .3 in here.
The \hspace command has no effect at a line boundary, but the \hspace*
inserts the space no matter what. For example, the previous
sentence is written as:
The \verb|\hspace| command has no effect at a line boundary, but
the \verb|\hspace*| \hspace*{1in} inserts the space no matter what.
That is why you see the 1 inch space at the beginning of the second line.
\hspace would not insert the 1 inch, but \hspace* does.
Two variations of \hfill are:
\dotfill . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
\hrulefill
Analogously, vertical spacing is controlled by \vspace, \vspace* and
\vfill. The height of one line of normal text is in the keyword \baselineskip,
so vspace{\baselineskip} skips one line at the next new line. The verti-
cal space is not added if this goes to the top of a new page; that is what
\vspace* does. In particular, at the very beginning of your document, if you
want to make your own title page, you use \vspace*{2in} to put a 2 inch
margin at the top (\vspace would not insert the space).
The easiest way to control line spacing throughout your document is to
specify usepackage{setspace} in your preamble. This gives you three
commands:
\singlespacing \onehalfspacing \doublespacing
Right after you specify one of these, that spacing will commence.
You can, however, specify \renewcommand{\baselinestretch}{1.2}
to increase the spacing modestly (actual increase depends upon the font
size). This acts like renewing the \arraystretch setting and remains in
effect until changed. However, one difference is that you need to change font
size before this change goes into effect. You could write
\renewcommand{\baselinestretch}{1.2}\small\normalsize
Exercises. Submit a printed copy of both the L
A
T
E
X source (tex file) and
the associated postscript result (ps file). Be sure your name is on each.
1. Write a paragraph in article [and letter] style with the following prop-
erties:
26 2 TEXT
(a) Each font style in Table 1 is used as one letter in a word that has
more than one letter.
(b) Each font style in Table 1 is used for one complete word.
(c) Each font style in Table 1 is used for two consecutive complete
words.
2. Write two paragraphs in article [and letter] style with each of the fol-
lowing properties:
(a) Default indentation on both paragraphs.
(b) No paragraph is indented.
(c) Both paragraphs are indented.
(d) There is added space between paragraphs.
3. Write a paragraph in article style and make a cover page with the
following properties (like the cover page of this document):
• All lines are centered.
• The title appears first in very large letters.
• Your name appears second in letters that are not as large as the
title, but larger than normal size, and it is preceded by extra
space.
• Your e-mail address appears third.
• Your web address appears fourth.
• Course number and title appears next, with extra space preceding
it.
• Date appears last, with extra space preceding it.
4. Give an enumeration of at least three things you like about mathemat-
ics. Give the same list without numbers.
5. Produce the following table:
Colors
Primary Secondary
Red Green
Blue Orange
Yellow Purple
6. Produce the following table (including the accents and alignments).
2.7 Spacing 27
Mathematician Birth Death
Gabrielle Emilie Le Tonnelier
de Breteuil Marquise
du Châtelet 1706 1749
Benjamin Banneker 1731 1806
Sophie Germain 1776 1831
Julius König 1849 1913
Rózsa Péter 1905 1977
7. Produce the following table.
Payoffs ($)
Player A Player B
1 2 3
4 5 6
1 2
3 4
5 6
8. How can you have an entire table whose columns are of fixed width?
9. Create a 3-column text such that each column is a paragraph of arbi-
trary length using about
1
3
of the page width each.
10. Use the tabbing environment to produce the following:
apples integral derivative
grapefruit sum difference
variables constants
11. Use the tabbing environment to produce what you see on page 30.
12. Produce the following:
rate of mass
accumulation
in the
compartment
=
net rate of mass
entering the
compartment by
convection
+
net rate of
mass entering
by diffusion.
28 3 BIBLIOGRAPHY WITH BIBT
E
X
3 Bibliography with BibT
E
X
3.1 Overview
It might seem strange to have this section so early, instead of with §7.4. That
is because I require students to produce an annotated bibliography early in
the semester, and I want them to use BibT
E
X. So here we are.
BibT
E
X [11] was developed by Oren Ptashnik and is available free of
charge. It reads a plain text file, called a bib file (plus one of the files
created by the latex compiler, about which you need not be concerned). The
bib file contains the bibliographic database, which could extend beyond one
document. The bibtex program that you apply to your source creates another
file (which you need not examine), from which a second latex compilation
causes the bibliography to be created. The execution looks like this (same
under unix and DOS):
latex myfile
bibtex myfile
latex myfile
You might have to compile with latex more times, until you do not have
any “unresolved” bibliography citations. Once this is successful, you do not
have to bibtex myfile again until you change your bib file or add a citation.
This added loop is illustrated in Figure 31.
myfile.tex myfile.dvi myfile.ps
create/edit view/print print/post
compile
latex
convert
dvips
bibtex
myfile.pdf dvipdf
Figure 31: Adding bibtex to the Command Sequence
3.2 The bib File
3.2.1 Main body
For purpose of this introduction, suppose your bibliography is in a file called
mybiblio.bib, but that name is arbitrary as long as it ends with .bib. We
3.2 The bib File 29
begin with the most important part of your bib file, which are the entries
you want to include in its database. Each entry has the following form:
@type {label ,
field = "value",
.
.
.
}
For example, Knuth’s book [8] would be entered as follows:
@article{tex,
author = "Donald E. Knuth",
title = "The {\TeX} Book",
publisher = "Addison-Wesley Publishing Company",
address = "Reading, MA",
year = "1989",
edition = "15th",
}
Most authors develop a style to labeling bibliographic entries. The use
of one keyword is somewhat simplistic and could cause problems with a
great number of entries because the labels must be unique. We cannot, for
example, have two entries with tex as their label. Here are two styles I have
seen, which you might consider:
Form Example
author.year knuth.89
author:first_keyword_in_title knuth:tex
With two authors, you can put both of their names; with more than 2,
you can add et al. (Linguistically, the use of the Latin et al. in formal writing
follows this rule.) In the first form, if there are two publications by the same
authors in the same year, some people add a, b, . . . after the year (no blank).
In the second form, if there are two publications by the same authors in the
same year, some people add another keyword. You must discover what style
works best for you.
Before listing each style (article is one style) and the fields they can or
must have (author is one field), here are a few things to note.
• The label is arbitrary, but do not use any L
A
T
E
X special characters or
blanks. In the example, the label is specified as tex and it must be
followed by a comma. Also, labels are case-sensitive, so tex is not the
same as TeX.
• Each bib entry must have a unique label, so it can be cited without
ambiguity in the source file.
30 3 BIBLIOGRAPHY WITH BIBT
E
X
• The order of the fields is arbitrary, and fields are separated by commas
(hence the comma after the terminal quote). The last field does not
require a comma at the end, but it will not hurt anything, and it gives
flexibility if you want to add a field or change the order.
• Fields do not have to be on separate lines, but it is more readable that
way.
• The field value can be anything recognized by L
A
T
E
X, even mathemat-
ical symbols in math mode.
• There is a final } to close the entry — @type{ . . . }.
Remember to put each author’s name as first last or last, first. If you
put Knuth Donald, the compiler will think the first name is Knuth and the
last name is Donald.
Here is a list of standard entry types with their required fields. What are
“optional fields” in BibT
E
X are not necessarily optional as far as having a
complete bibliography citation. For example, the volume and page numbers
of an article are necessary to include even though they are optional to sat-
isfy BibT
E
X syntax. (What is “necessary” depends upon the standard one
applies, but most journals require the volume of the journal and the page
numbers for the cited article.) Fields that are neither required nor optional
are ignored, even if they are valid fields in other types of entries.
article refers to an article from a journal or magazine.
Required fields: author, title, journal, year.
Optional fields: volume, number, pages, month.
book refers to a book with an explicit publisher.
Required fields: author or editor, title, publisher, year.
Optional fields: volume or number, series, address, edition, month.
booklet refers to a bound, printed document, but without an explicit
publisher.
Required fields: author or key, title.
Optional fields: author, howpublished, address, month, year.
inbook is a part of a book, such as a chapter or just some range of pages.
Required fields: author or editor, title, chapter and/or pages,
publisher, year.
Optional fields: volume or number, series, type, address, edition,
month.
incollection is a part of a book having its own title.
Required fields: author, title, booktitle, publisher, year.
Optional fields: editor, volume or number, chapter series,
type, pages. address, edition, month.
3.2 The bib File 31
inproceedings is an article in a conference proceedings.
Required fields: author, title, booktitle, year.
Optional fields: editor, volume or number, series, pages,
month, organization, publisher, address.
manual is some technical documentation.
Required fields: author or key (see note below). title.
Optional fields: author, organization, address, edition,
month, year.
mastersthesis is a Master’s thesis.
Required fields: author, title, school, year.
Optional fields: type, address, month.
misc is when nothing else fits.
Required fields: author or key (see note below).
Optional fields: author, title, month, howpublished, year.
phdthesis is a PhD thesis.
Required fields: author, title, school, year.
Optional fields: type, address, month.
proceedings
Required fields: title, year.
Optional fields: editor, volume or number, series, publisher,
organization, address, month.
techreport is a report published by some institution.
Required fields: author, title, institution, year.
Optional fields: type, number, address, month.
unpublished is a document with an author and title, but not formally
published, even as a technical report. (Some note of explanation is
required.)
Required fields: author, title, note.
Optional fields: month, year.
In addition to the optional fields listed, which vary by the type of entry,
the note field is always an option. This lets you enter a note that will appear
at the end of the citation. To have a comment that is not printed, enter an
unrecognized field, such as comment = "...", (this is ignored with no error
message given).
If a required field is missing when you compile, you will get an error
message. Possibly there will be some standard fixup, but it is best if you
provide the missing field value. If a document has no author, you must
provide a key for sorting. For example, consider the following entry for
L
A
T
E
X2
ε
[10], which has no person as an author. The bibliography will be
sorted with the key, LaTeX2e, used to order this entry relative to author
32 3 BIBLIOGRAPHY WITH BIBT
E
X
names. The key will not be printed.
@manual{usrguide,
key = "LaTeX2e",
title = "{\LaTeXe} for authors",
type = "World Wide Web site",
institution = "Comprehensive {\TeX} Archive",
address = "{CTAN\url{/macros/latex/doc/usrguide.ps}
(see~\cite{CTAN} for replacing CTAN)}",
year = "1995--99",
}
When there are multiple authors, we separate them with and (no com-
mas). For example, [5] in this document has the following BibT
E
X entry:
@Book{companion,
author = "Michel Goosens and Frank Mittelbach and
Alexander Samarin",
title = "The {\LaTeX} Companion",
publisher = "Addison-Wesley Publishing Company",
address = "Reading, MA",
year = "1994",
}
The use of the braces in {\LaTeX} is to tell the bibtex program to take
everything inside just as it is written (for the latex program to process).
Otherwise, the bibtex program might try to process it itself and produce an
unintended result. This applies to accents too. In ordinary L
A
T
E
X, we write
G\"{o}del to produce Gödel, but this will not work in BibT
E
X. Instead,
we write G{\"{o}}del, or simply G{\"o}del.
The use of braces to force a particular result is necessary in other in-
stances, such as writing {F}ourier analysis to force the capital F; oth-
erwise, the bibtex program will produce ‘fourier analysis’ (the plain style
produces article titles in lower case, except the first letter of the first word).
Some authors, however, use this feature inappropriately by putting braces
around everything. That defeats one of the primary advantages of using
L
A
T
E
X and BibT
E
X in the first place: we want to let the style files deter-
mine the final form, so we can switch styles and use the same source (tex
and bib files).
3.2.2 Web citations
When BibT
E
X was developed, the World Wide Web did not exist. Now it
is a major source of information. There is no universally accepted standard
for how to reference web documents, but here is one way.
If it is a book, use the book type and specify:
3.2 The bib File 33
publisher = "World Wide Web",
address = "url ",
Here is an example:
@book{Strunk,
author = "William Strunk{, Jr.}",
title = "Elements of Style",
publisher = "World Wide Web",
address = "http://www.columbia.edu/acis/bartleby/strunk/",
year = "1999",
note = "This is the web version of the classic book by
Strunk and White~\cite{StrunkWhite}",
}
The reference \cite{StrunkWhite} presumes there is the entry for the
original publication. The use of the braces in the name is to be sure that
the author appears as intended: William Strunk, Jr. Otherwise, without the
braces, the comma would signal the bibtex program that ‘Jr.’ is the first
name of the author, and it would appear as ‘Jr. William Strunk’.
If the document is a technical report, use that style but include the url
as a note or in the address field. Eventually, you will run into some difficulty
with writing urls. For one thing, the url could contain special characters; in
particular, ~ is in many urls, and writing it will produce a space, not the
tilde. Also, a url could become very long, and with latex having no place to
break, you will see a line with lots of spaces (for justification), followed by
the url. An unsightly line with spaces could also appear after the url. These
difficulties are overcome by specifying:
address = "\url{http://www.columbia.edu/acis/bartleby/strunk/}",
The \url specification is not actually an intrinsic command in L
A
T
E
X;
it is defined in a package. Its main use is to determine where the url can
be broken in order to put it on two lines, if needed. Another feature of
the url package is that \url prints special characters, like ~. To have the
\url command active in your document, put the following declaration into
your preamble: \usepackage{url}. The default font it uses is tt, but
you can change this to TimesRoman or sans serif font with the specification:
\urlstyle{rm} or \urlstyle{sf}, respectively.
There are occasions when we want to reference an entire web site. One
example is the L
A
T
E
X2
ε
reference [2], given by:
@misc{latex2e,
author = "Johannes L. Braams and David P. Carlisle and
Alan Jeffrey and Frank Mittelbach and Chris Rowley
and Rainer Sch{\" o}pf",
title = "{\LaTeXe} and the {LaTeX}3 Project",
34 3 BIBLIOGRAPHY WITH BIBT
E
X
howpublished = "World Wide Web,
\url{http://www.latex-project/org/latex3.html}",
year = "1994",
}
We have seen several packages so far, and you shall learn more about
packages in §6, where I describe enhancements for having graphics in L
A
T
E
X.
However, this is the first use of \renewcommand, about which I shall say more
when I describe ways to customize your document in §??.
3.2.3 Additional features
One element of good style is to be consistent in your terms, including abbre-
viations and names of publishers. One sometimes sees “Kluwer,” other times
“Kluwer Academic Publishers,” and still other times “Kluwer Pub.” To help
be consistent and to save some work in the long run when we write many
different documents and produce more bib files, we can define strings with
the entry:
@string{name = "string"}
Then, we can refer to the string anywhere in the value of a field by excluding
the quotes. (That is why we needed the quotes before, when we wrote
literals.)
For example, suppose we define:
@string{kluwer = "Kluwer Academic Publishers"}
Then, we can enter:
publisher = kluwer,
to produce the publisher value = “Kluwer Academic Publishers.” Besides
consistency, an advantage is that if some name changes, we merely change
the one string value and recompile.
We can concatenate strings and/or literals with #. For example, suppose
we write
@string( mom = "My Mother" )
@string( dad = "My Father" )
author = mom,
title = mom # dad,
editor = dad,
Then, the three field values are equivalent to:
author = "My Mother",
title = "My MotherMy Father",
editor = "My Father",
3.2 The bib File 35
Note the absence of a space between the string values in the title. To ensure
a space, use the space character, ~, as a literal:
title = mom # "~" # dad,
The same title as the above is obtained by any of the following:
title = "My Mother " # dad,
title = mom # " My Father",
Another useful feature of BibT
E
X is the crossref field for cross refer-
encing. For example, suppose we have the following entry (kluwer is a string;
the other values are literals):
@Proceedings{Byrnes:FAA-89,
editor = "J.S. Byrnes and Jennifer L. Byrnes",
title = "Recent Advances in {F}ourier Analysis and its
Applications: Proceedings of the {NATO}
{A}dvanced {S}tudy {I}nstitute",
publisher = kluwer,
year = 1990,
}
Then, we can have the following entry:
@InProceedings{Artemiadis:FAA-89-311,
crossref = "Byrnes:FAA-89",
author = "N.K. Art{\’e}miadis",
title = "Results on the Absolutely Convergent Series
of Functions and of Distributions",
pages = "311--316",
}
If these were the only references, the result would appear as follows:
[1] N.K. Artémiadis. Results on the absolutely convergent series of functions
and of distributions. In Byrnes and Byrnes [2], pages 311–316.
[2] J.S. Byrnes and Jennifer L. Byrnes, editors. Recent Advances in Fourier
Analysis and its Applications: Proceedings of the NATO Advanced Study
Institute. Kluwer Academic Press, 1990.
BibT
E
X also recognizes a preamble in our bib files to enable us to define
some L
A
T
E
X commands. The general form is
@Preamble{ string }
where string is any concatenation of literals and strings. Here is an ex-
ample [11] that is useful for guiding the sorting of references in a special
circumstance:
36 3 BIBLIOGRAPHY WITH BIBT
E
X
@Preamble{ "\newcommand{\noopsort}[1]{}" }
The \newcommand is something I shall describe more fully in §??. For
now, it is used to define a command, \noopsort, requiring one argument.
Command \noopsort ignores the argument that it receives, producing noth-
ing (indicated by {}). Here is how this can be used.
Suppose there is a 2-volume work by the same authors, originally pub-
lished 1971, but a second edition of volume 1 is printed in 1973. The bib
entries would have the years in the opposite order than we want because
sorting is first by the authors, which are the same, then by year. To force
the first volume to sort before the second, we fool the bibtex program with
the following specifications:
Volume 1 Volume 2
year = "{\noopsort{a}}1973", year = "{\noopsort{b}}1971",
This fools the bibtex program into thinking the years are a1973 and b1971,
thus putting volume 1 first. The definition of \noopsort, however, does not
actually print the letters, so just the years appear.
3.3 Declaration and Citation
At the end of your source file (where you want the bibliography to appear),
before \end{document}, put the following commands (in either order):
\bibliography{mybiblio}
\bibliographystyle{plain}
The first declares the bibliography to be in the bib file, mybiblio.bib.
The second command defines the format style of the bibliography to
be plain, which comes with every installation of latex. There are other
bibliography format styles, including some provided by publishers. Here is
a list of the most basic ones (included in every installation):
plain is the most common because it formats entries according
to accepted standards. Entries are sorted by the alphabetical
order of author names, breaking ties with the year of publi-
cation, and they are labeled with numbers.
abbrv differs from plain by abbreviating names of journals,
among other things (to give a more compact bibliography).
alpha differs from plain by citing by labels, rather than num-
bers.
unsrt differs from plain by sorting entries by the order in which
they are cited, rather than by the author names.
3.4 Some Controls 37
We shall use only the plain style here, but know that many other styles
have been written and are available free of charge.
To cite particular references, the L
A
T
E
X command is \cite{label [,. . . ]},
where label is what we put in our bib file entry. For example, [8] is produced
by specifying \cite{tex}. You can put more than one citation, separated by
commas. For example, \cite{tex,latex} produces [8, 9] for this document.
You can insert some further citation information as an optional input
argument to the \cite command. For example, \cite[p.~46]{latex} pro-
duces [9, p. 46] in this document. (In the option, delimited by [ ], the ~ is
used to ensure that there is a space but no line break when giving the page
number as “p. 46” in the citation.)
The rule is that only those bib entries that are cited appear in the final
document. The reason is that we can maintain one large bib file and write
many documents that use it. There are times when we want to be sure a
particular bib entry appears, but we do not want to cite it in the text. This
is done with the \nocite command. In particular, if we want to have every
entry in our bib file appear, we specify \nocite{*}. If we want only some
particular list of entries to appear, we use \nocite with their labels, such as
\nocite{tex} to be sure Knuth’s T
E
X book appears, even if it is not cited
explicitly. Figure 32 shows a complete source file for having all entries in
mybiblio.bib appear, and that is the entire document!
\documentclass[12pt]{article}
\begin{document}
\nocite{*}
\bibliographystyle{plain}
\bibliography{mybiblio}
\end{document}
Figure 32: A Document to Print the Bibliographic Database
We can specify more than one bib file, such as:
\bibliography{mybiblio,another}
The bibtex program will search them sequentially for any citation. If we have the
same label in both bib files, the entries must be identical; otherwise, we will get
a fatal error message, Repeated entry– telling us which label is repeated. If we
have the same entry with different labels, they will appear twice if both labels are
used (or if we used \nocite{*}).
3.4 Some Controls
You might want to have the bibliography single-spaced, even if the main document
is spaced differently. Just before the \bibliography statement, put
\renewcommand{\baselinestretch}{1} \selectfont
38 3 BIBLIOGRAPHY WITH BIBT
E
X
The \selectfont command is needed to activate the change in \baselinestretch.
If you also want to change the font size (or style), that will do. Thus, you could
specify
\renewcommand{\baselinestretch}{1} \small
to have the bibliography set in small font size (and single spaced); \selectfont is
not needed.
You can change the heading put by bibtex. The default is Bibliography, but
you might want to have it be References. You do this by specifying (sometime
before the bibliography):
\renewcommand{\refname}{References} for article style
\renewcommand{\bibname}{References} for book and report styles
You can have more than one bibliography by using the bibunits package, which
is available at CTAN [4]. This will let you put different bibliographies throughout
your document, and you can change the properties, such as the heading.
Exercises. Submit a printed copy of the L
A
T
E
X source (tex file), the BibT
E
X
data (bib file), and the associated postscript result (ps file). Be sure your name is
on each.
1. Produce a document with one paragraph that cites three bibliographic items,
one for each of the following types:
(a) An article in a journal.
(b) An entire book with at least three authors.
(c) A chapter in a book.
(d) A technical report.
2. Produce a document that lists your entire database, which consists of at least
one entry for each of six different document types. Further, at least one entry
must have more than two authors.
3. Produce a document with one paragraph that cites three bibliographic items,
one for each of the following types:
(a) A technical report on the web.
(b) A book on the web.
(c) An entire web site.
4. Produce a document that has only a bibliography composed of the following
three entries (in the order shown).
[1] I.M. Rich, editor. Impossible Dreams, volume I. MacTaco, second edition,
1999.
[2] I.M. Rich, editor. Impossible Dreams, volume II. MacTaco, 1990.
[3] I.M. Smart, U.R. Tu, and V.F. Money. How to Square a Circle, chapter 1.
Volume II of Rich [2], 1990.
39
5. Produce an annotated bibliography of the following form (note the indenta-
tions on left and right margins):
[1] P.R. Halmos, Naive Set Theory, Van Nostrand, Princeton, NJ, 1960.
This is a good book, which I assign to my Ph.D. students. The first
100 pages seem simple. The next 100 reveal lack of understanding
the first 100.
[2] G. Polya, How To Solve It, Princeton University Press, Princeton, NJ,
1945.
This is a seminal book that articulates the problem-solving — i.e.,
theorem-proving — process. There are many editions, and there are
modern descendants, such as . . .
4 Counters, Labels, and References
4.1 Basic Concepts
A counter is a numerical value that refers to something that is being numbered,
such as pages, sections, figures, and equations. A label is the identification of
a particular value, and a reference is a citation to a label. The L
A
T
E
X syntax
for labeling a counter is \label{label }, placed where the counter’s value is set,
where label is unique in the document. The L
A
T
E
X syntax for referencing a label is
\ref{label }. For example, in this book I defined:
\section{Bibliography with \Bibtex} \label{sec:Bibliography}
Now I can refer to §3 by \S\ref{sec:Bibliography}. The choice of label is ar-
bitrary, except do not use L
A
T
E
X special characters or blanks, just as the labels in
the bib file entries.
There are times when you just want to produce the counter value, without a
label. This is done by \thecounter. For example, \thepage produces the page
number. On the other hand, if you want to use the counter’s numerical value as an
argument in a command, specify \value{counter}.
In the next section I describe intrinsic counters and illustrate how to label and
reference them. Then, I shall introduce the figure and table environments, which
have intrinsic counters associated with them.
4.2 Intrinsic Counters
Anything to which L
A
T
E
X assigns a number has a counter associated with it. Here
I illustrate some of those that are in all document styles. Counters that depend
upon the style, like a chapter in a book, can be labelled and referenced in the same
manner.
You are looking at page 39, which I was able to print by writing \thepage.
Similarly, you are reading subsection 4.2 of section 4, whose numbers I could write
by \thesubsection\ and \thesection, respectively.
To illustrate how I can reference other parts of this document, the following
labels were defined (when the subsection and subsubsection were first written):
40 4 COUNTERS, LABELS, AND REFERENCES
\subsection{The bib File} \label{subsec:bibfile}
\subsubsection{Web citations} \label{subsubsec:webcite}
Then, I can refer to these as follows:
\S\ref{subsec:bibfile} ⇒ §3.2
\S\ref{subsubsec:webcite} ⇒ §3.2.2
I can also refer to their page numbers:
p.~\pageref{subsubsec:webcite} ⇒ p. 32
p.~\pageref{subsec:bibfile} ⇒ p. 28
For any counter, \pageref{counter}, gives the page number where its label is
defined, just as \ref{counter} gives its value. (Recall from p. 24 that ~ is used to
have a space without a linebreak, which is an element of good style.)
Equation (6), page 60, was labelled \label{eqn:hessian}, so in this sentence
I wrote its number by \ref{eqn:hessian} (with parenthesis added) and its page
number by \pageref{eqn:hessian}. In the exercise to list what you like about
mathematics, I entered the label \label{exer:likeaboutmath} (page 26), which
I can now reference as exercise #4 by writing \#\ref{exer:likeaboutmath}.
The choice of label, such as subsubsec:webcite, is any string you want to use
that does not contain embedded blanks or special characters used by L
A
T
E
X. In
my choice of label, I used the structure prefix:name. That is a matter of style,
and I used the prefix subsec here. This helps me to distinguish labels for different
things. For equations, I use the prefix eqn. Some people use this same form but
with different prefixes, such as ss for subsection and e for equation. You can choose
any labeling convention that is meaningful to you. (If you have a lot of labels and
need to keep track of them by printing each label and citation in your drafts, see
the showkeys package at CTAN [4].)
4.3 Figures and Tables
In this section I describe figure and table environments, which have the same
syntax:
\begin{figure}[options] \begin{table}[options]
[\caption{caption[\label{label }]}] [\caption{caption[\label{label }]}]
.
.
.
.
.
.
[\caption{caption[\label{label }]}] [\caption{caption[\label{label }]}]
\end{figure} \end{table}
The caption, if present, can go at the top or bottom; where you put it is where
it will appear. The label to reference a figure or table is put inside the caption.
(If you put it outside the caption, as given in [9], it will not be understood, even
though you will get no error message.)
Because figures and tables are not split, their exact location depends upon how
much room there is. For that reason they are called “floating objects,” or floats.
The environment options define where the float is to be located. The four choices
4.3 Figures and Tables 41
are shown in Table 6. In this document most tables and figures are specified with
[ht], which means they are to be placed “here,” the place where it is specified in
the source, if possible. If there is not enough room, it is to be located at the top of
the following page.
Table 6: Figure and Table Location Options
Option Meaning
h Locate here (where the environment is declared).
t Locate at the top of the next page.
b Locate at the bottom of the page (or the next page,
if this page does not have enough room).
p Locate on a separate page, called a float page,
which has no text, only figures and tables.
The placement of a float is sometimes a source of frustration. We might specify
[ht] and find the float in an unexpected place, perhaps on a page by itself. One
cause could be an accumulation of floats that should be cleared at some point
before continuing. This is done with the \clearpage command. This does the
same as \newpage, except that it also prints all remaining floating objects. It
is also advisable to specify \usepackage{float} in the preamble. One of the
enhancements is the placement option: [H], which insists that the float be placed
here (note the capital H and no other option specified). This option is used in many
places in this book, which is why you sometimes see pages with some blank space
in the lower portion, followed by a figure or table. I did this to avoid confusion by
having some float appear pages after it is cited and discussed.
The table environment is not to be confused with the tabular environment. The
latter makes tables, but the table environment does not have to contain a table; it
differs from a figure only in its name, and they have separate counters. The figures
and tables in this document appear as the form:
Figure number : caption vs. Table number : caption
That’s it.
As a matter of style, we generally use the figure environment to present what
we usually think of as figures, notably pictures, and we generally use the table
environment to present information in tabular form. However, neither of these
conditions is necessary for their L
A
T
E
X environments.
Floats can be framed, using the \fbox command. For example, Figures 33 and
34 illustrate how to frame a figure with a thick border.
42 4 COUNTERS, LABELS, AND REFERENCES
\begin{figure}[ht]
\begin{center}
\setlength{\fboxrule}{3pt} % make border lines thick
\setlength{\fboxsep}{.2in} % increase distance to border
\fbox{ This is a framed figure. }
\end{center}
\caption{Framed Figure with Caption at Bottom \label{fig:fboxbottom}}
\end{figure}
Figure 33: Framed Figure 34 Source
This is a framed figure.
Figure 34: Framed Figure with Caption at Bottom
The parameter settings have returned to their default values, upon leaving the
figure environment. (This is called a local setting.) Thus, the frame in Figure 35
has thin lines and no extra padding around the border. Also note how the caption
is put at the top (see exercise 3).
Figure 35: Framed Figure with Caption at Top
This is framed with default parameter values.
4.4 Defining Your Own
In the preamble you can define your own counter with the \newcounter command:
\newcounter{name}[within]
where name is the (unique) name of the counter (cannot be the same as one of
the intrinsic counter names). The initial value of the counter is 0. For example,
\newcounter{mycounter} defines a counter whose name is mycounter. You can
also define the counter to be within another counter. For example,
\newcounter{mycounter}[section]
defines mycounter to be within the section counter. This will cause the value of
mycounter to be reset to 0 when entering a new section. Further, instead of the
printed values being 1, 2, . . . , they will be 1.1, 1.2, . . . within section 1; more
generally, the values of mycounter will be of the form :.1, :.2, . . . when printed
within section :.
The counter values are printed in Arabic numerals, but you can specify the type
of numeral, shown in Table 7.
4.4 Defining Your Own 43
Table 7: Numerals to Print Counters
What you see What you write
a, b, c, d, . . . \alph{mycounter}
A, B, C, D, . . . \Alph{mycounter}
1, 2, 3, 4, . . . \arabic{mycounter}
i, ii, iii, iv, . . . \roman{mycounter}
I, II, III, IV, . . . \Roman{mycounter}
Counter values can be incremented with the \addtocounter command. For
example, \addtocounter{mycounter}{1} adds 1 to the value of mycounter. If we
just want to increment the counter by 1, we can specify \stepcounter{mycounter}.
Counter values can be set to some absolute value with the \setcounter command.
For example, \setcounter{mycounter}{5} sets the value of mycounter to 5. This
can also be used to transfer the value of one counter to another. For example,
\setcounter{mycounter}{\value{page}}
sets the value of mycounter to the current page number (value of the intrinsic
counter, page).
When using a counter for some non-intrinsic sequence, we want to be able to
label it for future reference. This is done with the \refstepcounter command,
which also increments its value. For example, to increment mycounter by 1 and
establish a label to its value at the place this is done, write
\refstepcounter{mycounter} \label{mylabel}
Then, we can use \ref{mylabel} and \pageref{mylabel} wherever we like.
The default numeral type is arabic, but you can change the appearance to be
any of those listed in Table 7 by applying the \renewcommand to \thecounter. For
example,
\setcounter{mycounter}{0}
\renewcommand{\themycounter}{\roman{mycounter}}
\stepcounter{mycounter} (\themycounter),
\stepcounter{mycounter} (\themycounter), \dots
⇒ (i), (ii), . . .
This can be used for intrinsic counters too. For example, consider the enumerate
list environment, where the types of numerals for the four levels are: arabic, alph,
roman and Alph. We can change these to be any type we want, such as illustrated
in Figures 36 and 37.
The second level, whose counter is enumii, had its label changed to what is
specified in the source: \renewcommand{\labelenumii}{\theenumii.} (the “ap-
pearance” parameter is \labelenumii). These changes remain in effect (called a
global setting), so we must change them back if we want to restore the defaults,
shown in Table 8.
44 4 COUNTERS, LABELS, AND REFERENCES
\renewcommand{\theenumi}{\Roman{enumi}}
\renewcommand{\theenumii}{\Alph{enumii}} % changes numeral type
\renewcommand{\labelenumii}{\theenumii.} % changes appearance
\begin{enumerate}
\item Introduction
\item Terms and Concepts
\begin{enumerate}
\item Groups and fields
\item Picnics and frolic
\end{enumerate}
\end{enumerate}
Figure 36: Alternative enumerate Symbols Source (Result in Figure 37)
I. Introduction
II. Terms and Concepts
A. Groups and fields
B. Picnics and frolic
Figure 37: Alternative enumerate Symbols Result (Source in Figure 36)
Table 8: Default Settings for enumerate Counters
What
Counter changes Command
enumi numeral \renewcommand{\theenumi}{\arabic{enumi}}
label \renewcommand{\labelenumi}{(\theenumi)}
enumii numeral \renewcommand{\theenumii}{\alph{enumii}}
label \renewcommand{\labelenumii}{(\theenumii)}
enumiii numeral \renewcommand{\theenumiii}{\roman{enumiii}}
label \renewcommand{\labelenumiii}{(\theenumiii)}
enumiv numeral \renewcommand{\theenumiv}{\Alph{enumiv}}
label \renewcommand{\labelenumiv}{(\theenumiv)}
Exercises. Submit a printed copy of the L
A
T
E
X source (tex file) and printed copy
of the associated postscript result (ps file). Be sure your name is on each.
1. Write a document with at least two pages and two sections. Put an enumer-
ated list of items near the beginning of your document, and use the \ref or
\pageref command to reference each of the following.
(a) Reference §2 by a label that you assign to section 2 (make whatever
45
label name you like).
(b) Somewhere near the end of your document reference the page number
of the first section.
(c) Reference item #2 of your enumerated list.
2. Include two tables and figures in your document, and reference them by label.
Also reference the page that they appear.
3. Produce Figure 35.
4. Produce lists using the enumerate environment with the following appear-
ance:
1. . . .
1.1 . . .
1.2 . . .
2. . . .
2.1 . . .
2.2 . . .
5. Produce two numbered lists such that the second starts its numbering where
the first leaves off. For example, produce the following:
1. List 1, item 1
2. List 1, item 2
Now we are out of list 1 . . . begin list 2.
3. List 2, item 1
4. List 2, item 2
5 Math Mode
One can write mathematical expressions by entering math mode, signified by de-
limiters $. . . $ or \[. . . \]. The $ delimiter keeps the mathematical expression in
the text, like this:
A consequence of Einstein’s postulates is that $E = mc^2$.
⇒ A consequence of Einstein’s postulates is that 1 = :c
2
.
The other form is math display mode, like this:
A consequence of Einstein’s postulates is that \[E = mc^2.\]
⇒ A consequence of Einstein’s postulates is that
1 = :c
2
.
46 5 MATH MODE
Table 9: Some Mathematical Operations
Example
Operation Symbol How it appears What you write
subscript _ r
3
x_3
superscript ^ r
3
x^3
multiply \times o / a\times b
divide \div o ÷/ a\div b
5.1 Mathematical Symbols
The example also illustrates the use of the superscript operator, ^. Table 9 shows
other common operations in math mode. (Each of the tables in this section applies
only to math mode.)
The braces enclose an expression that can be used to define a more com-
plex operand. For example, r
a+b
is written as $x_{a+b}$ and r
a
2
is written as
$x^{a^2}$. The order of subscripts and superscripts does not matter:
x_{a+b}^{c+d} ⇒ r
c+d
a+b
x^{c+d}_{a+b} ⇒ r
c+d
a+b
Table 10 shows some set notation. The complement of ¹ often appears as ∼ ¹,
produced by $\sim A$, but this is not universal notation; some authors use ¹
c
or
¹

, produced by $A^c$ and $A^\prime$, respectively, and some use ¹, produced
by $\overline{A}$.
Preceding any symbol by \not puts the line through the symbol, as in the
following examples:
A\not\subseteq B ⇒ ¹ ⊆ 1
x\not\in A\cup B ⇒ r ∈ ¹∪ 1
A\setminus B\not\supset B ⇒ ¹` 1 ⊃ 1
Table 10: Set Notation
What it is How it appears What you write
empty set ∅ \emptyset
intersection ∩ \cap
union ∪ \cup
set minus ` \setminus
element in ∈ \in
subset (proper) ⊂ \subset
subset or equal ⊆ \subseteq
superset (proper) ⊃ \supset
superset or equal ⊇ \supseteq
5.1 Mathematical Symbols 47
You can control the size of the font by using the usual specification before enter-
ing math mode. For example, {\Large $(x\div y) + z$} ⇒(x÷y)+z. Font
style, however, does not apply to math mode because math mode has its own, sep-
arate from text mode. You can make math fonts boldface by specifying \boldmath
before entering math mode. For example, {\boldmath $x^n+y^n=z^n$} ⇒ x
n
+
y
n
= z
n
. Note that \boldmath is surrounded by the braces; otherwise, math fonts
would remain bold, even when leaving and re-entering. The following illustrates
this, where 1 ∪ C is boldface in the first case, and returns to normal style in the
second case.
\boldmath$A\supset B$ text $B\cup C$ ⇒ A ⊃ B text B ∪ C
{\boldmath$A\supset B$} text $B\cup C$ ⇒ A ⊃ B text 1 ∪ C
Within math mode, we can control the font style of letters with the command,
\mathfont{expression}, where font is one of: {bf, cal, it, normal, rm, sf,
tt} (analogous to the `textfont command, p. 8). Unlike \boldmath, this applies
only to letters, digits and accents, but not to special mathematical symbols. For
example,
{\boldmath$\tilde A\times\vec{1}\otimes\overline{2}$} ⇒
˜

1 ⊗2
$\mathbf{\tilde A\times\vec{1}\otimes\overline{2} }$ ⇒
˜
A
˜
1 ⊗2
Table 11 illustrates the outcome of each font for this expression:
\mathfont{\tilde A\times\vec{1}\otimes\overline{2}}
Table 11: The `mathfont Commands
Font Style Command Example Result
boldface \mathbf
˜
A
˜
1 ⊗2
calligraphic \mathcal
˜
/ ∞⊗∈
italic \mathit
˜
A
˜
1 ⊗2
normal \mathnormal
˜
¹1 ⊗:
roman \mathrm
˜
A
˜
1 ⊗2
sans serif \mathsf
˜
A
˜
1 ⊗2
typewriter \mathtt
~
A
~
1 ⊗2
The calligraphic style applies only to capital letters, causing unintended results
when applied to other symbols, as shown in Table 11. The calligraphic alphabet
looks like this (and it is available only in math mode):
/B(TcT(H1./L´^O{O1oT |1¼A\Z.
Write ${\cal P} = A + B$ to produce { = ¹ + 1; without the braces, the calli-
graphic fonts remain in effect: $\cal P = A + B$ ⇒ { = /+B.
Greek letters are defined only in math mode, and they are specified by spelling
them as keywords. For example, to produce
α −β = ∆−δ
48 5 MATH MODE
write \[ \alpha - \beta = \Delta - \delta \]. As Lamport [9, p. 43] says,
“Making Greek letters is as easy as π (or Π)” (written $\pi$ or $\Pi$). (Not every
Greek letter is included — see Appendix Table 36.) The \mathbf does not make
Greek letters boldface. We could use \boldmath to achieve this, but there is a
package that not only provides the boldface font, but also produces proper spacing.
In the preamble specify \usepackage{bm}, then \bm{\beta} ⇒β.
5.2 Fractions and Variable Size Functionality
To make fractions, we could write $(x+y)/4$ to make (r + n)´4, but if we want
x+y
4
, we use the \frac command: $\frac{x+y}{4}$. We can make this appear
larger, as
x+y
4
, by preceding the math mode with \large.
The general form is \frac{numerator}{denominator}, where the numerator
and denominator can be any expression. Here is a more complex equation in math
display mode:
¹ =
r
2
+ n
α
1 +
η
x
2
+1
.
written as \[ A = \frac{x^2+y_\alpha}{1+\frac{\eta}{x^2+1}}, \]. Note how
the sizes of the fractions adjust automatically.
Some mathematical symbols adjust their size to fit the expression. Table 12
shows some of the most common of these, and I present more examples below. In
the case of the integrals, note the use of \, between the integrand and dr. This
inserts a thin space (compare the results by writing the expression with and without
the \,).
In L
A
T
E
X, symbols whose size you would want to adapt to expressions are gen-
erally designed to do so. Figures 38 and 39 illustrate this with another example,
which uses the \sqrt and \prod functions:
\[ \sqrt{\frac{\prod_{n=1}^N \left( \sum_{i\in I_n} x_i^n\right)}
{\sqrt[3]{\sum_{i\in I_\infty} x_i}}
}
\]
Figure 38: Variable Sizes Source (Result in Figure 39)

¸
N
n=1
¸
i∈I
n
r
n
i

3

¸
i∈I

r
i
Figure 39: Variable Sizes Result (Source in Figure 38)
Notice that even though it is written in math display mode, the indices on the
sums and product appear as they would in line. L
A
T
E
X compilers make judgments
about the layout, but you can force either of the two styles with the \displaystyle
and \textstyle commands. Figures 40 and 41 illustrate this.
5.2 Fractions and Variable Size Functionality 49
Table 12: Variable Size Mathematical Operation Symbols
Operation How it appears What you write
sum
¸
\sum
n
¸
i=1
r
i
\sum_{i=1}^n x_i
integral

\int

b
a
1(r) dr \int_a^b f(x)\,dx
parentheses () \left( \right)

r
1 + n

\left(\frac{x}{1+y} \right)
braces ¦¦ \left\{ \right\}

¸
i
r
i
¸
\left\{\sum_i x_i \right\}
brackets [] \left[ \right]
¸

0
1(r) dr

\left[\int_0^\infty f(x)\,dx\right]
\[ \sqrt{\frac{\displaystyle
\prod_{n=1}^N \left( \sum_{i\in I_n} x_i^n\right)}
{\sqrt[3]{\displaystyle\sum_{i\in I_\infty} x_i}}
}
\]
Figure 40: \displaystyle Source (Result in Figure 41)
In text mode you can force the display style of placing these subscripts and
superscripts on functions, as well as sizing the expression, as though it were in
display mode. Figure 42 gives more examples to compare in line text and display
mode, using \textstyle and \displaystyle to override the default form for the
mode. The “default” is not always predictable; in particular, math display mode
does not always use displaystyle.
50 5 MATH MODE

N
¸
n=1

¸
i∈I
n
r
n
i

3

¸
i∈I

r
i
Figure 41: \displaystyle Result (Source in Figure 40)
What to write What to write
Appearance in text mode in display mode
x
2
\frac{x}{2} \textstyle\frac{x}{2}
r
2
\displaystyle\frac{x}{2} \frac{x}{2}
max
x∈X
\max_{x\in X} \textstyle\max_{x\in X}
max
x∈X
\displaystyle\max_{x\in X} \max_{x\in X}
Figure 42: Examples to Compare Text and Display Modes
Table 13 shows symbols used in logical expressions. For example, to have
(r ∈ ¹ ⇒r ∈ 1) ⇔(¹ ⊆ 1).
write \[ (x\in A\Rightarrow x\in B) \Leftrightarrow (A\subseteq B). \]
To have
∀r∃n ÷ [1(r) ∧ Q(n)].
write \[ \forall x\exists y\ni [P(x)\wedge Q(y)]. \]
The quantifiers in this last example seem a bit crowded, so we might want
to add some spacing between terms. In math mode a full space is obtained by
specifying \; and a half space by \,. Here is how each looks:
\forall x \exists y ⇒ ∀r∃n
\forall x\, \exists y ⇒ ∀r∃n
\forall x\; \exists y ⇒ ∀r ∃n
(There are other spacing commands, including negative spacing, shown in Appendix
Table 35.)
Table 14 shows some relations for ordered sets (besides those on the keyboard:
< = >). Here are some examples:
(-\infty,0] = \{x\ni x \le 0\} ⇒ (−∞. 0] = ¦r ÷ r ≤ 0¦
a_j\prec b_i \equiv b_i \succ a_j ⇒ o
j
≺ /
i
≡ /
i
~ o
j
\forall y\,\{x: x\not\prec y\}
\not\subset {\cal A} ⇒ ∀n ¦r : r ≺ n¦ ⊂ /
5.3 Arrays and Equations 51
Table 13: Some Symbols in Logic
Logical Term How it appears What you write
existential quantifier ∃ \exists
universal quantifier ∀ \forall
negation \neg
disjunction ∨ \vee
conjunction ∧ \wedge
implication → \rightarrow
⇒ \Rightarrow
equivalence ⇔ \Leftrightarrow
≡ \equiv
such that ÷ \ni
Table 14: Order Relations
Relation How it appears What you write
less than or equal ≤ \le
greater than or equal ≥ \ge
not equal = \ne
precedes ≺ \prec
precedes or equals _ \preceq
succeeds ~ \succ
succeeds or equals _ \succeq
We have seen how to embed math mode into text, but we can also do the reverse
with the \mbox command. Compare each of the following:
r
i
< 01o:o||i = 1. . . . written as $x_i < 0 for all i=1,\dots$
r
i
< 0 for all i = 1. . . . written as $x_i < 0$ for all $i=1,\dots$
r
i
< 0 for all i = 1. . . . written as $x_i < 0 \mbox{ for all } i=1,\dots$
The first line points out that blanks mean nothing in math mode, and all letters
are in the math form of italic (not quite the same as the italic in text mode). The
use of \mbox is particularly convenient in math display mode, which I shall illustrate
in the next section.
5.3 Arrays and Equations
The array environment is to math mode what tabular environment is to text mode,
and more. It has the form:
52 5 MATH MODE
\begin{array}{column specs}options
first row spec \\
.
.
.
last row spec [\\ options]
\end{array}
The column specifications and options are the same as in the tabular environment,
but the body is in math mode. The following table has text headers and math
body, so it can be generated in either of two ways: with the tabular environment,
using the math mode designation for each body entry: $. . . $, or with the array
environment, using \mbox for each header entry.
Variable Current Value Limit
r 1.234567 1
n −9.87 −12.2
This can be generated by either of the following two ways:
\[\begin{array}{ccc}
\mbox{Variable} & \mbox{Current Value} & \mbox{Limit} \\ \hline
x & 1.234567 & 1 \\
y & -9.87 & -12.2 \\ \hline
\end{array}
\]
or
\begin{center}
\begin{tabular}{ccc}
Variable & Current Value & Limit \\ \hline
$x$ & $ 1.234567$ & $ 1 $ \\
$y$ & $-9.87 $ & $-12.2 $ \\ \hline
\end{tabular}
\end{center}
You can align a series of equations to appear this way:
r = 5.2
n = 2.5
. = 7.7 (= r + n)
The above was produced by the following use of math display mode (which is always
centered):
\[ \begin{array}{lcl}
x &=& 5.2 \\
y &=& 2.5 \\
z &=& 7.7 \; (= x+y)
\end{array}
\]
The \; specifies a space; otherwise, 7.7 (= x+y) ⇒ 7.7(= r + n).
5.3 Arrays and Equations 53
Another environment is eqnarray. This is like a 3-column array with specifica-
tions {lcl}, as above, but each row is numbered:
r = n (1)
n = . (2)
(Another difference is that the eqnarray environment uses displaystyle.) We use
the eqnarray environment directly (without entering math display mode), so the
above is produced by the following:
\begin{eqnarray}
x &=& y \label{eqn:xy} \\
y &=& z \label{eqn:yz}
\end{eqnarray}
The \label statements are to illustrate that we can reference these by writing
(\ref{eqn:xy}) to produce (1) and (\ref{eqn:yz}) to produce (2). (Note that
\ref gives just the number; parentheses are added.)
The relation need not literally be an equation; anything could be used for the
middle column. Further, there are times when we need to use more than one line
for an ‘equation,’ in which case we need to suppress the numbering of all but one
of the rows. Figures 43 and 44 give an example. The \nonumber command causes
no number to be assigned to the first part of the second equation.
\begin{eqnarray}
x &\mbox{is equal to}& y \\
y & \preceq & \frac{a+b+c+d}{\Psi} + \frac{e+f+g+h}{\Phi}
+ \nonumber \\
& & I+K+J+L
\end{eqnarray}
Figure 43: eqnarray Environment Source (Result in Figure 44)
r is equal to n (3)
n _
o + / + c + d
Ψ
+
c + 1 + o + /
Φ
+
1 + 1 + J + 1 (4)
Figure 44: eqnarray Environment Result (Source in Figure 43)
There is also an eqnarray* environment, which is the same as eqnarray, but
without the equation numbers. There is no apparent advantage to this since the
same result can be produced by the ordinary array environment, specialized to this
column specification. It does, however, let us change our mind easily as to whether
or not to include equation numbers by simply adding or removing the * from the
environment specification.
54 5 MATH MODE
For a single, numbered equation, there is the equation environment. This poses
no particular advantage over specifying eqnarray and merely entering one row
(except that column separators (&) are not used in the equation environment).
Analogous to eqnarray*, there is the equation* environment, which suppresses the
equation numbering. To illustrate, Figures 45 and 46 show how to present a matrix
equation. Also, note how r

is specified. Table 15 shows other ways to denote the
transpose of a vector.
Table 15: Transpose of a Vector
What you write How it appears
x’ ⇒ r

x^t ⇒ r
t
x^T ⇒ r
T
x^{\mathsf{T}} ⇒ r
T
x^{\mbox{\tiny $T$}} ⇒ r
T
\begin{equation*}
Ax^\prime = \left[ \begin{array}{rrr}
1.1 & 1.2 & 1.3 \\
21.0 & 22.0 & -2.1 \\
\end{array} \right]
\left( \begin{array}{cc} x_1 \\ x_2 \\ x_3
\end{array} \right).
\end{equation*}
Figure 45: Matrix Equation Source (Result in Figure 46)
¹r

=
¸
1.1 1.2 1.3
21.0 22.0 −2.1

¸
r
1
r
2
r
3

.
Figure 46: Matrix Equation Result (Source in Figure 45)
Array environments can be nested, as illustrated in Figures 47 and 48. Notice
how the vertical line was drawn by the column specification, {c|c}, and the hori-
zontal line separating the blocks is obtained by specifying \hline before the second
row of the outer array.
5.3 Arrays and Equations 55
\[ \left[ \begin{array}{c|c}
\begin{array}{ccc} A_{11} & A_{12} & A_{13} \\
A_{21} & A_{22} & A_{23}
\end{array}
& 0 \\ \hline
0 & \begin{array}{cc} B_{11} & B_{12} \\ B_{21} & B_{22}
\end{array}
\end{array} \right]
\]
Figure 47: Nested Arrays Source (Result in Figure 48)

¹
11
¹
12
¹
13
¹
21
¹
22
¹
23
0
0
1
11
1
12
1
21
1
22
¸
¸
¸
¸
Figure 48: Nested Arrays Result (Source in Figure 47)
We can enclose mathematical expressions in a box, sometimes used for empha-
sis. For example,
\fbox{$ \begin{array}{lcl}
\displaystyle\int_0^\infty xe^{-\tau x}\,dx
&=& \displaystyle\frac{1}{\tau} \\ \\
&=& \displaystyle\oint_a^{b+c} \Psi(x)\,dx
\end{array}
$}


0
rc
−τx
dr =
1
τ
=

b+c
a
Ψ(r) dr
We can use \fbox within math mode, such as writing $x = \fbox{y} + z$ to
produce r = y +.. Note how the line height does not adjust to the frame, causing
an undesirable clash. This could be overcome by putting a vertical space command
just after the expression. In particular, putting \vspace{.2\baselineskip} after
r = y +. causes extra vertical space equal to 20% of the value of \baselineskip,
which is the height of one line of normal text. (In the longrun, it is better to use
parameters, like \baselineskip, rather than absolute measurements for spacing,
because the former takes into account the font size, which you might change.)
56 5 MATH MODE
Now consider the following conditional assignment:
1(r) =

−1 if r < 0;
0 if r = 0;
1 if r 0.
produced by the following L
A
T
E
X code:
\[ f(x) = \left\{ \begin{array}{rll}
-1 & \mbox{if} & x < 0; \\
0 & \mbox{if} & x = 0; \\
1 & \mbox{if} & x > 0.
\end{array}\right.
\]
Note the use of \right. after the array. This is because \left and \right
must balance — i.e., there must be an equal number of each. It is not necessary
that the left symbol be related to the right one — i.e., \left\{ does not require
\right\} to balance; any right symbol will do. The period is not printed in this
case, used specifically for this purpose of balance.
We have seen the use of \left and \right for brackets around a matrix. Now
the use of the \left L
A
T
E
X command for conditional assignment raises related uses
of the underbrace and overbrace. Figures 49 and 50 illustrate these, along with
\overline, \underline, \widehat and \widetilde.
\[ \begin{array}{cc}
\mbox{This sum has} \\ \mbox{an overbrace} \\
\overbrace{\overline{i\dots j} + \underline{k\cdots l}}
& \underbrace{\widehat{xy} - \widetilde{ab}}
\\ & \mbox{This difference}
\\ & \mbox{has an underbrace}
\end{array} \]
Figure 49: Horizontal Braces Source (Result in Figure 50)
This sum has
an overbrace
. .. .
i . . . , + / | ´ rn −
¯
o/
. .. .
This difference
has an underbrace
Figure 50: Horizontal Braces Result (Source in Figure 49)
We often need to mix mathematical notation and text. We could use the
tabular environment and specify in-line math mode where needed (with $), or we
5.3 Arrays and Equations 57
could use the array environment and use either the \mbox or \parbox (see p. 18)
to enter the text. There are, however, some nuances to understand. Figures 51 and
52 show the problem with using \flushleft to make the text within the parbox
flush left. (Try it with the default justify and you will see that the spacing gives a
poor appearance.) The problem is that \flushleft skips a line, which ruins the
alignment (even though [t] is specified). The solution is to use the \raggedright
command, as shown in figures 53 and 54. In addition, the \raisebox command is
used to lower the small matrix, giving it some space below the horizontal line.
\renewcommand{\arraystretch}{1.2}
\begin{center} \begin{small} \begin{tabular}{lll}
Matrix & Definition & Example \\ \hline
\parbox[t]{.9in}{Covariance}
& \parbox[t]{2in}{\flushleft
$A_{ij} = E[(X_i-\mu_i)(X_j-\mu_j)]$, where $\{X_i\}$
are random variables, and $E[\cdot]$ is the expected
value operator with $\mu_i=E(X_i)$. }
& \parbox[t]{1.4in}{\scriptsize
$\left[\begin{array}{rrrrr}
\frac{1}{2}&0 \\ 0&\frac{1}{2}
\end{array}\right]$
\flushleft for $X_2=X_1^2$ and
$Pr[X_1=x]$ \\
$= \left\{ \begin{array}{lll}
\fourth &\mbox{for}& x=-1~ \vspace{.05in} \\
\half &\mbox{for}& x=~~0~ \vspace{.05in}\\
\fourth &\mbox{for}& x=~~1.
\end{array}\right.$ } \vspace{.1in} \\ \hline
\end{tabular} \end{small} \end{flushleft}
\renewcommand{\arraystretch}{1}
Figure 51: \flushleft in parbox Source (Result in Figure 52)
Matrix Definition Example
Covariance
¹
ij
= 1[(A
i
−j
i
)(A
j
−j
j
)],
where ¦A
i
¦ are random
variables, and 1[] is the
expected value operator with
j
i
= 1(A
i
).

1
2
0
0
1
2

for X
2
= X
2
1
and
Pr[X
1
= x]
=









1
4
for x = −1
1
2
for x = 0
1
4
for x = 1.
Figure 52: \flushleft in parbox Result (Source in Figure 51)
58 5 MATH MODE
\begin{center} \begin{small} \begin{tabular}{lll}
Matrix & Definition & Example \\ \hline
\parbox[t]{.9in}{Covariance}
& \parbox[t]{2in}{\raggedright
$A_{ij} = E[(X_i-\mu_i)(X_j-\mu_j)]$, where $\{X_i\}$
are random variables, and $E[\cdot]$ is the expected
value operator with $\mu_i=E(X_i)$. }
& \parbox[t]{1.4in}{\scriptsize\raisebox{-.1in}{
$\left[\begin{array}{rrrrr}
\frac{1}{2}&0 \\ 0&\frac{1}{2}
\end{array}\right]$}
\flushleft for $X_2=X_1^2$ and
$Pr[X_1=x]$ \\
$= \left\{ \begin{array}{lll}
\fourth &\mbox{for}& x=-1~ \vspace{.05in} \\
\half &\mbox{for}& x=~~0~ \vspace{.05in}\\
\fourth &\mbox{for}& x=~~1.
\end{array}\right.$ } \vspace{.1in} \\ \hline
\end{tabular} \end{small} \end{center}
Figure 53: \raggedright in parbox Source (Result in Figure 54)
Matrix Definition Example
Covariance ¹
ij
= 1[(A
i
−j
i
)(A
j
−j
j
)],
where ¦A
i
¦ are random
variables, and 1[] is the
expected value operator with
j
i
= 1(A
i
).

1
2
0
0
1
2

for X
2
= X
2
1
and
Pr[X
1
= x]
=









1
4
for x = −1
1
2
for x = 0
1
4
for x = 1.
Figure 54: \raggedright in parbox Result (Source in Figure 53)
5.4 Special Functions and Alphabets
Math mode recognizes a collection of special functions. Table 16 shows some com-
mon ones. These special functions are used to make the source clearer, rather than
using \mbox to achieve the same result.
Among the special functions are the complete set of trigonometric functions.
For example, we write $\tan\theta = \frac{\sin\theta}{\cos\theta}$ to pro-
duce: tan θ =
sin θ
cos θ
. Appendix Table 40 (p. 115) has a much longer list of special
functions, as well as the arrows used in some of the examples shown in Table 17.
There is also a package of AMS symbols, which you declare in your preamble
with \usepackage{amssymb}. This gives the following alphabet with the mathbb
font:
$\mathbb{ABCDEFGHIJKLMNOPQRSTUVWXYZ}$
⇒ ABCDEFGHIJKLMNOPQRSTUVWXYZ
5.4 Special Functions and Alphabets 59
Table 16: Some Common Mathematical Functions
Function How it appears What you write
limit lim \lim
lim inf liminf \liminf
log log \log
maximum max \max
tangent tan \tan
Table 17: Examples of Mathematical Functions
How it appears
textstyle displaystyle What you write
lim
n→∞
r
n
lim
n→∞
r
n
\lim_{n\rightarrow\infty}x_n
liminf
n↓0
log r
n
liminf
n↓0
log r
n
\liminf_{n\downarrow 0}\log\,x_n
max
x∈X
1(r) max
x∈X
1(r) \max_{x\in X}f(x)
tan(θ+π)
ln x
tan(θ + π)
ln r
\frac{\tan(\theta + \pi)}{\ln\,x}
For example, the real line is sometimes denoted by R, rather than 1, which is
the L
A
T
E
X special symbol, $\Re$. Table 18 shows how \mathbb can be used for
specifying other numerical spaces.
Table 18: Notation Using mathbb Fonts from amssymb Package
What you write

How it appears What it means
\mathbb{R} ⇒ R Real values
\mathbb{C} ⇒ C Complex values
\mathbb{Z} ⇒ Z Integer values
\mathbb{Q} ⇒ Q Rational values

In math mode.
Another alphabet is \mathscr, for which you specify \usepackage{mathrsfs}
in the preamble. This gives the following alphabet:
$\mathscr{ABCDEFGHIJKLMNOPQRSTUVWXYZ}$
⇒ABCDE FGH IJK LMN OPQRST U V W XY Z
In particular, L is often used to denote the Laplace transform or the Lagrangian,
60 5 MATH MODE
and H is sometimes used to denote the Hamiltonian. (Compare with \mathcal,
H, which is also used by some authors.)
5.5 Derivatives and Integrals
We can express a total derivative, d1(r)´dr, by writing $df(x)/dx$; or, we can use
the \frac command to produce
df(x)
dx
. The partial derivative symbol, ∂, is written
\partial, so you can write \partial f(x)/\partial x to produce ∂1(r)´∂r, and
{\large$\frac{\partial f(x)}{\partial x}$} to produce
∂f(x)
∂x
.
The usual notation for the gradient of a function is the nabla, denoted by
the symbol ∇, (also called “del”), which is an upside down delta (introduced by
Hamilton in 1853). In L
A
T
E
X it is produced by \nabla, and its mathematical
definition is the vector of first partial derivatives:
∇1(r) = (∂1(r)´∂r
1
. . . . . ∂1(r)´∂r
n
). (5)
I leave it as an exercise to show the L
A
T
E
X code that produced equation (5).
The Hessian is the matrix of second partial derivatives:

2
1(r) =
¸

2
1(r)
∂r
i
∂r
j

.
This was produced by the following code:
\[ \begin{array}{lll}
\nabla^2f(x) &=& \left[ \displaystyle \frac{\partial^2f(x)}
{\partial x_i \partial x_j}
\right].
\end{array}
\]
There seems to be some crowding in this direct specification. Compare with the
following and see if you can produce it:

2
1(r) =
¸

2
1(r)
∂r
i
∂r
j

. (6)
There are two integral signs: \int⇒

and \oint⇒

, which are both variable-
size symbols. For example, note how the outer integral is large in the following
expression:

b
a
lim
λ→∞

X(v)
rc
λf(x)
dr

X(v)
c
λf(x)
dr

Φ(·) d·.
This was obtained by the following code:
\[ \int_a^b \lim_{\lambda\rightarrow\infty} \left|
\frac{\oint_{X(v)} xe^{\lambda f(x)}\,dx}
{\oint_{X(v)} e^{\lambda f(x)}\,dx}\right|
\,\Phi(v)\,dv .
\]
5.6 Theorems and Definitions 61
(Note the use of the thin space, \,.)
Definite multiple integrals are no problem. To have


0

x
n
0

x
n−1
0

x
2
0
H(r
1
. . . . . r
n
) dr
1
dr
n
write
\[ \int_0^\infty \int_0^{x_n} \int_0^{x_{n-1}} \cdots
\int_0^{x_2} H(x_1,\dots,x_n)\,dx_1\cdots dx_n
\]
However, consider the following:

S
(n∇· −·∇n) do =

τ
(n∇ ∇· −·∇ ∇n) dτ.
The domains of integration, and the spacing of the integral signs, are better with
the following, which is not produced by standard L
A
T
E
X2
ε
, but by specifying
\usepackage{amsmath} in the preamble (see The L
A
T
E
X Companion [5, p. 223]):

S
(n∇· −·∇n) do =

τ
(n∇ ∇· −·∇ ∇n) dτ.
Note how the domains are centered on the multiple integrals and the spacing of the
integral signs.
5.6 Theorems and Definitions
The foundations of mathematics are axioms and rules of inference. The rules
create theorems, which are statements whose truths are established relative to the
underlying logic. This is so fundamental that L
A
T
E
X has the facility to define a
special environment that includes a keyword, like “Theorem,” and a name, which
is not only the name of the environment, but is also the name of the associated
counter. Consider the following example:
Theorem 5.1 For n 2, there is no solution to r
n
+ n
n
= .
n
for
r. n. . ∈ Z
++
.
Notice how “Theorem 5.1” appears, all text is in italic, and we have the
counter value: \thetheorem=5.1. This was defined in the preamble by:
\newtheorem{theorem}{Theorem}[section]
Then, the theorem was produced by the following L
A
T
E
X code:
\begin{theorem}
For $n > 2$, there is no solution to $x^n + y^n = z^n$ for
\newline $x,y,z\in \LZ_{++}$.
\end{theorem}
62 5 MATH MODE
Other theorem-like environments can be defined to have the same properties.
This requires both a keyword, like Theorem, and a unique name for the environment,
like theorem, also used as a counter. Here is the syntax:
\newtheorem{name}{keyword}[within]
The name defines the environment name, and it defines a counter, so it must be
different from all other environment and counter names. The within option defines
the counter to be within some other, which can be intrinsic or some other counter
defined by the \newcounter command (p. 42) or by some other \newtheorem.
In this document, I defined the theorem environment to be numbered within the
section, so you see Theorem 5.1, rather than Theorem 1. To further illustrate,
here is a corollary environment:
Corollary 5.1.1 The sum of cubes cannot be a cube.
It was defined in the preamble as follows:
\newtheorem{corollary}{Corollary}[theorem]
Note that this is within the theorem counter, which is valid by having been defined
by its own \newtheorem. Then, the above corollary was written as:
\begin{corollary} The sum of cubes cannot be a cube.
\end{corollary}
The following creates an axiom environment that is not within any other counter.
\newtheorem{axiom}{Axiom}
The “Axiom of Choice” can then be stated thusly:
Axiom 1 From any (infinite) family of sets a new set can be created that contains
exactly one element from each set in the family.
This was created by the following code:
\begin{axiom} \label{axm:choice}
From any (infinite) family of sets a new set can be created
that contains exactly one element from each set in the family.
\end{axiom}
The label allows us to refer to the Axiom of Choice as ‘Axiom 1 on page 62’ by
writing Axiom~\ref{axm:choice} on page~\pageref{axm:choice}.
The environment created by \newtheorem puts the text in italics, but this is
generally not desirable for a definition. Consider the following example:
Definition 5.1 The circumference of a sphere is the circumference of any great
circle on the sphere.
This was created by first entering (in the preamble):
5.7 Refinements 63
\newtheorem{defn}{Definition}[section]
Then, in the text:
\begin{defn} The circumference of a sphere is the
circumference of any great circle on the sphere.
\end{defn}
Compare this with the following:
Definition 5.1 The circumference of a sphere is the circumference of any great
circle on the sphere.
This was created by first entering (in the preamble):
{\theorembodyfont{\rmfamily} \newtheorem{mydefn}{Definition}[section]}
Then, in the text:
\begin{mydefn} The \textit{circumference of a sphere} is the
circumference of any great circle on the sphere.
\end{mydefn}
For more customization, the theorem package enables a wide range of varia-
tions over the font style (among other things).
5.7 Refinements
Mathematical delimiters, like parentheses and braces, must be varied to en-
close some expressions. Whereas \left and \right commands adjust the
size of a mathematical delimiter to fit the enclosed expression, we can also
enlarge these delimiters ourselves. One way is with a size command — for
example,
{\large(}$E=mc^2${\large)} ⇒ (1 = :c
2
).
There are, however, delimiter size control commands, which apply to a single
character: \big, \Big, \bigg, and \Bigg. For example,
$\big(E=mc^2\big)$ ⇒

1 = :c
2

.
The use of text font environments comes close to the corresponding math
size, (large↔big, . . . , Huge↔Bigg), but they are different, especially the
thicknesses. This is more evident with the square and angular brackets:
$\Big[E=mc^2\Big]$ ⇒

1 = :c
2

.
{\Large[}$E=mc^2${\Large]} ⇒[1 = :c
2
].
$\bigg\langleE=mc^2\bigg\rangle$ ⇒

1 = :c
2

.
64 5 MATH MODE
{\LARGE$\langle$}$E=mc^2${\LARGE$\rangle$} ⇒'1 = :c
2
`.
The remaining refinements use the amsmath package (introduced on p. 61
for obtaining better multiple integrals).
The gather and gather* environments allow the new line specification,
\\, in math mode. They behave like the eqnarray and eqnarray* environ-
ments, respectively, except the equations are not aligned. Figures 55 and 56
illustrate this. The same result with equation numbers is obtained by the
gather environment.
\begin{gather*}
(a+b)^2 = a^2 + 2ab + b^2 \\
{\cal L} \oplus M^\varepsilon - V = H_0 \\
A(x) = \{y: \phi(y) = \cup_{a\in \cal A} \Psi(x)\}
\end{gather*}
Figure 55: gather* Environment Source (Result in Figure 56)
(o + /)
2
= o
2
+ 2o/ + /
2
L ⊕`
ε
−\ = H
0
¹(r) = ¦n : φ(n) = ∪
a∈A
Ψ(r)¦
Figure 56: gather* Environment Result (Source in Figure 55)
When writing a matrix within text, we could produce

o /
c d

by spec-
ifying $\left(\begin{array}{cc} a&b \\ c&d \end{array}\right)$. An al-
ternative is with the amsmath smallmatrix environment:

a b
c d

is obtained
by $\left(\begin{smallmatrix} a&b \\ c&d \end{smallmatrix}\right)$.
(Note that there are no column specifications.) This is not equivalent to
preceding the array specification with a text size environment; in particular,
\scriptsize produces

a b
c d

. While the letters inside the matrix are ap-
proximately the smallmatrix size, the spacing and parentheses are not the
same.
The amsmath package has a command to put dots across any number
of columns in an array. Its syntax is \hdotsfor[spacing]{n}, where spacing
determines the spacing between the dots, and n is the number of columns it
spans. For example,
5.7 Refinements 65
\left|\begin{array}{ccccc}
1 & 2 & 3 & 4 & 5 \\
\hdotsfor{3} \\
& \hdotsfor{3} \\
\hdotsfor[2]{5} \\
\hdotsfor[.5]{5} \\
\end{array}\right|

1 2 3 4 5
. . . . . . . .
. . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . . . . . . . . .

The \stackrel command lets us put characters over a relation: For
example, n^+\stackrel{\mathrm{def}{=}n+1 ⇒ n
+
def
= n + 1. With more
generality, the \overset and \underset amsmath commands enable us to
put any characters over or under any character. For example,
$\overset{a}{X}$ ⇒
a
A
$\underset{b}{Y}$ ⇒ Y
b
$\overset{a}{\underset{b}{Z}}$ ⇒
a
7
b
This can be used to stack subscripts:
$\displaystyle{\sum_{ \stackrel{ \mbox{\scriptsize$i\in I$} }
{j\in J}
} } A_{ij}
= \underset{j\in J}
{ \underset{i\in I}{\sum} } A_{ij} $

¸
i ∈ I
j∈J
¹
ij
=
¸
i∈I
j∈J
¹
ij
Nesting the \underset command can be unwieldy, but you can use the
smallmatrix environment or the \substack command:
$\displaystyle\sum_{\substack{i\in I\\j\in J\\k\in K}} A_{ij}
= \underset{\begin{smallmatrix} i\in I\\j\in J\\k\in K
\end{smallmatrix}} {\sum} A_{ij}$

¸
i∈I
j∈J
k∈K
¹
ij
=
¸
i∈I
j∈J
k∈K
¹
ij
Another package in the ams family is amscd, which makes it easy to
draw commutative diagrams. Figures 57 and 58 illustrate this. (Specify
\usepackage{amscd} in the preamble.) The horizontal arrows are specified
by @>>> (left to right) with any expression placed above or below. The
66 5 MATH MODE
vertical arrows are specified by @VVV (down) or @AAA (up), with an expression
placed to its left or right. All possible horizontal and vertical placements are
illustrated.
\[ \begin{CD}
A @>a>> B @>>> C \\
\alpha @VV\beta V \gamma @AAA @VVV\delta \\
D @>>d> E @>e>> F
\end{CD} \]
Figure 57: Commutative Diagram Source (Result in Figure 58)
¹
a
−−−−→ 1 −−−−→ C
α

β γ
¸

δ
1 −−−−→
d
1
e
−−−−→ 1
Figure 58: Commutative Diagram Result (Source in Figure 57)
There are many more refinements, and more packages to make things
nicer. Many of these are described in The L
A
T
E
X Companion [5, Chapter 8],
and you can see an online catalog of packages at CTAN [4].
5.8 Grammar
When writing mathematical expressions, people make some common errors.
The general guide is to treat a mathematical expression linguistically. In
English this means that every sentence has a subject and predicate, clauses
are separated by commas, and phrases are appropriately punctuated. Here
are some of the most common elements of grammar to consider.
1. Punctuate math display mode. The expression usually needs a
comma or period. For example, note the colon before the display and
the comma at its end, which is incorrect to omit.
A symmetric rearrangement of a matrix has the following form:
1 = 1
t
`1.
where 1 is a permutation matrix.
2. Define before use. As you read articles notice that those that are
among the most confusing are when the authors used a term that is
not defined until pages later. For example, we might see “The distin-
guishing property of an abelian group is the commutivity . . . ” But a
group had not yet been defined.
5.8 Grammar 67
3. Reference object is located after the reference. For example, a
figure appears after its first reference. L
A
T
E
X does this automatically,
but you might want to take control over locating figures.
4. An object has only one definition. For example, if we write Φ =
on + /·, we cannot later refer to Φ(n. ·). Sometimes we define the
complete object, Φ(n. ·) = on + /·, then tell the reader something
like, “We shall use Φ
k
, instead of Φ(n
k
. ·
k
), when there is no risk of
confusion.” The overriding principle is clarity, and it is important that
the reader be told of this.
5. If . . . then . . . is not correct. Suppose ¹ and C are expressions. We
can write either ‘If ¹, C.’ or ‘Suppose ¹. Then, C.’ The first form
is preferred if ¹ and C are simple expressions. If either ¹ or C are
compound, the second form is clearer. The form, ‘If ¹, then C.’ seems
like it ought to be all right, and the comma is used to clarify where
the antecedent (¹) ends and the consequent (C) begins. In English,
however, this is not correct.
6. Equivalence needs commas. The expression, ‘¹ if and only if 1.’
should be written as ‘¹ if, and only if, 1.’
Exercises. Submit a printed copy of the L
A
T
E
X source (tex file) and of the
associated postscript result (ps file). Be sure your name is on each. (Lookup
special symbols in the Appendix.)
1. Produce each of the following in math display mode.
(a) r
2
= 1
2
−4¹C implies r = ±

1
2
−4¹C.
(b) If ∆1
n+1
= 1
n
, it follows that ∆
2
1
n+1
= ∆1
n
.
(c) r
+
=

r if r ≥ 0;
0 otherwise.
2. Produce the following in math display mode with the array environ-
ment and/or with the eqnarray* environment.

2
1
n
= 1
n+2
−21
n+1
+ 1
n
= 21
n
−1
n+1
.
3. Produce each of the following formulas in line with text (construct your
own sentences that contain them, and include proper punctuation).
(a) ln c
x
= r
(b) sin¦θ + 2π¦ = sin θ
68 5 MATH MODE
(c) n
n
=
¸
n−1
i=i
0
r
i
⇒n
n+1
−n
n
= r
n
−r
i
0
(d) 1(r) =

¸
n=0
1
(n)
(0)
r
n
n!
(e)

∂r

x
2
a
1(n)dn = 2r1(r
2
)
(f) ´
Fe(H
2
O)
6
= 6´
H
2
O

Fe
4. Produce the following equation in math display mode.
¸
1.1 −1.2 −1.3
−2.1 2.2 2.3

+
¸
o
1
o
2
o
3
/
1
/
2
/
3

=
¸
α −β γ
−δ λ θ

.
5. Produce the expression in the Preface.
6. Produce equations (5) and (6).
7. Produce each of the following expressions:
(a) r = n mod n
def
= r −n = /n for some / = 0. 1. . . .
(b)
. .. .
α
1
+

β
2
− x
2
+y
3
. .. .
8. Produce each of the following in line with text (that you compose) and
in math display mode.
(a) /
?
= ¦o ∈ o : o ∈ o¦
(b)

[T {[ ≤ π
(c)

¸
¸
¸

o
11
o
12
o
21
o
22

(B
1C

9. Produce each of the following formulas in math display mode (with
punctuation):
(a) c

(G) ≥ max

∆(G).
2:(G
A
)

¹−1

if G = ∅.
(b) (0. r
T
)T = (0. r
T
)

¹ 1
0 C

=

0
r
T
C

= (0. C
T
r)
T
,
(c) \ =
3

3
2

a
0
(−r
1
3
+ o)
2
dr;
5.8 Grammar 69
10. Combine your knowledge of derivatives, conditional assignment (with
array environment), and mathematical symbols to produce the follow-
ing (called the truncated gradient):

+
1(r)
j
=

max¦0. ∂1(r)´∂r
j
¦ if r
j
= o
j
∂1(r)´∂r
j
if o
j
< r
j
< /
j
min¦0. ∂1(r)´∂r
j
¦ if r
j
= /
j
11. Produce the following symbols:
(a) Extended reals: R

.
(b) Strictly positive integers: Z
++
.
(c) Complex n-vectors: C
n
.
(d) Non-negative rational n-vectors: Q
n
+
.
12. Produce the following:
∂f(x)
∂x
j

x=¯ x
.
13. Produce the following:
Definition 1 A matrix is singular if its determinate is zero.
Definition 2 A matrix is non-singular if it is not singular.
Theorem 1 Every non-square matrix has an inverse.
Proof: The determinate of a non-square matrix cannot be zero
because it is not defined. Therefore, the matrix is non-singular.
This implies it has an inverse.
14. What is grammatically wrong with each of the following segments.
(a) A key is how to add velocities the formula is
(n + ·)
1 +

c
2
where c is the velocity of light.
(b) A result of these assumptions is the following equation
1 = :c
2
Einstein first noticed this equivalence between energy (1) and
mass (:).
70 5 MATH MODE
(c) Let r be an n-vector and ω a scalar, and define
n = ¹r −ω/.
where ¹ is an : n matrix and / is an :-vector. Now suppose
n(ω) is specified and we want to find r.
(d) Now we consider adding velocities.
v
u+v
u
Figure 1. Adding velocity vectors: n + ·.
Figure 1 (above) shows how to add velocities simply as vectors.
(e) Theorem If r. n. . ∈ Z
+
and r
n
+ n
n
= .
n
, then n < 3.
The remaining exercises are more difficult. You are to produce the mathe-
matical expressions shown in math display mode.
15. The following is tricky to get the evaluation expression, t =
1
2
to be
the right size and location.
d
dt
1(r + t∇
+
1(r))

t =
1
2
= −

(1´2)
1/4
+ 1.
16. Note the row and column labels outside the matrix.
¹ =
o / c d c
1
2
3
4

1 0 0 0 1
1 1 0 1 0
0 1 1 0 0
0 0 1 1 1
¸
¸
¸
¸
17. Row pointers:
¹ =
¸
11 12
21 22

←rows in 1
←rows in 2 (this arrow is closer to matrix)
18. Column pointers:
¹ =
¸
11 12
21 22

↑ ↑
columns columns
in 1 in 2
71
19. Row and column pointers:
¹ =
¸
11 12
21 22

←rows in 1
←rows in 2
↑ ↑
columns columns
in 1 in 2
6 Graphics
Graphics may be part of a L
A
T
E
X document by one of three ways:
1. Use standard L
A
T
E
X2
ε
commands, notably the picture environment;
2. Use a graphics package to draw within the document;
3. Use a package to import some standard graphics file.
I illustrate each, but I do not provide a complete list of the relevant packages
(see CTAN [4] and The L
A
T
E
X Companion [5]).
6.1 Picture Environment
If all we want is a series of boxes and arrows, we can do this simply with
\fbox and a long arrow in math mode, as follows:
\fbox{left}$\longrightarrow$\fbox{center}$\longrightarrow$\fbox{right}
⇒ left −→ center −→ right
The \framebox command can be used instead of \fbox to produce the
same result. However, \framebox also has two optional arguments to control
the length of the box and the position of the text within it. For example,
\framebox[2cm][l]{left}$\longrightarrow$\framebox[2cm][c]{center}$%
$\longrightarrow$\framebox[2cm][r]{right}
⇒ left −→ center −→ right
The % at the end of the first line is to avoid having a blank between the
center box and the $\longrightarrow$ that follows it. The first optional
argument of this \framebox command is the width of the box, given as 2 cm
for each box. The second optional argument is the position of the inscribed
text: l = left, c = center, and r = right.
We can make the contents of a box obey all paragraph controls in text
mode by the \parbox command. By itself, it lets us stack short phrases,
72 6 GRAPHICS
like
top
middle
bottom
(note how the paragraph spacing adjusts). Combined with
\framebox, we can create vertical diagrams easily, as illustrated in Figures
59 and 60.
\begin{center} \parbox{2cm}{
\framebox[2cm]{top} \\ \centerline{$\downarrow$} \\
\framebox[2cm]{middle} \\ \centerline{$\downarrow$} \\
\framebox[2cm]{bottom}
} \end{center}
Figure 59: Vertical Diagram Source (Result in Figure 60)
top

middle

bottom
Figure 60: Vertical Diagram Result (Source in Figure 59)
The box created by \parbox has its center aligned with the text, but it
has an optional argument to align its top or bottom with the text. This is
done by specifying \parbox[t]{width}{text} or \parbox[b]{width}{text},
respectively.
These commands can be combined, along with other box commands, but
there is a need for more versatility, like ovals and diagonal arrows, and more
control over positioning. A basis for this is the picture environment. To
begin, Figure 61 shows a more elaborate chart, which was created by the
picture environment, whose source is shown in Figure 62. Going through its
parts will serve to explain the various commands.
,
top left
center
bottom right
`
1
o
/
c

oval
-
?
Z
Z
Z
Z~
- -

Figure 61: Variety of Objects in Picture Environment
The first command begins a center environment, and I use the \setlength
command to set the units of measurement to be 1 inch. This means that when
6.1 Picture Environment 73
\begin{center} \setlength{\unitlength}{1in}
\begin{picture}(0,0)
\put( 0, 0){\circle*{.1}}
\put( 0,-.5){\framebox(.7,.3){center} }
\put(-1,-.5){\dashbox{.01}(.7,.3)[tl]{top left} }
\put( 1,-.5){\dashbox{.1}(1.2,.3)[br]{bottom right} }
\put(-.65, -1){\circle{.2}} \put(-.7,-1.05){1}
\put( 1, -1){\oval(.5,.25)} \put(.85,-1.05){oval}
\put(0,-1){\fbox{$\begin{array}{c}a\\b\\c\end{array}$}}
\put(-.3,-.35){\vector(1,0){.3}}
\put(-.65,-.5){\vector(0,-1){.4}}
\put( .35,-.5){\vector(4,-3){.5}}
\put(-.55,-1){\vector(1,0){.55}}
\put( 0,-1){\vector(-1,0){.55}}
\put(.32,-1){\vector(1,0){.43}}
\put(1.2,-.895){\line(1,1){.3975}}
\end{picture}
\end{center} \vspace{1in}
Figure 62: Source for Figure 61
I specify some length = 5, I am specifying 5 inches. The parameter that de-
termines this is \unitlength, and the default for the picture environment is
1 pt. Then, we enter the picture environment stating that the point of entry
is the origin, indicated by the coordinates (0,0). (There is an alternative way
to begin the picture environment, which is not described here.) The filled
circle shows where (0,0) is in this picture.
Every picture command begins with \put, which is exclusively for the
picture environment. The complete syntax is: \put(r,n){stuff }, where stuff
can be text or some picture object. The (r. n) coordinates are relative to
where the position is when the picture environment is entered. This could
be at the left margin, as in beginning a paragraph with \noindent; it could
be a column in a table defined within the tabular environment; or


it
could be in the middle of a sentence, just as the smiley face appears here
(see Exercise 1).
The first \put in Figure 62 specifies the position at the origin, and the
stuff is a filled circle with diameter .1 inches (centered at the origin):
\put(0,0){\circle*{.1}} ⇒ ,
The next three commands put three different kinds of boxes, each be-
ginning at .5 inches below the origin (i.e., n = −.5). The first is similar
to \framebox in text mode, but its syntax is different. In picture mode it
enables control over not only the width, but also the height, and this ex-
tends the position options to a second character: t = top; b = bottom. The
74 6 GRAPHICS
general form of the \framebox command in the picture environment is as
follows:
\framebox(width,height)[posn]{text}
In the example shown in Figure 62, the specifications are width = .7 inches
and height = .3 inches; the position is centered because that is the default.
The next \put puts a dashed box, having the same dimensions as the
framed box, with the length of the dash set to .01 inches. The next dashed
box has the dash length set equal to .1 inches, resulting in fewer dashes to
compose the box. The box length is set to 1.2 inches, and the text is at the
bottom right because of the optional specification, [br].
Now we come to the \circle specification, located at coordinates (−.65. −1)
(from \put), with diameter = .2 inches. The “1” inside the circle required an-
other \put, and some trial and error was needed to establish its position. We
know the center of the circle is at (−.65. −1), but that is not where we want
to put the inscribed text to be centered. Unlike the box family, we cannot
include the centering of text within the circle command. The same applies
to the \oval specification, followed by putting text that required some trial
and error to locate. The oval, itself, has dimensions .5 .25 (inches), where
.5 measures the entire width:

width
-
height
6
?
After the \oval specification, I use the \fbox command. This is the
same as I used in text mode, except here I use it to frame an array, defined
as usual in math mode: the array has three rows and one column, which is
centered.
Now the code begins to draw the vectors, which are lines with arrow
heads. Both \vector and \line have the same syntax:
\line(∆r,∆n){len} \vector(∆r,∆n){len}
If ∆r = 0, the line is vertical, and len is the amount of change above or below
the original point (it does not matter what the magnitude of ∆n is; only its
sign matters). If ∆n = 0, the line is horizontal, and len is the amount of
change to the right or left of the original point (it does not matter what the
magnitude of ∆r is; only its sign matters). Otherwise, if ∆r = 0, the actual
change in r is still |cn, and the slope of the line is
∆n
∆r
. This is undoubtedly
confusing, so consider Figure 63. The new point is determined by moving
6.1 Picture Environment 75
from (r
0
. n
0
) along the line with slope
∆n
∆r
until the new r-coordinate is
r
0
+ |cn. Then, the new n-coordinate is n
0
+ |cn
∆n
∆r
. The actual length of
the line segment is |cn

1 +

∆n
∆r

2
.
slope =
∆n
∆r
r
0
n
0
r
0
+ ∆r
n
0
+ ∆n
r
0
+ |cn
n
0
+ |cn
∆n
∆r
Original
point
New
point
Figure 63: Line Parameters
As if this unnatural definition of the line segment were not enough, there
is an important restriction: ∆r. ∆n must be integer-valued and within −6
to 6. Suppose our original point is (r
0
. n
0
), and we want our destination
point to be (r
t
. n
t
). If r
t
= r
0
, the calculation is simple: set ∆r = 0,
|cn = [n
t
−n
0
[, and
∆n =

1 if n
t
n
0
;
−1 otherwise.
If r
t
= r
0
, we could have problems with approximating the results. Suppose,
for example, we want (r
t
. n
t
) = (r
0
+ 1.3. n
0
+ 1.5). If we we set |cn = 1.3
to obtain the correct r-coordinate, how should we set the slope parameters?
Ideally, we would set
∆r
∆n
=
13
15
, but the restrictions do not permit this. The
closest we could come is
4
5
.
Fixing |cn = r
t
− r
0
, then searching for a nearest slope approximation,
is not necessarily the best overall approximation. We could setup a least-
squares estimation problem, but trial and error in selecting the parameters
tends to be just as efficient. Either way, we have some work to do.
The first \vector command in Figure 62 starts at (−.65. −1), which I
calculated to be from the “top left” box to the “center” box.
76 6 GRAPHICS
width
c
box begins here →
-
height
6
?
In Figure 61 the “top left” box starts at (−1. .5), and its width is .7, so
the right edge of that box is at r = −.3. Starting at n = −.35, the vertical
position changes by moving up half of the height, so the coordinate where
the arrow begins (called its tail ) is (−.3. −.5 +
1
2
.3) = (−.3. −.35). That
accounts for the initial position given by \put(-.3,-.35). The arrow is
to be horizontal, drawn left to right, so ∆r = 1 and ∆n = 0, as specified
with \vector(1,0). Finally, we need to determine the length, specified as
{.3}. We want the coordinate of the end of the arrow (called the head) to
be flush to the left side of the “center” box. That box begins at (0. −.5),
so ∆r = 0 − (−.3) = .3. It required these computations to determine the
complete picture command: \put(-.3,-.35){\vector(1,0){.3}}.
Now consider the next \vector, which is a vertical arrow from the same
box to the circle below it. The initial position is calculated simply as the
midpoint of the bottom edge of the box: (r. n) = (r
0
+
1
2
/. n
0
), where the
box starts at (r
0
. n
0
) and / = /cio/t. In this case, (r
0
. n
0
) = (−1.. −.5) and
/ = .3, so we obtain the coordinates of the arrow’s tail: (r
t
. n
t
) = (−1 +
1
2
.3. −.5) = (−.65. −5), as specified. Since the arrow is downward, ∆r = 0
and ∆n < 0, given by \vector(0,-1). The length is determined by where
we want the arrowhead: at the top of the circle. The circles n coordinate is
−.65, which is its center. We must add the radius, which is
1
2
(.2) since .2 is
the diameter specified by \circle{.2}. Thus, the position of the arrowhead
is (r
h
. n
h
) = (r
t
. n
c
− n
t
+ :) = (−.65. −1 − (−.5) + .1) = (−.65. −.4), and
we set |cn = [∆n[ = .4, which is what is specified:
top left
`
1
?
⇐ `vector(0,-1){.4}
The next arrow is double-headed, so we use two \vector commands to
draw one arrow left to right, then an arrow at the same end points, but drawn
right to left. Further, this involves more calculations because the arrow is
not simply horizontal or vertical. We begin the same way, by computing the
coordinates of the tail and head. The left end point is at the n-coordinate of
the center of the circle, and its r-coordinate is to the right by the length of
the radius: (r
1
. n
1
) = (r
c
+ :. n
c
) = (−.65 + .1. −1) = (−.55. −1), so that is
where we \put the first arrow. The head is to be flush with the left edge of
the \fbox, and this needs some trial and error. The uncertainty is the width
of the box; we know only that the center of the box was put at (0. −1), but we
6.2 PSTricks 77
do not know the width of the box. With just a few iterations, the end point
was determined to be r = 0, so |cn = ∆r = .55. The reverse arrow begins at
(0. −1), and its slope is (−1. 0), which is why we have \vector(-1,0){.55}.
The last vector also required trial and error, due to not having the corner
of the oval coordinates. In this case, the end points were determined to be
from (1.2. −.895) to (1.6. −.5). The former was found by trial and error, but
the latter was computed by knowing that the “bottom right” box starts at
(1. −.5) and has a width of 1.2, so the midpoint of the bottom edge is at
(1.6. −.5). Now the true slope of the line we want is
.395
.4
, but the restrictions
do not allow this. The closest slope we can have is with (∆r. ∆n) = (1. 1),
which is what is specified. Given this slope, the best choice of |cn can be
found as the average of the deviations:
|cn =
1
2
(.4 + .395) = .3975.
Thus, we specify \line(1,1){.3975} to obtain the line shown in Figure 61.
There are packages to extend the picture environment, and we can plot
curves, called Bezier approximations, to a set of points. However, I shall
cover these in the next section with a powerful package called PSTricks.
Table 46 (in the Appendix, p. 117) gives the commands in the picture
environment, but here are some things to note:
• Only boxes can have inscribed text; the circles and ovals require sepa-
rate \put commands, which can take some trial and error to position.
• Some calculations and some trial and error are needed to align objects
and lines.
• Moving a portion of the picture can be tedious, requiring re-calculations
and more effort for the new positions.
• There is no direct way to control the size or style of the arrow heads,
and there is very limited control over line thicknesses.
These can make using the picture environment time consuming and rather
unpleasant. There is a better way!
6.2 PSTricks
PSTricks [15] was written by Timothy Van Zandt, and is provided free of
charge. (It is not standard with MiKTeX, but you can obtain it at CTAN [4].)
In the preamble specify \usepackage{pst-all} for the entire system. (You
can use parts, in which case you specify the parts you use instead of pst-all
— see [15] for loading individual portions.)
78 6 GRAPHICS
One thing you need to know is that not all of the pst results can be seen
with a dvi viewer. Some require converting to postscript and viewing the ps
file. This is especially true of commands that involve rotations.
PSTricks (pst, for short) is designed to overcome difficulties with using
the picture environment, some of which were listed above. Here are some of
the features of PSTricks that I shall illustrate.
• Circles and ovals, in addition to boxes, can have inscribed text.
• Lines and arrows have the same command, identifying any of a great
variety of arrowheads simply.
• Only one command is needed to put lines through a sequence of points,
and slopes need not be calculated.
• Objects can be named (as nodes) and lines and arrows can be drawn
between them by naming the tail and head, thereby eliminating the
need for calculation or trial and error.
• Arrow heads are adjustable.
• Shapes are highly variable.
• Drawing curves is simple, including plots of points that can come from
a data file, and Bezier approximations of four points are available.
Another widely distributed picture-drawing system is MetaPost [6, 7],
written by John D. Hobby, also provided free of charge. It is more difficult to
learn than PSTricks, but MetaPost is more open-ended in its design, which
makes it potentially more versatile, especially on varying the types of file
outputs (PSTricks is tied to postscript). In particular, pdflatex (not covered
here; see [4]) does not work with PSTricks, but it does with MetaPost.
There are many packages [4], typically available free of charge, that do
many of the things done by PSTricks (and some additional things). Many
of these are described in The L
A
T
E
X Companion [5].
All of the pst commands have options to override default settings for
relevant parameters. The defaults, themselves, can be set with the \psset
command: \psset{parameter = value[,. . . ]}. For example, the default unit
of measurement is 1 cm, and the default fill color is white, but we can change
them by specifying:
\psset{unit=1in,fillcolor=gray}
A fundamental command in pst is \rput, but unlike the \put command
in the picture environment, this is not the only way to put objects. The
commands, themselves, can specify where to put them. Table 19 gives
6.2 PSTricks 79
some of the common commands to draw objects and lines. For those ex-
amples, the unit of measurement was set to 1 mm. For each command,
we can specify relevant options as [parameter = value]. For example, to
produce a solid circle with radius .1 cm, centered at the origin, we write
\pscircle[fillstyle=solid](0,0){.1} (having already set fillcolor=gray).
The origin is determined by where you are when issuing a pst command;
no environment is entered. Thus, I can put that circle right here: All
commands use the linewidth parameter to control the thickness of the lines
used in the drawing, and objects that could be made solid, like boxes and
circles, use the fillstyle parameter. I shall illustrate the commands in
Table 19 first, showing the ease and versatility of PSTricks, then I shall show
some additional shapes and commands. This is meant to be an introduction,
so many features are not presented here. The User’s Guide [15] is freely
available and clearly written.
Table 19: Some Basic Drawing Commands in PSTricks
psframe(r
0
. n
0
)(r
1
. n
1
) Draws rectangle with a corner at
\psframe(0,1)(10,-2)
(r
0
. n
0
) and opposite corner at (r
1
. n
1
).
pscircle(r. n){:} Draws circle centered at (r. n) with
\pscircle(5,0){2}
radius = :.
psellipse(r. n)(:
x
. :
y
) Draws ellipse centered at (r. n) with
\psellipse(3,0)(5,2)
horizontal radius = :
x
and
vertical radius = :
y
psline{o}(r
0
. n
0
) . . . (r
n
. n
n
) Draws line or arrow, determined by o:
\psline{-}(0,0)(10,0)
- no arrow; -> forward arrow;
\psline{<->}
<-> double arrow’ <- backward arrow;
(0,0)(5,-2)(1,0)
(there are more!), along path given by
\psline{|-*}
coordinates.
(0,0)(10,-2)
pspolygon(r
0
. n
0
) . . . (r
n
. n
n
) Draws closed polygon with given
\pspolygon(0,0)
coordinates; same as \psline{-}. . . ,
(0,-3)(6,-3)
except figure is closed by drawing
line from (r
n
. n
n
) to (r
0
. n
0
).
In using these commands, we do want the \rput command in order to put
text into various objects. The idea of a box is to have some shape enclose text.
PSTricks extends the rectangle in \framebox by having a variety of shapes,
shown in Table 20. A parameter used by these commands is the distance
80 6 GRAPHICS
between the border and the text inside, called framesep=len, where the
default value of len is 3 pt. (As usual, other parameters include linewidth,
linestyle, linecolor, and fillcolor.) The pst figures are drawn after
specifying \psset{unit=1mm,fillcolor=white}.
Table 20: Boxes in PSTricks
psframebox{stuff} Draws rectangle but could have rounded corners
framebox \psframebox{framebox}
framebox \psframebox[framearc=.4]{framebox}
psshadowbox{stuff} Adds shadow to psframebox
shadow added \psshadowbox{shadow added}
psdblframebox{stuff} Draws double frame
double frame \psdblframebox{double frame}
pscirclebox{stuff} Draws circle around stuff
circle \pscirclebox[linewidth=2pt]{circle}
psovalbox{stuff} Draws oval around stuff
oval \psovalbox[linestyle=dotted]{oval}
These commands can be used in the text. For example, we obtain
this oval by writing: . . . we obtain \psovalbox{this oval}. . .
Boxes need not be enclosed (like \makebox), and they can be scaled by
specifying one of the following:
\scalebox{size}{stuff} scales stuff keeping the same aspect ratio
\scalebox{width,height}{stuff} scales the width and height individually
Here are some examples:
6.2 PSTricks 81
Halving
the
circle
\scalebox{.5}{\pscirclebox{
\begin{tabular}{c}
Halving \\ the \\ circle
\end{tabular} } }
Doubling
\scalebox{2}{\psframebox{
\textsl{Doubling} }}
Tall \scalebox{1 3}{Tall}
Wide \scalebox{3 1}{Wide}
There are times when we want to rotate stuff. Here is how:
Left
Down Right
\rotateleft{Left}\rotatedown{Down}
\rotateright{Right}
One application is given by the following:
Who is the founder of T
E
X? Who is the founder of \TeX?
Answer: Donald E. Knuth
\rotatedown{Answer: Donald E. Knuth}
So far I have described a variety of shapes, by themselves and as enclo-
sures for boxes. These can be connected by \psline, with a great variety
of styles, including variations of arrowhead shape. To avoid the tedious cal-
culations in locating the coordinates of the tail and head, the objects being
joined can be referenced by name. In PSTricks, the named objects are called
nodes. Consider the following example:
Node A
Node B Node C
The source code is shown in Figure 64. After entering the centering envi-
ronment and setting the default units of measurement, the \rput command
puts a node, with the \rnode command. The name is set to A, and the text
Node A is put there (with no frame). The syntax for \rnode is:
\rnode{name}{stuff }
The next two commands put nodes named B and C, each enclosed with a
frame. The \ncline command has the same arrow options as \psline, but
with the following syntax:
\ncline{o}{name of node A}{name of node B}
82 6 GRAPHICS
The first \ncline in Figure 64 draws a plain line from node A to node B.
The [nodesepA=3pt] option gives 3 pt separation between the end of the
line and node A. Otherwise, the line would touch Node A text, which is not
what we want. The separation is exaggerated to 5 pt in the arrow from
node C to node A. The default value is nodesep=0pt, which is what we
want when the nodes are enclosed boxes, like B and C. In general, node
separation can be specified for either end point, or for both end points, by
specifying nodesepA=n, nodesepB=n, or nodesep=n, respectively. (nodesepA
and nodesepB are keywords and have nothing to do with the names we assign
to our nodes.)
\begin{center}
\psset{unit=1cm}
\rput( 0, 0){\rnode{A}{Node A}}
\rput(-2,-1){\rnode{B}{\psframebox{Node B}}}
\rput( 2,-1){\rnode{C}{\psovalbox{Node C}}}
\ncline[nodesepA=3pt]{A}{B}
\ncline[nodesepA=5pt]{<-}{A}{C}
\ncline{<->}{B}{C}
\end{center} \vspace{.5in}
Figure 64: PSTricks Source for Connecting Nodes
Figure 66 shows a graph that could represent any number of things. Its
source, using PSTricks, is shown in figure 65. (Try adding one line at a time
and observe each effect.)
\begin{center} \psset{unit=1cm}
% Nodes
\cnodeput(-2, 0){1}{1} \cnodeput(0, 0){2}{2}
\cnodeput( 2, 1){3}{3} \cnodeput[doubleline=true]( 2,-1){4}{4}
\pnode(-3, 0){1tail} \ncline{->}{1tail}{1} % tailess arc into (1)
% Arcs (with labels)
\ncline{->}{1}{2} \aput{:U}{1/2} % \aput puts label above arc
\ncline{->}{2}{3} \aput{:U}{2/3}
\ncline{->}{2}{4} \bput{:U}{2/4} % \bput puts label below arc
\ncarc{->}{3}{4} \Aput{\small 3$\rightarrow$4} % \Aput keeps
\ncarc{->}{4}{3} \Aput{\small 4$\rightarrow$3} % label horizontal
\ncloop[angleB=180,loopsize=.5,arm=.2,linearc=.2]{->}{3}{3}
\Bput[5pt]{loop} % \Bput keeps label horizontal and 5pt is
% the space added between label and arc
\ncloop[angleA=180,loopsize=.5,arm=.2,linearc=.2]{<-}{4}{4}
\Bput[5pt]{loop}
\end{center} \vspace{1cm}
Figure 65: Graph Source (Result in Figure 66)
6.2 PSTricks 83
1 2
3
4
1/2 2/3 2/4 3→4 4→3 loop loop
Figure 66: Graph Result (Source in Figure 65)
Now I describe curves that go through, perhaps approximately, given
points. The examples that follow use the following pst settings:
\psset{unit=.5cm,showpoints=true}
(The showpoints=true setting is what causes the points to be included in
the picture you see.)
We begin with the parabola, whose command syntax is:
\parabola{o}(r
0
. n
0
)(r
1
. n
1
),
where (r
0
. n
0
) is one point on the parabola, and (r
1
. n
1
) is the (unique) point
having dn´dr = 0. \parabola* specifies filling the parabola. For example,
\psgrid[subgriddiv=1,griddots=10,gridlabels=7pt](-1,0)(4,4)
\parabola{<->}(4,3)(2,0)
\parabola*[fillcolor=black,showpoints=false](1,1)(2,3)

Question: What is the pst command to draw the parabola given by
n = or
2
+ /r + c, where o = 0?
Answer: `parabola(0. c)(−
b
2a
. −
b
2
4a
+ c)
The following shows two commands: pscurve and psccurve, the latter
being a closed curve that joins the last point with the first.
\pscurve{(->}(0,0)(1,1)(1,-1)(-1,1)(-1,-1)
\psccurve(0,0)(1,1)(1,-1)(-1,1)(-1,-1)
84 6 GRAPHICS
The Bezier curve joins two end points and comes as close as possible to
two intermediate points. The command syntax is:
\psbezier[parameters]{o}(r
0
. n
0
)(r
1
. n
1
)(r
2
. n
2
)(r
3
. n
3
)
\psbezier(0,0)(1,3)(2,1)(3,4)
We can read data from a file, perhaps produced by mathematical soft-
ware like gnuplot
c
, Octave
c
, Maple
c
, Mathematica
c
, MATLAB
c
, and
S-PLUS
c
. The data file just needs pairs of coordinates, which can be sepa-
rated by a comma or just blank and can have parenthesis, braces, or nothing
around each pair. The following histogram was plotted by the source code in
Figure 67, which I shall explain. The data file had n = number of students
with test score = r+50. (The offset of 50 was used in establishing the origin
in the plot.)
50 60 70 80 90 100
0
5
10
15
score
F
D
C
B
A
\psset{unit=2mm, showpoints=false}
\fileplot[plotstyle=dots]{mydata.dat}
\psaxes[Ox=50,Oy=0,Dx=10,Dy=5,dx=10,dy=5,ticks=y]{<->}(60,17)
\rput[r](60,-2){\large score}
\psline(1,0)(1,6)(10,6)(10,0) \rput( 5, 8){\textsf{F}}
\psline(11,0)(11,2)(19,2)(19,0) \rput(14, 4){\textsf{D}}
\psline(20,0)(20,11)(29,11)(29,0) \rput(25,13){\textsf{C}}
\psline(30,0)(30,8)(39,8)(39,0) \rput(35,10){\textsf{B}}
\psline(40,0)(40,5)(50,5)(50,0) \rput(45, 7){\textsf{A}}
Figure 67: Source Code for Drawing Histogram of Test Scores
After setting the units of measurement to 2 mm, the data file is read and
its points plotted with the \fileplot command. (Setting showpoints=false
6.2 PSTricks 85
suppresses plotting the points in the \psline commands.) The data file is
plain text and has the following entries:
% This is mydata.dat
5 2 9 4 % F = [0,65)
15 2 18 1 % D = [65,70)
22 6 27 4 % C = [70,80)
30 2 31 1 35 4 39 1 % B = [80,90)
40 2 45 2 50 1 % A = [90,100]
The plot, itself, is just the points, specified by plotstyle=dots. There
are other plot styles, such as plotstyle=line, and there are 11 dot styles.
Here is one of the alternatives:
\fileplot[dotstyle=+,plotstyle=dots]{mydata.dat} ⇒
Next, axes are superimposed with the \psaxes command:
\psaxes[params]{o}(r
0
. n
0
)(r
1
. n
1
)(r
2
. n
2
)
where (r
0
. n
0
) is the origin, (r
1
. n
1
) is the Southeast corner, and (r
2
. n
2
)
is the Northwest corner. As in \psline, if (r
0
. n
0
) is absent, the origin is
assumed to be at (0. 0). If (r
1
. n
1
) is absent, it is assumed to be equal to the
origin. Here are some examples:
\psaxes[unit=.5cm]{->}(4,0)(8,3) ⇒
0 1 2 3
0
1
2
\psaxes[unit=.5cm]{->}(0,0)(-1.1,0)(2.1,2.1) ⇒
0 1 2 −1
1
2
Note that ticks are uniformly spaced on each axes. This is suppressed for
the r-axis in Figure 67 by specifying the option, ticks=y. The other pa-
rameter settings are described in Table 21. (The default values, dx=dy=0,
cause the spacing to be equal (approximately) by using Dx÷\psxunit and
Dy÷\psyunit, respectively.)
86 6 GRAPHICS
Table 21: Parameters for \psaxes
Horizontal Vertical Default Meaning
Ox=n Oy=n 0 Label at origin
Dx=n Dy=n 1 Label increment
dx=n dy=n 0 Label spacing
The next command, rput[r](60,-2){\large score} puts “score” in
large font, flush right (indicated by [r]) at the coordinates (60. −2). Thus,
when I superimpose the commands \fileplot, \psaxes and \rput, we ob-
tain the data plot. The remaining commands draw the histogram boxes and
put the letter grade above each box in sans serif font. Leaving off the “score,”
Figure 68 shows the sequence of how each \psline and \rput adds to the
picture. To fit the picture and the code next to it, this is scaled (simply, by
specifying \psset{unit=1mm}):
We shall stop here, but this does not exhaust the PSTricks commands.
See [15] for lots more, including many examples.
6.3 Importing pictures
The way to import a picture into L
A
T
E
X is to convert it to encapsulated
postscript (eps). An exceptionally clear description of this, including his-
torical context, is given by Keith Reckdahl [12]. (He also goes deeper
into customizing placements of pictures in figures.) Many systems that
let us draw figures, and those that plot mathematical functions or data,
have an option to export an eps file. (If you can get a ps file, you could
use \psfig, or there is a unix conversion utility, \ps2epsi.) On unix,
xfig is an excellent system to draw figures, and the export options in-
clude the eps file format. A basic plotting system for functions and data,
for both unix and DOS that produces eps files, is gnuplot. This is avail-
able free of charge at FTP://ftp.dartmouth.edu/pub/gnuplot/. Octave ex-
tends the capabilities of gnuplot and is also available free of charge, at
http://www.che.wisc.edu/octave/. There are also commercial systems, like
Maple
c
, Mathematica
c
, MATLAB
c
, and S-PLUS
c
, which can produce
eps files of plots.
Another way to obtain an eps file is with conversion. The unix systems
xv and Image Magick can do this for a large variety of graphic file formats,
including bitmap (xbm), gif and jpeg files. There are free conversion systems
on MS Windows, notably jpeg2ps, which converts jpeg files to eps, and
emftoeps, which converts Windows Metafiles (wmf) to eps.
Once the file is in eps format, we can import it using the Graphics Bun-
dle [3], written by David P. Carlisle, provided free of charge. It comes with
MiKTeX and basic unix installations. There are two packages that provide
6.3 Importing pictures 87
essentially the same capabilities but with different syntax. One is called
graphics, the other is graphicx. Here I use graphicx, as specified in the
preamble by \usepackage{graphicx}.
\fileplot[plotstyle=dots]{mydata.dat}
50 60 70 80 90 100
0
5
10
15
\psaxes[Ox=50,Oy=0,Dx=10,Dy=5,
dx=10,dy=5,ticks=y]{<->}(60,17)
50 60 70 80 90 100
0
5
10
15
F
\psline(1,0)(1,6)(10,6)(10,0)
\rput(5,8){\textsf{F}}
50 60 70 80 90 100
0
5
10
15
50 60 70 80 90 100
0
5
10
15
F
D
\psline(11,0)(11,2)(19,2)(19,0)
\rput(14,4){\textsf{D}}
50 60 70 80 90 100
0
5
10
15
50 60 70 80 90 100
0
5
10
15
F
D
C
\psline(20,0)(20,11)(29,11)(29,0)
\rput(25,13){\textsf{C}}
50 60 70 80 90 100
0
5
10
15
50 60 70 80 90 100
0
5
10
15
F
D
C
B
\psline(30,0)(30,8)(39,8)(39,0)
\rput(35,10){\textsf{B}}
50 60 70 80 90 100
0
5
10
15
50 60 70 80 90 100
0
5
10
15
F
D
C
B
A
\psline(40,0)(40,5)(50,5)(50,0)
\rput(45,7){\textsf{A}}
Figure 68: Sequence of PSTricks Commands to Draw Histogram
To include an eps file, simply specify \includegraphcs[options]{filename}.
For example, Figure 69 shows a figure that was imported with the following
statement:
88 6 GRAPHICS
\begin{center}\includegraphics[scale=.5]{sin.eps}\end{center}
−8 −6 −4 −2 0 2 4 6 8
−1
−0.8
−0.6
−0.4
−0.2
0
0.2
0.4
0.6
0.8
1
Figure 69: Applying \includegraphics to Import an eps File
In this case I specified the option, scale=.5, which prints the figure
half the size it was produced (in this case by MATLAB, by specifying
print sin -deps after plotting the sin function over the indicated grid).
Figure 70 shows the same eps file, but with the width and height set as
follows:
\begin{center} \includegraphics[width=2in,height=1in]{sin.eps}
\end{center}
−8 −6 −4 −2 0 2 4 6 8
−1
−0.8
−0.6
−0.4
−0.2
0
0.2
0.4
0.6
0.8
1
Figure 70: Specifying Dimensions in \includegraphics
For a very large picture, we might want to specify width=\textwidth,height=!,
and let it fill the entire width of the page. The height specification (!) says
to maintain the aspect ratio.
If you find yourself importing eps files but would like to make some
changes in L
A
T
E
X, read about the PSfrag package, by Michael C. Grant and
6.3 Importing pictures 89
David Carlisle, which comes with a basic installation (including MiKTeX),
whose documentation is at CTAN [4]. It has two basic operations: (1) edit
some string or position in the figure (i.e., the eps file), and (2) translate L
A
T
E
X
commands that you put in the figure in the first place. The documentation
gives examples, with eps files produced by MATLAB and xfig.
Importing graphics is only one of the functions of graphicx. It can also
perform scaling, rotation, and sizing of an arbitrary box. The box could
contain text, pictures, or almost any stuff. Here are examples:
Double your fun \scalebox{2}{Double your fun}
Open wide \resizebox{1in}{!}{\fbox{Open wide}}
R e fl e c t o n t h i s \reflectbox{Reflect on this}
L
a
n
d
s
c
a
p
e
\rotatebox[origin=c]{90}{Landscape}
W
a
s
P
y
t
h
a
g
o
r
a
s
a
s
q
u
a
r
e
?
\rotatebox[origin=rt]{45}
{\psframebox{
\begin{tabular}{c}
Was\\Pythagoras\\a square?
\end{tabular}
} }
These operations are available because the programs that perform them
are used in the \includegraphics command. Although it is feasible to
perform the operation after importing a graphic, it is more efficient to specify
that option in the \includegraphics. Here are some examples:
90 6 GRAPHICS

\includegraphics
{protractor.eps}

\includegraphics[width=.25\textwidth,
height=!]{protractor.eps}











\includegraphics[height=.5in,width=!,
angle=90,origin=c]{protractor.eps}
Exercises. Submit a printed copy of the L
A
T
E
X source (tex file) and printed
copy of the associated postscript result (ps file). Be sure your name is on
each.
1. Use the picture environment to draw the smiley face on page 73.
2. Draw the following graph with the picture environment, where \thicklines
is specified and \unitlength = 1mm
m
·
1
m
·
2
m
·
3
m
·
4
E
c
'

©
3. Use PSTricks to draw Figure 3 (p. 4).
6.3 Importing pictures 91
4. Use PSTricks or the picture environment to draw the following.
rhombus
5. Use PSTricks or the picture environment to draw the following.
α
β
6. Make a figure in some system that lets you save it as an eps file (or
use some conversion program). Then, include it in your document.
7. Use whatever means you prefer (or that your instructor requires) to
include each of the following figures in your document. (They were
drawn here with PSTricks, but this section did not describe all that is
needed, so you must obtain the PSTricks User’s Guide [15].)
(a) Graphic view of Pythagorean Theorem:
hypotenuse leg 1 leg 2
square of
leg 1
square
of
leg 2
square of
hypotenuse
92 6 GRAPHICS
(b) Network with arc data:
1
2
3
4
5
(25, 30) (35, 50) (45, 10) (15, 40) (15, 30) (25, 20) (35, 50)
(45, 60)
(c) The sin function:
y = sin

π
2
x

x
(d) Bernoulli family tree:
Nikolaus
(1623–1708)
Jacob I
(1654–1705)
Nikolaus
(1662–1716)
Nikolaus I
(1687–1759)
Jahann I
(1667–1748)
Nikolaus II
(1695–1726)
Daniel
(1700–1782)
Johann II
(1710–1790)
Johann III
(1746–1807)
Jocob II
(1759–1789)
93
7 Making Special Parts
7.1 Cover Page
The easiest way to make a cover page is with the \maketitle command.
This is usually done just following \begin{document}. The necessary pa-
rameters are \author and \title, which can be defined anyplace before
the \maketitle. Typically these are put into the preamble, or right after
\begin{document}, followed immediately by \maketitle; it depends upon
your management style.
If you see a page number on your cover page, suppress this by adding
\thispagestyle{empty}, right after \maketitle. You might also need to
initialize the page counter with \setcounter{page}{0}.
Multiple authors are separated by \and, such as in the example shown
in Figures 71 and 72. (The jagged edges in Figure 72 mean that there is
more space between the title and the top of the paper.)
Specifying \date is optional (\maketitle puts in the current date if the
date is not defined). The cover page is by itself and is not numbered.
\title{The \LaTeX\ Companion}
\author{Michel Goosens \and Frank Mittelbach \and Alexander Samarin}
\date{1994}
\maketitle
Figure 71: Title Page Source (Result in Figure 72)
The L
A
T
E
X Companion
Michel Goosens Frank Mittelbach Alexander Samarin
1994
Figure 72: Title Page Result (Source in Figure 71)
Since articles often have this information on the first page of the article
(rather than a separate page), titlepage must be specified as an option in
the \documentclass command. For example, the following does this while
specifying 12pt font as another option:
94 7 MAKING SPECIAL PARTS
\documentclass[12pt,titlepage]{article}
Addresses, affiliations, and other information about each author can be
added, using \\ to create new lines. For example, Figure 73 shows how the
authors appear when the \author definition in Figure 71 is changed to the
following:
\author{Michel Goosens \\ Geneva, Switzerland
\and Frank Mittelbach \\ Mainz, Germany
\and Alexander Samarin \\ Geneva Switzerland}
As illustrated in Figure 73, \maketitle puts the third author on a sep-
arate line. This is because the added width of author information makes it
too long to fit on one line. All three authors would be put on separate lines
if the address information were extended further, or if the names were very
long.
The L
A
T
E
X Companion
Michel Goosens Frank Mittelbach
Geneva, Switzerland Mainz, Germany
Alexander Samarin
Geneva, Switzerland
1994
Figure 73: Adding Addresses to Authors
There are times when we want to acknowledge support for one or more of
the authors. The \thanks command does this by creating a footnote, using
different footnote marks for each one. Figures 74 and 75 illustrate this along
with some variation in the date.
7.1 Cover Page 95
\title{Pieces of $\pi$\thanks{Renamed.}}
\index{\texttt{$\backslash$thanks}} \index{footnote}
\author{Archimedes\thanks{Supported by the army.}\\ Syracuse, Sicily
\and Pythagoras \\ Samos, Ionia }
\date{210 {\sc bc} (revision of earlier version, 510 {\sc bc})}
Figure 74: Footnotes in the Cover Page Source (Result in Figure 75)
Pieces of π
1
Archimedes
2
Pythagoras
Syracuse, Sicily Samos, Ionia
210 bc (revision of earlier version, 510 bc)
1
Renamed.
2
Supported by the army.
Figure 75: Footnotes in the Cover Page Result (Source in Figure 74)
Figure 76 shows how to use only one footnote for authors having the same
affiliation. This uses the \footnotemark command, which puts the footnote
mark without any new text. Prior to that, the value of the footnote counter
is set back to the first footnote mark.
\title{Doing \LaTeX}
\author{I.M. Rich \thanks{Smart University}
\and U.R. Grand \thanks{Street-Smart}
\and I.C. All \addtocounter{footnote}{-2} \footnotemark
}
Figure 76: Authors with same footnote (Result in Figure 77)
96 7 MAKING SPECIAL PARTS
Doing L
A
T
E
X
I.M. Rich
1
U.R. Grand
2
I.C. All
1
1
Smart University
2
Street-Smart
Figure 77: Authors with same footnote (Source in Figure 76)
7.2 Abstract
The abstract environment is in all document styles, except article. To have
it, specify titlepage as an option in \documentclass (even if you do not in-
tend to use \maketitle). This environment is defined to produce an abstract
on a separate page (placed wherever you put the environment specification),
with the header: Abstract, in boldface and centered. The abstract, itself,
is one paragraph and is printed without indentation. Figures 78 and 79 il-
lustrate this. (Like the cover page, the abstract is placed far from the top of
the paper, which is not shown in Figure 79.)
\begin{abstract}
This shows that the ratio of the circumference to the diameter
of any circle is the same constant value, denoted $\pi$.
We further prove that this constant is bounded by
$\frac{223}{71} < \pi < \frac{22}{7}$.
\end{abstract}
Figure 78: Making an Abstract Source (Result in Figure 79)
7.3 Other Front Matter 97
Abstract
This shows that the ratio of the circumference to the diameter
of any circle is the same constant value, denoted π. We further
prove that this constant is bounded by
223
71
< π <
22
7
.
Figure 79: Making an Abstract Result (Source in Figure 78)
7.3 Other Front Matter
The \tableofcontents command makes a table of contents; it is placed
wherever you put the command, which should be right after the cover page.
Then, you can include lists of figures and tables with the \listoffigures
and \listoftables commands, respectively.
The table of contents generally includes numbered parts, like sections and
subsections. To include other front matter, L
A
T
E
X provides the \addcontentsline
command. For example, the table of contents in this document was obtained
with the specifications given in Figure 80.
\newpage \pagenumbering{roman} \pagestyle{myheadings}
\tableofcontents \newpage
\addcontentsline{toc}{section}{List of Figures}
\listoffigures \newpage
\addcontentsline{toc}{section}{List of Tables}
\listoftables \newpage
Figure 80: Some Front Matter Specifications for This Document
The \pagenumber specification causes the page numbers for the front
matter to be put into Roman numerals. That is why you see the Table
of Contents on page i (first numbered page, just after the cover). Then, I
declare \listoffigures, which is on page v, followed by the list of tables.
Each of these are put on a new page. Just above each declaration, I use the
\addcontentsline to add it to the table of contents, indicated by the toc
specification. The section parameter tells the latex program to format it
like a section — flush left.
The page numbering is reset when we finish the front matter by specifying
\newpage \pagenumbering{arabic} \pagestyle{headings}
This switches to the Arabic numerals and initializes the page counter.
98 7 MAKING SPECIAL PARTS
The same format as the abstract can be used for other front matter that
we want to format the same way. The only change we require is another
header name. This is done by re-defining the \abstractname parameter
used by the abstract environment. The \renewcommand enables us to do
this. §?? has more to say about using this command to customize many
things. For now, consider the following example that illustrates how to have
an Acknowledgments page:
\renewcommand\abstractname{Acknowledgments}
\begin{abstract}
I thank my family and friends for all of their support.
I also thank the contributors to the Comprehensive \TeX\
Archive Network (CTAN).
\end{abstract}
Alternatively, we might want something to look like a section (and au-
tomatically added to the table of contents), but we do not want it to have
a section number. This is achieved by the \section* command, where the
* suppresses the numbering. For example, \section*{Preface} puts “Pref-
ace” in the same style as any section, but with no number (and the section
counter remains unchanged).
7.4 Back Matter
After the main part of the document is finished, we put the bibliogra-
phy (see §3 and §8.6). We might first want to have appendices that fol-
low the main text. This could be done with the appendix environment:
\begin{appendix} . . . \end{appendix}.
The last portion in the back of any book is its index. This could also
be desirable in a long report. To make an index, we have three things to put
into our source file:
1. Put \usepackage{makeidx} in the preamble.
2. Put \makeindex at the end of the preamble.
3. Put \printindex just before \end{document}.
After a successful compilation, with all references resolved, enter at the
command line:
makeindex myfile
Then, compile again. This is analogous to the use of bibtex (p. 28), and is
illustrated in Figure 81.
To have entries in the index, you use \index{entry}. For example, the
index in this book contains several occurrences of ‘flushleft’. In the text,
7.5 Footnotes 99
at each occurrence, I specify \index{flushleft}. To put the entry as a
subordinate, use !. For example, \index{package!makeidx} puts the entry
‘makeidx’ under the entry ’package’.
myfile.tex myfile.dvi myfile.ps
create/edit view/print print/post
compile
latex
convert
dvips
makeindex
myfile.pdf dvipdf
Figure 81: Adding makeindex to the Command Sequence
There are packages to make other back matter: acronym makes a list of
acronyms, nomencl makes a list of nomenclature, and gloss makes a glossary.
7.5 Footnotes
The \thanks command is one way to have a footnote on the cover page,
and was shown on page 95. More generally, the \footnote command can be
used anywhere. Figures 82 and 83 illustrate.
Here is my first footnote\footnote{first};
here is my second\footnote{second}.
Figure 82: Setting a Footnote Source (Result in Figure 83)
Here is my first footnote
1
; here is my second
2
.
1
first
2
second
Figure 83: Setting a Footnote Result (Source in Figure 82)
\thefootnote gives the counter for footnotes, and you can change from
numbers to letters by \renewcommand{\thefootnote}{\alph{footnote}}.
You can also change to common footnote symbols by specifying
\renewcommand{\thefootnote}{\fnsymbol{footnote}}, as shown in fig-
ures 84 and 85.
The distance between the line that underlies part of the last line of text
to the footnote is \footnotesep, which can be controlled by \setlength.
100 8 TAKING CONTROL
\renewcommand{\thefootnote}{\fnsymbol{footnote}}
Here is my first footnote\footnote{first};
here is my second\footnote{second}.
Figure 84: Setting a Footnote Source (Result in Figure 85)
Here is my first footnote

; here is my second

.

first

second
Figure 85: Setting a Footnote Result (Source in Figure 84)
Exercises. Submit a printed copy of both the L
A
T
E
X source (tex file) and
the associated postscript result (ps file). Be sure your name is on each.
1. Write an article with a title page and abstract. Make the main body
have at least three sections: Introduction, Main Results, and Conclu-
sions.
2. Extend exercise 1 to have acknowledgments and references (using BibT
E
X).
3. Combine exercises 1 and 2 and add a table of contents showing not only
all sections and subsections, but also the abstract, acknowledgments
and references.
8 Taking Control
This section introduces you to fundamentals of customizing your document.
It is still in the context of an introduction, choosing only a few of the
things you can change. A key to these changes are the \newcommand and
\renewcommand commands, which enable you to define your own commands
and change parameter values of existing commands.
8.1 Your Own Abbreviations and Commands
The command that gives us the ability to make our own has the following
form: \newcommand{\name}[n]{whatever}, where n is the number of argu-
ments, and whatever is whatever you want the command to do. Here are
two examples simply to abbreviate commands with long names:
\newcommand{\ul}{\underline}
\newcommand{\mc}{\multicolumn}
8.1 Your Own Abbreviations and Commands 101
The first lets us write \ul{something} to underline something. The second
lets us write \mc{3}{c}{stuff} to enter a multicolumn, in either a tabular
or an array environment, spanning 3 columns and centered.
The latex compiler will not let you use a name that is already being used.
For example, if you specify \newcommand{\fbox}..., you will get a fatal error
message since there is already a \fbox command.
A related use is when the command requires some lines of code. Consider
the following example:
\newcommand{\Box}{\mbox{\begin{picture}(0,0)
\put( 2,0){\framebox(7,7)}
\end{picture} }}
(\mbox is used to ensure text mode). Now \Box ⇒ and, having defined
the \Box command, we can use it in other new commands. For example,
in the preamble I specified: \newcommand{\chkbox}{$\Box^\surd\;$} Then,
\chkbox ⇒

Some commands are specifically for math mode, but we want them to
work in any mode. This is achieved by the command: \ensuremath{math
stuff }. For example, consider \newcommand{\Gs}[1]{G_{#1}}. If we are
already in math mode, \Gs{i+j} is replaced by G_{i+j} to produce G
i+j
;
otherwise, if \Gs{i+j} is specified in text mode, it is replaced by $G_{i+j}$
to put it into math mode first. Thus, we can specify \Gs{subscript}, no
matter which mode we are in, and obtain the correct result.
Another reason to have our own commands is for consistency, particularly
of notation. Suppose we have a key term, like the null space of a matrix.
Some authors write ^(¹), some write nul ¹, and there are still more symbols
people use. We can choose one and define
\newcommand{\nul}{\ensuremath{\mathcal{N}}}.
Then, we can write \nul(A) to obtain ^(A) (and we can be in text or math
mode when we write this). Some publishers have their own notation, so we
must be careful not to override them with ours. A way to do this is to choose
a different name, like mynul, then add to the preamble:
\newcommand{\usenul}{\mynul}
and specify \usenul in the document. If you need to use the publisher’s,
simply change the one line to:
\newcommand{\usenul}{\nul}
(where the publisher’s command name is \nul).
The preamble can become very long as we add our commands, so it
is useful to put them in a separate file, say mydefs.tex (note the .tex
suffix). Then, we use the \input command to have the latex compiler read
102 8 TAKING CONTROL
it wherever it is placed. In particular, the preamble of this document contains
the command:
\input{mydefs}
(The suffix .tex is assumed.) Different source files could simply input this
same file, so duplication of work is avoided.
8.2 Your Own Names, Titles and Numbers
There are times when we prefer some name other than the default. Table 22
shows the common names we might want to change. For example, in this
document the “Table of Contents” was obtained by specifying the following
in the preamble:
\renewcommand{\contentsname}{Table of Contents}.
Table 22: Intrinsic Name Parameters
What it is How it is called (keyword)
Abstract \abstractname
Appendix \appendixname
Chapter \chaptername
Contents \contentsname
Index \indexname
List of Figures \listfigurename
List of Tables \listtablename
Part \partname
References \refname for article style
\bibname for book and report styles
You might want to change the numbering of some intrinsic counter. We
saw an example of this in changing the counters for enumeration lists (p. 43).
The general form is
\renewcommand{\thecounter}{something}
Another example is to change section numbering in a report document
style. The first level of division is assumed to be a chapter, so the numbering
will be chapter.section[.subsection]. . . If you have no chapters, it will number
the first section as 0.1. Making the first level division chapters will overcome
the numbering problem, but the format of chapters is different, similar to
a book. Making the document class an article will also solve the problem,
since the section is the first level division. This might not be appropriate due
to other considerations, such as entering the document into a database using
8.3 Your Own Environments 103
BibT
E
X, where you want it to be counted as a report, not as an article.
The way to do this is as follows:
\makeatletter
\renewcommand\thesection{\@arabic\c@section}
\makeatother
The preceding command, \makeatletter, is to make the @ character a letter.
The succeeding command, \makeatother, restores @ to its special meaning
(\@ is for certain spacing, equal to about 2 spaces).
8.3 Your Own Environments
The \newenvironment command enables us to define our own environments,
and the \renewenvironment command enables us to revise an existing en-
vironment. They have the same syntax:
newenvironment{name}[n]{begin}{end}
renewenvironment{name}[n]{begin}{end}
where name is the name of the environment, n = number of arguments (omit
[0] for n = 0), begin is what is executed upon entering the environment,
and end is what is executed upon leaving the environment. For example,
the following creates a proof environment:
\newenvironment{proof}
{\begin{flushleft} \begin{description}
\item \textit{\textbf{Proof:}}~ } % begin proof
{\hfill\rule{2.1mm}{2.1mm}
\end{description}\end{flushleft} } % end proof
Then,
\begin{proof}
First, suppose \dots \linebreak
Thus, the theorem follows.
\end{proof}
produces:
Proof: First, suppose . . .
Thus, the theorem follows.
8.4 Your Own Margins and Spacing
The default margins and spacing are set with purposeful values, and you
will usually not need to change them. When you do, however, they can
be changed by setting certain parameters in the preamble. The margins
104 8 TAKING CONTROL
of the document are controlled by the parameters shown in Figure 86 and
described in Table 23. (See Table 25 for conversion factors; in particular,
1 pt = 72.27 in.)
For example, if we are using 8
1
2
11 paper, the current settings (shown
in Table 23) break down the horizontal parts as follows:
6.14.295 pt
8.5 in
72.27 pt
1 in
39 pt
54 in
113.025 pt
1.564 in
390 pt
5.396 in
begin body stuff end body stuff
We can increase the text width by setting \textwidth=length in the
preamble. For example, \textwidth=6in increases the text width to 6 inches.
The body expands to the right unless we also change \oddsidemargin.
Margin settings can be negative; for example, we raise the body 1 inch by
specifying \topmargin=-1in in the preamble. This might be accompanied
by increasing the text length. The geometry package provides easy specifica-
tions for page layout.
The \hspace* and \vspace* commands provide a great deal of control
over horizontal and vertical spacing, respectively. We might want some global
settings to make repeated use of these unnecessary. Table 24 lists some you
can set with the \setlength command, showing their default values (used
in this document).
Table 23: Margin Parameters
Current
Parameter Setting

Meaning
\footskip 30.0pt space between bottom of body and top of
footer
\headsep 25.0pt space between bottom of header and top of
body
\headheight 12.0pt height of header
\hoffset 0.0pt horizontal offset to add to indentation of body
\oddsidemargin 29.0pt extra space added at left (applies only to odd
numbered pages if the style is two-sided, in
which case there is also an `evensidemargin
parameter)
\paperheight 794.96999pt height of the paper
\paperwidth 614.295pt width of the paper
\textheight 585.38744pt height of the body
\textwidth 360.0pt width of the body
\topmargin 21.0pt space added before the top of the header
\voffset 0.0pt vertical offset to add to indentation of body

Printed using \theparameter.
8.4 Your Own Margins and Spacing 105
\pagewidth
\pageheight
1 in +
\hoffset
1 in +
\voffset
header
body
footer
\textwidth
\textheight
\headheight
\footskip
\headsep
\topmargin
\oddsidemargin
Figure 86: Document Margins
106 8 TAKING CONTROL
Table 24: Spacing Parameters
Parameter Meaning
\itemsep space added to \parsep between items in a list.
\parindent indentation at beginning of paragraph.
\parsep space between paragraphs in the same item of a list.
\parskip space between paragraphs.
In the case of list parameters, they must be set within the list environ-
ment. For \itemsep this can occur just after the \begin statement. (De-
faults are restored after leaving.) For example, the lists in §2.2 (p. 12) are
spaced by default values. Here is what happens when we change \itemsep:
• The default value of \itemsep is 4.5pt plus 2.0pt minus 1.0pt, and I
have saved it by: \setlength{\mylength}{\itemsep}.
• See the above spacing between items. What you see next is with
\setlength{\itemsep}{0pt}.
• What you see next is with \setlength{\itemsep}{10pt}.
• Next is back to normal by \setlength{\itemsep}{\mylength}.
• We are back to normal with \itemsep = 4.5pt plus 2.0pt minus 1.0pt.
Figure 87 shows two enumerate lists, varying by the value of \itemsep.
1. This list specified
\setlength{\itemsep}{0pt}
right after \begin{itemize}
2. This item is fairly close to the
first one.
1. This list specified
\setlength{\itemsep}{12pt}
right after \begin{itemize}
2. This item is farther from the
first one.
Figure 87: Varying \itemsep to control item spacing in a list
The left margin parameter is \leftmargini for a level-1 list (note the i
at the end). This must be set before the \begin command. Figure 88 gives
an example.
8.4 Your Own Margins and Spacing 107
{\setlength{\leftmargini}{2in}
\begin{itemize} \setlength{\itemsep}{0in}
\item Item 1 of list with 2in left margin
\item Item 2 of list with 2in left margin
\end{itemize}
}
Figure 88: Changing the Left Margin of a List (Result in Figure 89)
• Item 1 of list with 2in left margin
• Item 2 of list with 2in left margin
Figure 89: Changing the Left Margin of a List (Source in Figure 88)
Figures 90 and 91 show the presentation of an array with a p-column to
put horizontal space between the other two columns. Note how congested it
is, so we want to increase its vertical spacing.
\[ \begin{array}{lp{.3in}l}
\,B\,x_B = b_N + \frac{1}{2}\theta\delta b_N
&& \pi_N B = c_B + \frac{1}{2}\theta\delta c_B \\
B^*x_B > b_B + \frac{1}{2}\theta\delta b_B
&& \pi_N N < c_N + \frac{1}{2}\theta\delta c_N
\end{array}
\]
Figure 90: Array with Fixed Width Column Source (Result in Figure 91)
1 r
B
= /
N
+
1
2
θδ/
N
π
N
1 = c
B
+
1
2
θδc
B
1

r
B
/
B
+
1
2
θδ/
B
π
N
· < c
N
+
1
2
θδc
N
Figure 91: Array with Fixed Width Column Result (Source in Figure 90)
Here is 1.3 line spacing: \renewcommand{\arraystretch}{1.3}
1 r
B
= /
N
+
1
2
θδ/
N
π
N
1 = c
B
+
1
2
θδc
B
1

r
B
/
B
+
1
2
θδ/
B
π
N
· < c
N
+
1
2
θδc
N
Here is 1.6 line spacing: \renewcommand{\arraystretch}{1.6}
1 r
B
= /
N
+
1
2
θδ/
N
π
N
1 = c
B
+
1
2
θδc
B
1

r
B
/
B
+
1
2
θδ/
B
π
N
· < c
N
+
1
2
θδc
N
108 8 TAKING CONTROL
Back to default: \renewcommand{\arraystretch}{1}
1 r
B
= /
N
+
1
2
θδ/
N
π
N
1 = c
B
+
1
2
θδc
B
1

r
B
/
B
+
1
2
θδ/
B
π
N
· < c
N
+
1
2
θδc
N
Line spacing is controlled by \baselinestretch , but it is easier to use
the setspace package. You will have control of the whole document, or just
some lines whose space you want to control.
8.5 Your Own Output Control
You can write conditionals to do one thing or another with the ifthen package.
The syntax is
\ifthenelse{condition}{action if true}{action if false}
For example, suppose your preamble contains
\newcommand{ \printsol}[2]{ \ifthenelse{\equal{#1}{1}}%
{\textsf{\textbf{Solution. }#2}}{} }
This will test the condition if your first argument equals 1. If so, it will print
Solution. followed by your second argument; if not, it will do nothing.
This is illustrated in figures 92 and 93.
How much is 2+2?
\printsol{\sol}{4}
Figure 92: \ifthenelse Source (Results in Figure 93)
`def`sol{0}
.
.
.
How much is 2+2?
`def`sol{1}
.
.
.
How much is 2+2? Solution. 4
Figure 93: \ifthenelse Results (Source in Figure 92)
Another example is to decide on page numbering. Suppose that you
want no page numbering if there is only one page. If there is more than
one page, suppose you want to use headings in the header with page num-
bers. You can do this as follows. First, just before \end{document}, put
\label{lastpage}. Just after \begin{document}, put the following:
\ifthenelse{\equal{\pageref{lastpage}}{1}}{\pagestyle{empty}}%
{\pagestyle{headings}}
This will test if the value of \pageref{lastpage} is equal to 1. If so,
there is only one page, and the first statement applies: \pagestyle{empty}.
If not, there is more than one page, and the second statement applies:
\pagestyle{headings}.
8.6 Your Own Bibliography 109
8.6 Your Own Bibliography
You can choose not to use BibT
E
X, and use thebibliography environment
instead. You will have complete control over the formatting, and there
will be no sorting — the list of references will appear in the order you put
them. Instead of the BibT
E
X commands, \bibliography{mybiblio} and
\bibliographystyle{plain}, specify the following:
\begin{thebibliography}{n}
\bibitem[what appears]{label (that you cite)} entry
.
.
.
\end{thebibliography}
where n is the width of the widest label you want to allow. (It works if
you specify 99.) Each \bibitem is an entry, as described for BibT
E
X in §3
(p. 28), with label the unique identifier used by the \cite command. The
option is an alternative to having the references numbered, and you can enter
whatever you like.
Here is a complete example with two references, which I formatted to
agree with the plain style of BibT
E
X:
\begin{thebibliography}{99}
\bibitem{companion} Michel Goosens, Frank Mittelbach and Alexander
Samarin, \textit{The \LaTeX\ Companion}, Addison-Wesley
Publishing Company, Reading, MA, 1994.
\bibitem{tex} Donald E. Knuth, \textit{The \TeX\ Book},
Addison-Wesley Publishing Company, Reading, MA,
15th edition, 1989.
\end{thebibliography}
These will appear in the document’s list of references even if they are not
cited. They can be cited in the same way described in §3: by \cite{companion}
and \cite{tex}, respectively. When citing Knuth’s book, for example, we
obtain [2] in the text. Alternatively, we can exercise the option:
\bibitem[Knuth, 1989]{tex} Donald E. Knuth, . . .
in which case \cite{tex} ⇒ [Knuth, 1989].
Some publishers give you no choice, but if you are writing a report and
have control over the formatting, it generally helps the reader to know some-
thing about the citation. Thus, [Knuth, 1989] is preferred to [2] because
it immediately gives the reader information about the document without
having to flip to the bibliography section.
With you in control, there is no format monitoring, so each entry appears
however you put it, even if there are inconsistencies in style. This is one
reason it is usually better to use BibT
E
X, even though you lose control
over what appears (i.e., they will be numbers). The bib style file, such as
110 8 TAKING CONTROL
plain.bst, applies in either case. Most installations come with more than
the basic plain, and its three variations (given on p. 36). Alternative bst files
are achicago (from the frankenstein package), apalike and plainnat (from
natbib), which give the author and year, such as [Knuth, 1989] instead of [8].
These packages provide even more versatility in how the citations appear
(see [4] or [5, Chapter 13]).
If you want to have several bibliographic units in one document, such as
at the end of each chapter of a book, use the bibunits package, which you
obtain from CTAN [4].
Closing Remarks
Now you know how to write a mathematical document in L
A
T
E
X2
ε
and you
know there is much more you can learn to gain refinements. Besides what
you can do yourself to elevate the quality of the results, there are many
packages, available from CTAN [4]. Figure 94 shows the preamble used for
this document. As you begin to use packages, it is necessary to become aware
of updates. You can visit CTAN from time to time, or you can join their
newsgroup at comp.text.tex (can be done in Google by clicking on Groups).
You will find other packages useful, depending upon your technical area.
Here are some packages that give you special symbols: chemsym, qsymbols,
SIunits, wasysym, and xypic. Also, the algorithm package enables an environ-
ment to write source code with standard language elements, and there are
others with similar properties or for particular programming languages (viz.,
c-pascal and listings). The graphtex package specializes in all sorts of graphs,
including those commonly found in automata theory.
APPENDIX 111
\usepackage{amsmath} % formerly amstex
\usepackage{amssymb} % ams symbols (\mathbb fonts)
\usepackage{amscd} % draws commutative diagrams
\usepackage{bm} % bold math fonts (\mathbm)
\usepackage{graphicx,pst-all} % graphics
\usepackage{hyphenat} % enables control over hyphenation
\usepackage{fancyvrb,moreverb} % verbatim
\usepackage{float} % enable float [H] option
\usepackage[T1]{fontenc} % ...to write \textbf{\textsc{..}}
\usepackage{ifthen} % ifthenelse {condition}{true}{false}
\usepackage{makeidx} % index
\usepackage{mathrsfs} % more math symbols (viz., \mathscr)
\usepackage{multirow} % like multicolumn
\usepackage{theorem} % enables more control over newtheorem
\usepackage{url} % \url{...}
\renewcommand\contentsname{Table of Contents} % Change ‘Contents’
\renewcommand\url{\begingroup\urlstyle{sf}\Url} % put url in sf font
\input{mydefs} % My commands and environments
\makeindex % make myfile.idx (input to makeindex at command line)
Figure 94: Most of the Preamble for this Book
Appendix
This contains complete tables of font information and basic L
A
T
E
X commands.
It is designed like a reference manual for easy lookup, beginning with Ta-
ble 25, which gives conversion among three common units of measurement.
Table 25: Conversions of Common Units of Measurement
pt in cm
pt 1 .01384 .03515
in 72.27 1 2.54
cm 28.45 .3937 1
Table 26 is a guide to how most of the remaining tables are organized.
Afterwards, Table 45 gives special symbols that can be used in either text or
math mode, and Table 46 gives the commands for the picture environment.
112 APPENDIX
Table 26: Reference Tables
Table Contents
Text mode 27 Commands/Environments for Font Appearance
28 Commands/Environments for Controlling Position
29 Text Accents and Symbols
30 Commands for Counters
31 Commands/Environments to Organize Document
32 Commands to Control Document Style
Math mode 33 Commands to Control Fonts in Math Mode
34 Accents in Math Mode
35 Spacing Commands in Math Mode
36 Greek and Special Letters
37 Frequently Used Mathematical Symbols
38 Binary Operations
39 Operators and Quantifiers
40 Special Functions
41 Relation Symbols
42 Arrows
43 Dots Circles, Triangles and Lines
44 Variable Size Symbols
Table 27: Commands/Environments for Text Font Appearance
textbf textit textrm textsc textsf texttt
tiny scriptsize footnotesize small normalsize large
Large LARGE huge Huge underline verb
verbatim
Table 28: Commands/Environments for Controlling Text Position
bigskip center centerline clearpage flushleft
flushright hfill hspace hspace* linebreak
medskip newpage noindent nolinebreak nopagebreak
pagebreak quotation quote raisebox samepage
smallskip tabbing tabular verse vfill
vspace vspace*
APPENDIX 113
Table 29: Text Accents and Special Symbols
á \’{a} ˘ u \u{u} ç \c{c} ˙ x \.{x}
è \‘{e} ñ \~{n} d
.
\d{d} ¯z \={z}
î \^{i}
˝
H \H{H} b
¯
\b{b} ˇ v \v{v}
ö \"{o} oo \t{oo} . . . \dots
æ \ae œ \oe å \aa ø \o
Æ \AE Œ \OE Å \AA Ø \O
Ł \L ß \ss ¿ ?‘ ¡ !‘
Table 30: Commands for Counters
addtocounter label newcounter pageref
ref refstepcounter setcounter stepcounter
thecounter value
Table 31: Commands/Environments to Organize Document
abstract addcontentsline addtocontents
appendix bibliography bibliographystyle
listoffigures listoftables makeindex
maketitle printindex section
subsection subsubsection subsubsubsection
tableofcontents thanks thebibliography
Table 32: Commands to Control Document Style
markright markboth pagenumbering pagestyle
renewcommand setlength thispagestyle
Table 33: Commands to Control Fonts in Math Mode
left boldmath (set in text mode)
cal displaystyle mathbf mathcal
mathit mathnormal mathrm mathsf
mathtt mbox overbrace overline
right textstyle underbrace underline
114 APPENDIX
Table 34: Accents in Math Mode
ˇ o \check{a} ˘ c \breve{e}
´
i \acute{i} ` o \grave{o}
˙ r \dot{x} ¨ n \ddot{y} ¯ . \bar{z} · \vec{v}
ˆı `hat{`imath} ˜ , `tilde{`jmath} \hbar
¯ rn. `widehat{xyz}

o/c `widetilde{abc}
(Note that it is better style to use \imath, rather than i, and \jmath,
rather than ,, to avoid the clash between the accent and dot.)
Table 35: Spacing Commands in Math Mode
What you write What you see
x y ⇒ rn no space
x\,y ⇒ rn thin space
x\;y ⇒ r n medium space
x\quad y ⇒ r n space = 1em
x\qquad y ⇒ r n space = 2em
x\!y ⇒ rn negative thin space
x\negmedspace y ⇒ rn negative medium space
x\negthickspace y ⇒ rn negative thick space
Table 36: Greek and Special Letters
α \alpha θ \theta o o τ \tau
β \beta ϑ \vartheta π \pi υ \upsilon
γ \gamma ι \iota c \varpi φ \phi
δ \delta κ \kappa ρ \rho ϕ \varphi
c \epsilon λ \lambda · \varrho χ \chi
ε \varepsilon j \mu σ \sigma ψ \psi
ζ \zeta ν \nu$ ς \varsigm ω \omega
η \eta ξ \xi
Γ \Gamma Λ \Lambda Σ \Sigma Ψ \Psi
∆ \Delta Ξ \Xi Υ \Upsilon Ω \Omega
Θ \Theta Π \Pi Φ \Phi
ℵ \aleph / \ell 1 \Re · \Im
/. . . 7 {`mathcal A...Z}
APPENDIX 115
Table 37: Frequently Used Mathematical Symbols
{superscript} ^{} / \prime ∞ \infty ∅ \emptyset
{subscript} _{}
Table 38: Binary Operations
± \pm ∩ \cap ∪ \cup \odot
∓ \mp ¯ \sqcap . \sqcup ⊗ \otimes
\times ∧ \wedge ¬ \uplus . \oslash
÷ \div ∨ \vee ⊕ \oplus \ominus
` \setminus
¸
\bigcap
¸
\bigcup
¸
\bigodot
` \backslash

\bigvee
¸
\bigoplus
¸
\bigotimes
¸
\biguplus

\bigwedge
¸
\bigsqcup
Table 39: Operators and Quantifiers
∇ \nabla ∂ \partial

\surd ℘ \wp
∀ \forall ∃ \exists \neg
Table 40: Special Functions
arccos arcsin arctan arg cos cosh cot coth
csc det dim exp gcd hom inf ker
lg lim liminf limsup ln log max min
Pr sec sin sinh sup tan tanh
Table 41: Relation Symbols
≤ \leq ≥ \geq = \neq ≡ \equiv
≺ \prec ~ \succ ∼ \sim
.
= \doteq
_ \preceq _ \succeq · \simeq [= \models
< \ll \gg

= \cong ∝ \propto
⊂ \subset ⊃ \supset · \asymp ∈ \in
⊆ \subseteq ⊇ \supseteq ≈ \approx ÷ \ni
_ \sqsubseteq _ \sqsupseteq
116 APPENDIX
Table 42: Arrows
← \leftarrow ←− \longleftarrow ↑ \uparrow
⇐ \Leftarrow ⇐= \Longleftarrow ⇑ \Uparrow
→ \rightarrow −→ \longrightarrow ↓ \downarrow
⇒ \Rightarrow =⇒ \Longrightarrow ⇓ \Downarrow
↔ \leftrightarrow ←→ \longleftrightarrow | \updownarrow
⇔ \Leftrightarrow ⇐⇒ \Longleftrightarrow ¨ \Updownarrow
→ \mapsto −→ \longmapsto \nearrow
← \hookleftarrow → \hookrightarrow ` \searrow
÷ \leftharpoonup ÷ \rightharpoonup \swarrow$
÷ \leftharpoondown ÷ \rightharpoondown ` \nwarrow
\rightleftharpoons
Table 43: Dots, Circles, Triangles and Lines
◦ \circ ( \bigcirc
\cdots
.
.
.
\ddots
.
.
. \vdots • \bullet
· \frown \smile
´ \triangle \diamond
> \triangleright \triangleleft
´ \bigtriangleup \bigtriangledown
> \bowtie ⊥ \perp
· \top ⊥ \bot
¬ \dashv ¬ \vdash
∠ \angle | \|
[ \mid | \parallel
Some Tips 117
Table 44: Variable Size Symbols
¸
\sum

\int

\oint
¸
\prod
....
. \overbrace{ . } .
....
\underbrace{ . }
¸
\coprod . \overline{ . } . \underline{ . }
n
d
\frac{n}{d}

. \sqrt{ . }
These use `left and `right
() ( ) ¦¦ \{ \} [] [ ]
[ | \lfloor | \rfloor
` \backslash \lceil | \rceil
' \langle ` \rangle
Table 45: Special Symbols in Both Text and Math Modes

\dag § \S c ( \copyright
‡ \ddag ¹ \P £ \pounds
. . . \ldots
Table 46: Commands and Parameters in Picture Environment
put(r. n){stuff } multiput(r. n)(∆r. ∆n){nn:/c:}{stuff }
line(r. n){|cnot/} framebox(nidt/. /cio/t)[j]{tcrt}
vector(r. n){|cnot/} dashbox{do:/
s
i.c}(nidt/. /cio/t)[j]{tcrt}
circle{:odin:) makebox(nidt/. /cio/t)[j]{tcrt}
circle*{:odin:) oval(nidt/. /cio/t)[j]
linethickness{di:cn:ion}
j ∈ ¦l,r,t,b,lt,lb,rt,rb¦. For oval, it is the portion selected;
for boxes, j is where the text goes.
Some Tips
Here are some tips that apply in special situations.
Structuring large documents. Theses, books, and other large documents
are best managed with the \input command. The document might
look like this:
118 Some Tips
Preamble
\begin\{document}
\input{Abstract}
\input{Chapter1}
\input{Chapter2}
.
.
.
\end\{document}
Preamble
\begin\{document}
%\input{Abstract}
\input{Chapter1}
%\input{Chapter2}
.
.
.
\end\{document}
(a) Separate chapters in a document
(input files are
Abstract.tex, Chapter1.tex, . . . )
(b) Compiling just Chapter 1
This is also a good structure for co-authoring.
Counting words. Under linux enter dvi2tty file.dvi | wc --words
This takes file.dvi as input into the program dvi2tty, which con-
verts it into a plain text file (without latex commands). That output
file is passed to the command wc, which counts the words. If you need
to apply this to a portion of the document, such as an abstract, you
can put that portion into a skeleton latex file and compile it. Using
the above structure, this is particularly easy to obtain the word or line
count of any section.
Commenting text. Besides putting % in column 1, you can logically delete
a block of text by \if 0 ... \fi. This is a good way to manage a
document when you might want to remove some text with the idea
that you it might be restored, at least in part.
\if 0
stuff here is ignored
\fi
(See §8.5 for more general control.)
\newcommand vs. \def. It is generally better to use \newcommand because
you will get an error if the command name is already defined, such as by
some package. The syntax is \newcommand{\name}[#]{what to do};
for example, \newcommand{\note}[1]{\texttt{Note: #1}} defines a
\note command that has one argument. When you specify \note{hello},
you will get Note: hello in the text.
\def is to be used only when you want to allow an over-ride (without
warning). Its syntax is \def\name{what} or \def\name#1{what} if
there is one argument. To suppress printing all notes, defined by the
above \note command, you can specify \def\note#1{ }. This says to
do nothing when seeing \note{stuff}. You can have an \input file with
many commands and environment definitions that you use repeatedly.
It might have something like the following:
REFERENCES 119
\newcommand{\note}[1]{\texttt{note: #1}}
% \def\note #1{ } % activate to suppress printing notes
A “better” alternative (avoiding \def) is to use
%\renewcommand{\note}[#1]{} % activate to suppress printing notes
The \note command is useful when collaborating, and you could define
a separate one for each author, showing initials.
There is also the \providecommand. This has the same syntax as
\newcommand, but it does not issue an error if the command is al-
ready defined. It does nothing if the command exists; otherwise, it
defines the command.
Making / clear. It is good practice to use \ell in math mode when you
want the letter ell because it is easier to distinguish \ell from the
number 1. Compare / vs. | with 1 — also compare the source:
$\ell$ vs. $l$ with $1$.
Also consult the T
E
X Catalog Online [14].
References
[1] Johannes L. Braams. Babel, a multilingual style-option system for use
with L
A
T
E
X’s standard document styles. TUGboat, 12(2):291–301, 1991.
Available at CTAN [4].
[2] Johannes L. Braams, David P. Carlisle, Alan Jeffrey, Frank Mittelbach,
Chris Rowley, and Rainer Schöpf. L
A
T
E
X2
ε
and the LaTeX3 project.
World Wide Web, http://www.latex-project.org/latex3.html, 1994–99.
[3] David P. Carlisle. Packages in the ‘graphics’ bundle. World Wide Web,
CTAN /macros/latex/required/graphics/ (see [4] for replacing CTAN),
1994–99.
[4] Comprehensive T
E
X archive (CTAN). UK: ftp.tex.ac.uk/tex-archive/;
Germany: ftp.dante.de/tex-archive/; USA: ftp.tug2.cs.umb.edu/tex-
archive/ . These are host sites, which contain a list of mirror sites.
[5] Michel Goosens, Frank Mittelbach, and Alexander Samarin. The L
A
T
E
X
Companion. Addison-Wesley Publishing Company, Reading, MA, 1994.
[6] John D. Hobby. A User’s Manual for MetaPost. Computing Sci-
ence Technical Report no. 162, AT&T Bell Laboratories, Murray Hill,
New Jersey, 1992. Available at http://cm.bell-labs.com/who/hobby/
MetaPost.html/.
120 REFERENCES
[7] John D. Hobby. Drawing Graphs with MetaPost. Computing Science
Technical Report no. 164, AT&T Bell Laboratories, Murray Hill, New
Jersey, 1993. Available at http://cm.bell-labs.com/cs/cstr/164.ps.gz.
[8] Donald E. Knuth. The T
E
X Book. Addison-Wesley Publishing Company,
Reading, MA, 15th edition, 1989.
[9] Leslie Lamport. L
A
T
E
X: A Document Preparation System. Addison-
Wesley Publishing Company, Reading, MA, 1986 (also see 2nd edition,
1994).
[10] L
A
T
E
X2
ε
for authors. CTAN /macros/latex/doc/usrguide.ps (see [4] for
replacing CTAN), 1995–99.
[11] Oren Patashnik. BibT
E
Xing. World Wide Web, http://www.uic.edu/
depts/adn/infwww/ps/btxdoc.ps, 1988.
[12] Keith Reckdahl. Using imported graphics in L
A
T
E
X2
ε
. World Wide Web
site Version 2.0, Comprehensive T
E
X Archive, CTAN /info/epslatex.ps
(see [4] for replacing CTAN), 1995–97.
[13] Christian Schenk. MiKT
E
X Local Guide. World Wide Web, http:
//www.miktex.de/, 1998–99 (version 1.2).
[14] Graham Williams. T
E
X Catalogue Online. Technical report, CTAN,
http://www.ctan.org/tex-archive/help/Catalogue/catalogue.html.
[15] Timothy Van Zandt. PSTricks: PostScript macros for Generic TeX.
World Wide Web, http://www.tug.org/applications/PSTricks/, 1993–98.
Index
`Bigg, 63
`Big, 63
`addtocounter, 43
`arraystretch, 107
`author, 93
`baselineskip, 25, 55
`baselinestretch, 25, 108
`bibliographystyle, 36
`bibliography, 36
`bigg, 63
`bigskip, 11
`big, 63
`boldmath, 47
`cdots, 56, 61
`centerline, 7, 19
`cite, 37
`clearpage, 41
`cline, 16
`dashbox, 72
`date, 93
`def, 118
`displaystyle, 48, 55
`documentclass, 1, 93
`dotfill, 25
`dots, 8
`dvi2tty, 118
`dvipdf, 4, 28, 99
`dvips, 3
`ensuremath, 101
`equal, 108
`fboxrule, 42
`fboxsep, 42
`fbox, 9, 41, 55, 71, 74
`footnotesep, 99
`frac, 48, 60
`framebox, 71, 74
`frame, 9
`hfill, 10, 12, 25
`hline, 15, 54
`hrulefill, 25
`hspace*, 23, 25
`hspace, 25
`imath, 114
`input, 101, 117
`itemsep, 106
`jmath, 114
`kill, 23
`label, 39, 53
`left, 56, 117
`linebreak, 23
`line, 74
`listoffigures, 97
`listoftables, 97
`makebox, 72
`makeindex, 98
`maketitle, 93
`mathbb, 59
`mathscr, 59
`mathfont, 47
`mbox, 51, 57, 58
`medskip, 11
`multicolumn, 19, 101
`newcommand, 36, 100, 118
`newcounter, 42
`newenvironment, 103
`newline, 24
`newpage, 24, 41
`newtheorem, 61, 62
`nocite, 37
`nohyphens, 24
`noindent, 10
`nolinebreak, 24
`nopagebreak, 24
`oddsidemargin, 104
`overbrace, 56
`overline, 56
`overset, 65
`pagebreak, 24
`pagenumbering, 97
`pageref, 40, 62
121
122 INDEX
`pagestyle, 97, 108
`parbox, 18, 19, 72
`parindent, 106
`parsep, 106
`parskip, 106
`partial, 60
`prime, 54
`printindex, 98
`prod, 48
`providecommand, 119
`psset, 78
`raggedright, 57
`raisebox, 57
`refstepcounter, 43
`ref, 39, 40, 53, 62
`renewcommand, 25, 34, 43, 98, 100,
119
`renewenvironment, 103
`right, 56, 117
`samepage, 24
`section*, 98
`section, 5
`selectfont, 38
`setcounter, 43
`setlength, 42, 72, 99
`smallskip, 11
`sqrt, 48
`stackrel, 65
`stepcounter, 43
`subsection, 5
`substack, 65
`tableofcontents, 97
`textstyle, 48
`textwidth, 104
`textfont, 8
`thanks, 94
`theenumi, 43
`thefootnote, 95
`thecounter, 39, 43
`title, 93
`underbrace, 56
`underline, 9, 16, 56
`underset, 65
`unitlength, 72, 73
`url, 24, 33
`usepackage, 25, 41, 58, 59, 61, 77
`value, 43
`vector, 74
`verb, 21, 118
`vfill, 25
`vspace*, 25
`vspace, 25, 55
`widehat, 56
`widetilde, 56
``, 15
pdftex, 78
titlepage, 93
accents, 22
AMS, 1
amsmath, 61
Bezier curve, 84
bibtex program, 28
body, 1
boldmath, 47
box, 72
co-authoring, 118
column specification, 16
command line, 1
comment, 20
comments, 1, 118
compile, 1
conditional assignment, 56
counting words, 118
cover page, 93
cross referencing, 35
dash, 10
debugging, 2
derivative, 60
document styles, 1
DOS, 3, 28
dvi viewer, 3
dvips, 4
emftoeps, 86
INDEX 123
environment, 1, 6
abstract, 96, 98
appendix, 98
array, 51, 54, 57
axiom, 62
center, 6
corollary, 62
description, 12, 21
document, 1, 93
enumerate, 14
eqnarray, 53
eqnarray*, 53
equation, 54
figure, 40, 41
flushleft, 6, 18, 19, 57
flushright, 6
gather, 64
gather*, 64
itemize, 13
large, 9
longtable, 20
picture, 72
quotation, 10
quote, 9
smallmatrix, 64, 65
tabbing, 22
table, 40, 41
tabular, 15, 21, 41, 56
thebibliography, 109
theorem, 61
verbatim, 21
verse, 11
file
bib, 28, 29, 32, 34–36
bst (bib style), 36, 110
dvi, 1, 78
eps, 86
jpeg, 86
ps (postscript), 3, 78, 86
tex, 1, 32
wmf, 86
float, 40
float page, 41
floating object, 40
font size, 4, 9, 55
font style
bold small caps, 9
boldface, 8
boldmath, 47
calligraphic, 47
Greek, 47
italic, 8, 16
math mode, 47, 54, 58
non-English, 22
Roman, 16
sans serif, 8
slanted, 8
small caps, 8, 16
typewriter, 8
underlined, 8
footnote, 95, 99
fractions, 48
ghostview, 3
global setting, 43
graph, 82
Hamiltonian, 60
horizontal fill, 10
hyphenation, 24
index (making of), 98
jpeg2ps, 86
Lagrangian, 59
landscape, 81
Laplace transform, 59
latex command, 1
line spacing, 11, 25
bibliography, 38
list environment, 12
description, 12
enumerate, 14
itemize, 13
local setting, 42
124 INDEX
make index, 98
math display mode, 45, 48, 52
mathbf, 47
matrix equation, 54
message, 2
Overfull . . . , 2
Repeated entry, 37
Underfull . . . , 2
warning, 2
MetaPost, 78
MiKTeX, 3, 77, 86
nodes, 81
package, ix, 1, 4, 8
acronym, 99
algorithm, 110
amscd, 65
amsmath, 61, 64
amssymb, 58
babel, 22
bibunits, 38, 110
bm, 48
c-pascal, 110
chemsym, 110
fancyvrb, 111
float, 41
fontenc, 9
frankenstein, 110
geometry, 104
gloss, 99
graphicx, 87
graphtex, 110
hyphenat, 24
ifthen, 108
listings, 110
longtable, 20
makeidx, 98
mathrsfs, 59
moreverb, 111
nomencl, 99
psfrag, 88
pstricks, 77
qsymbols, 110
setspace, 25, 108
showkeys, 40
SIunits, 110
theorem, 63
url, 33
wasysym, 110
xypic, 110
page numbering, 108
paragraph positions, 6
preamble, 1, 35, 103, 110
quotation marks, 10
rotate, 81
section, 5
SIAM, 1
spacing, 12, 104
~, 24
horizontal, 25
line, 11, 25
list, 106
math mode, 48, 50, 107
vertical, 11, 25
special character, 1, 12, 15, 20, 21,
24, 29, 45, 103
`textspec , 21
in url, 33
special function, 58
stacking, 65
subscript, 46
stack, 65
subsection, 5
superscript, 46
tabbing commands, 22
table, 15
ticks, 85
transpose, 54
trigonometric functions, 58
units of measurement, 4, 16, 72,
78, 79, 81, 84, 111
INDEX 125
unix, 3, 28, 86
wc, 118
word count, 118
xdvi, 3
YAP, 3

i

Table of Contents
List of Figures List of Tables Preface Acknowledgments
A Sources of L TEX Software

iii vi viii ix ix 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 6 12 15 20 22 23 25 25 28 28 28 28 32 34 36 37 38 39 39 39 40 42 44

1 Overview 2 Text 2.1 Fonts and Paragraphs . . . . 2.2 Lists . . . . . . . . . . . . . . 2.3 Making Tables . . . . . . . . 2.4 Special Characters . . . . . . 2.5 Tabbing . . . . . . . . . . . . 2.6 Line, Page, and Word Breaks 2.7 Spacing . . . . . . . . . . . . Exercises . . . . . . . . . . . . . . 3 Bibliography with BibTE X 3.1 Overview . . . . . . . . . 3.2 The bib File . . . . . . . . 3.2.1 Main body . . . . 3.2.2 Web citations . . . 3.2.3 Additional features 3.3 Declaration and Citation . 3.4 Some Controls . . . . . . Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . .

4 Counters, Labels, and References 4.1 Basic Concepts . . . . . . . . . . 4.2 Intrinsic Counters . . . . . . . . . 4.3 Figures and Tables . . . . . . . . 4.4 Defining Your Own . . . . . . . . Exercises . . . . . . . . . . . . . . . .

ii 5 Math Mode 5.1 Mathematical Symbols . . . . . . . . . . . 5.2 Fractions and Variable Size Functionality 5.3 Arrays and Equations . . . . . . . . . . . 5.4 Special Functions and Alphabets . . . . . 5.5 Derivatives and Integrals . . . . . . . . . . 5.6 Theorems and Definitions . . . . . . . . . 5.7 Refinements . . . . . . . . . . . . . . . . . 5.8 Grammar . . . . . . . . . . . . . . . . . . Exercises . . . . . . . . . . . . . . . . . . . . . 6 Graphics 6.1 Picture Environment 6.2 PSTricks . . . . . . . 6.3 Importing pictures . Exercises . . . . . . . . . 7 Making Special Parts 7.1 Cover Page . . . . 7.2 Abstract . . . . . . 7.3 Other Front Matter 7.4 Back Matter . . . . 7.5 Footnotes . . . . . Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 46 48 51 58 60 61 63 66 67 71 71 77 86 90

. . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . .

93 . 93 . 96 . 97 . 98 . 99 . 100 . . . . . . 100 100 102 103 103 108 109 110 111 117 119 121

8 Taking Control 8.1 Your Own Abbreviations and Commands 8.2 Your Own Names, Titles and Numbers . . 8.3 Your Own Environments . . . . . . . . . . 8.4 Your Own Margins and Spacing . . . . . . 8.5 Your Own Output Control . . . . . . . . . 8.6 Your Own Bibliography . . . . . . . . . . Closing Remarks Appendix Some Tips References Index

. . . . . . . . Positioning Paragraphs Source (Result in Figure 7) . . . . . . . . . . . Command Sequence from Source to Postscript or PDF . . . \parbox Source (Result in Figure 25) . . . . . . . . Some Font Sizes Source (Result in Figure 11) . A Your First L TEX Source File . . . . . . . . . Nested Tables Result (Source in Figure 22) . . . . . . . . . . . . \parbox Result (Source in Figure 24) . A Table with Partially Spanning Horizontal and Vertical Lines Nested Tables Source (Result in Figure 23) . . Skipping Line Spaces Source (Result in Figure 13) . . . . . . . . \displaystyle Result (Source in Figure 40) . Variable Sizes Result (Source in Figure 38) . . . . . . . Centering Source (Result in Figure 9) . . . . . . . . . . . . Multicolumn Source (Result in Figure 27) . . . . . Tabbing Result (Source in Figure 29) . . . . . . . . . . . \displaystyle Source (Result in Figure 41) . . . . . . . . . . . . Framed Figure with Caption at Top . . . . . . . . . . . . . . . . 1 2 4 5 6 6 7 7 8 9 9 11 11 13 13 14 14 15 15 16 16 17 17 19 19 20 20 21 23 23 28 37 42 42 42 44 44 48 48 49 50 . . . . . . . . . . . . . Skipping Line Spaces Result (Source in Figure 12) . . . . . . . . . . . . A 2 × 3 Table . . . A 2 × 3 Table with Horizontal and Vertical Lines . . . . . . . . . . . . . . Alternative enumerate Symbols Source (Result in Figure 37) Alternative enumerate Symbols Result (Source in Figure 36) Variable Sizes Source (Result in Figure 39) . Adding bibtex to the Command Sequence . . . Enumerate List Environment Result (Source in Figure 17) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Obtaining Brackets in a Description List Environment . . . . . . . . . . Itemize List Environment Source (Result in Figure 16) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Enumerate List Environment Source (Result in Figure 18) . . . . . . . . . . . . . . . . . An Introductory Document Source (Result in Figure 5) . Centering Result (Source in Figure 8) . . .iii List of Figures 1 2 3 4 5 6 7 8 9 10 11 12 13 14 16 15 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 A The Structure of a L TEX Document. . . . . An Introductory Document Result (Source in Figure 4) . . . . . . . Framed Figure 34 Source . . . Description List Environment . Some Font Sizes Result (Source in Figure 10) . . . . . . . Itemize List Environment Result (Source in Figure 15) . . . . . . . . . . . . . . . . . . . . . . . Positioning Paragraphs Result (Source in Figure 6) . . . . . Multicolumn Result (Source in Figure 26) . . . . . . . . . . . . . . . . . . . Tabbing Source (Result in Figure 30) . . . . . . Framed Figure with Caption at Bottom . A Document to Print the Bibliographic Database . . . . . . . . . . . . . . .

. . . . eqnarray Environment Result (Source in Figure 43) . . . . . Specifying Dimensions in \includegraphics . . . . 93 . . . . Commutative Diagram Source (Result in Figure 58) . . . . . . . . . 72 . . . . . . . . . . . 58 . . . . . . . . . . . . . . Making an Abstract Result (Source in Figure 78) . . . . . . 64 . . Graph Source (Result in Figure 66) . . 58 . . . . 55 . . . . . . Graph Result (Source in Figure 65) . . Applying \includegraphics to Import an eps File . . . Authors with same footnote (Source in Figure 76) . . . . . . . . . 54 . Adding Addresses to Authors . . . . . . . . . 97 . . . . . . . . . 50 . gather* Environment Result (Source in Figure 55) . . . . .iv 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 Examples to Compare Text and Display Modes . 82 . gather* Environment Source (Result in Figure 56) . 84 . . . 53 . . . . . . . . 95 . . 93 . . . . . . . . . . . . . . Setting a Footnote Result (Source in Figure 82) . 55 . . . Vertical Diagram Result (Source in Figure 59) . 87 . Line Parameters . . . . . 100 . . 57 . . . . . 99 . . Horizontal Braces Result (Source in Figure 49) . 99 . 66 . . . 53 . . . . . . . . . . . . . Source for Figure 61 . . . Sequence of PSTricks Commands to Draw Histogram . . . . . Title Page Source (Result in Figure 72) . . PSTricks Source for Connecting Nodes . 73 . . . 54 . . Setting a Footnote Source (Result in Figure 83) . . . . . Variety of Objects in Picture Environment . . . . . . . . 96 . . . . . . . . . . . . . . . . . . . . . . . 88 . . . . . . Source Code for Drawing Histogram of Test Scores . . 88 . \flushleft in parbox Result (Source in Figure 51) . . . . . . . . . 56 . . . . 72 . Matrix Equation Source (Result in Figure 46) . 99 . . . . . . . . . Commutative Diagram Result (Source in Figure 57) . . . . . . . . . . Nested Arrays Result (Source in Figure 47) . . . . . . Vertical Diagram Source (Result in Figure 60) . . . . . . . . . eqnarray Environment Source (Result in Figure 44) . . . Nested Arrays Source (Result in Figure 48) . 95 . . Some Front Matter Specifications for This Document . Setting a Footnote Source (Result in Figure 85) . . . . . . 64 . . . . . 66 . . 82 . \raggedright in parbox Result (Source in Figure 53) . . Making an Abstract Source (Result in Figure 79) . Adding makeindex to the Command Sequence . . . . . \raggedright in parbox Source (Result in Figure 54) . . . . 97 . 95 . Matrix Equation Result (Source in Figure 45) . . \flushleft in parbox Source (Result in Figure 52) . . . 96 . 75 . 83 . . Title Page Result (Source in Figure 71) . . . . . . . . 56 . Footnotes in the Cover Page Source (Result in Figure 75) Footnotes in the Cover Page Result (Source in Figure 74) Authors with same footnote (Result in Figure 77) . . . 57 . . . . 94 . 72 . . Horizontal Braces Source (Result in Figure 50) . . . . . . .

. . . . . . . . . . . . . . Changing the Left Margin of a List (Source in Figure 88) . . . Array with Fixed Width Column Source (Result in Figure 91) Array with Fixed Width Column Result (Source in Figure 90) \ifthenelse Source (Results in Figure 93) . . . . . . . . . . . . . . 100 105 106 107 107 107 107 108 108 111 . . Document Margins . . . . . . Changing the Left Margin of a List (Result in Figure 89) .v 85 86 87 88 89 90 91 92 93 94 Setting a Footnote Result (Source in Figure 84) . . \ifthenelse Results (Source in Figure 92) . . . . . . . . . . . . . . . . . . . . . . Most of the Preamble for this Book . . . . . . . . . . . Varying \itemsep to control item spacing in a list .

. . . . . . . . . . . . . . . Default Settings for enumerate Counters . Operators and Quantifiers . . . . . . . . . . . . . . Commands to Control Fonts in Math Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Variable Size Mathematical Operation Symbols . . . . . . . . . . . . . . . . . . . . . . . . . . Accents in Math Mode . . . The Tabbing Environment . . . . . . . . . . . . Figure and Table Location Options . . Spacing Commands in Math Mode . . . . . . . . . . . . . . . . . . . . . . . . . . Commands/Environments for Controlling Text Position Text Accents and Special Symbols . . . . . . . . . . Spacing Parameters . . . . . . . . . Parameters for \psaxes . . . . . . . . . . . . . . . . . . . . . . . . Numerals to Print Counters . . . . . . . . Transpose of a Vector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Commands/Environments to Organize Document . . . . . . . . . . . . . . . . . . . . . . Conversions of Common Units of Measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Notation Using mathbb Fonts from amssymb Package . . . . . . . . . . . . . . . . . . . . . . . . The \mathfont Commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Reference Tables . . . . Binary Operations . . . . . . . . . . . . . . . . . . . Some Basic Drawing Commands in PSTricks . . . . . . . . . . Order Relations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Frequently Used Mathematical Symbols . . . Some Common Mathematical Functions . . . . . . . . . . . . . . . . . . . . . Some Mathematical Operations . . . . . . . . . . . Margin Parameters . . . . . . . . Boxes in PSTricks . . . . . . . . . . . Intrinsic Name Parameters . . . . . . . . . . . . . Some Accents for Letters . . . . . . . . . . . Some Symbols in Logic . . . . . . . . . . . . . . . . . . . . . Commands to Control Document Style . . . . . . Special Functions . . . . . . . . . . . . Commands for Counters . . . . . . . . . . . . . Relation Symbols . . . . . . . . . . . . . . . . . . . . Writing Special Characters . . . . . . . Commands/Environments for Text Font Appearance . . . . . . . . . . . 8 21 22 22 23 41 43 44 46 46 47 49 51 51 54 59 59 59 79 80 86 102 104 106 111 112 112 112 113 113 113 113 113 114 114 114 115 115 115 115 115 . . . . . . . . . . . . . . . . . . . . . . . Greek and Special Letters . . . . . . . . . . . . . . .vi List of Tables 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 Intrinsic Font Styles . . . . The \kill Tabbing Command . . . . . . . . . . . . . . . Set Notation . . Examples of Mathematical Functions . . .

. . . .vii 42 43 44 45 46 Arrows . . . . . . . Triangles and Lines . . . . . . Circles. . . . . . . 116 116 117 117 117 . . Commands and Parameters in Picture Environment . . . . . . . . . . . . . . . . Variable Size Symbols . . . . . . . . . . . . . . . Dots. . . . Special Symbols in Both Text and Math Modes . . . . . . . . . . . . . . . . . . . . .

Exercises are provided for guided instruction. not a word processor. 10]. You enter some editor that saves plain text files. and one could do all of the exercises. Chris Rowley. It was 10 years later that he published his seminal book [8]. Braams. and Rainer Schöpf [2]. Greenberg Denver. a descendant of A L TEX. especially getting used to functioning at the command line. Then. and he permanently changed the way mathematical A documents are prepared. Knuth [8]. Carlisle. thesis. A Happy L TEXing. like ai +ε 1 + (x − µ)2 dx Φ(ε) . but he published articles along the way. David P. designed by Leslie Lamport [9]. For one who is just learning how to use a computer. In any case. using L TEX 2ε to produce high quality results. or article. such as italic font or a complex mathematical expression. By contrast. which should be just a few classes. L TEX (pronounced Lah-tek or Lay-tek) is a collection of macros built on top of TEX that “represents a balance between A functionality and ease of use” [9. the finer points require more study. it will take longer. Originally believing that he could write a program in less than a year that could typeset documents. omitting many of the things L TEX 2ε can do. Colorado .viii Preface A The majority of this book is about using L TEX 2ε [2. you type text freely until you need something special. the basics that are covered should take less than 10 hours. this book is designed to given by the The L TE A be a succinct introduction. developed by a team of volunteers: Johannes L. particularly unix. xiii]. originated by Donald E. For one who is well acquainted with computers. My goal is to offer enough of an introduction that someone not acquainted with A A L TEX (or with TEX) can write a term paper. p. he actually ended up defining an entire branch of research in computer science. low cost publications in mathematics and related disciplines that caused Knuth (pronounced Kah-nooth) to invent TEX (pronounced Tek) in the late 1970’s. based on TEX. L TEX 2ε is the current version. This is a typesetting program. A A comprehensive coverage of L TEX and the many enhancements to it is A X Companion [5]. Frank Mittelbach. — Harvey J. Alan Jeffrey. ε→0+ lim ai It was the desire to have high quality.

Dan Luecking.ix Acknowledgments The author thanks the many contributors in the comp. Andrea Dean. Timothy Murphy. I thank Allen G. Bernd Schandl. notably William Briggs. Herman Bruyninckx. MA.tug.tex.uk/ in Cambridge. I also received useful comments from people who read an earlier draft that I made available on the web. UK. David Carlisle. Mårten Svantesson. These all describe how to search and browse the FTP sites for software and documents. Anton Schwaighofer.de/ in Mainz. A Sources of L TEX Software A The basic L TEX software system is available free of charge for unix systems. Germany (in German). Robin Fairbairns. . I especially thank Kasper B. Last. Jonathan Fine. is at the Comprehensive TeX Archive Network (CTAN) [4]. who were very generous with taking time to answer so many questions on a regular basis. Graversen whose in-depth review has made this version much better than my original. who A taught me L TEX in the first place. One student. and additional packages that extend the L TEX capabilities (to which I refer in this book). David Haller. 2. provided feedback that led to several points of clarification.text. Denis Girou. USA. The best A source of these. at three host sites (and many mirrors): 1. and Matt Swift.tex newsgroup. http://www. and 3. Axel Reichert. http://ctan. Sebastian Rahtz. http://www. and MiKTeX [13] is available free of charge for DOS systems.dante. Thomas Ruedas.org/ in Boston. but not least. particularly Donald Arseneau.ac. Holder.

some of the options.) This is called compiling your source. declaring use of packages \begin{document} . book. which creates several output files.1 1 Overview A You will create a file. which you enter by running Start−→Programs−→MS-DOS Prompt. All that follows is ignored (could be used for comments). which I shall describe in greater detail throughout this book. . called the L TEX source. concentrating on getting started A with using L TEX as quickly as possible. the command line is the DOS command line. and slides. For example. A Are you ready to write your first L TEX document? Copy the source file shown in Figure 2 and name it myfile. which they provide for you to download. .g. Figure 1 shows the structure of this file. which is plain text. The focus throughout this book is on the article style. I refer to myfile. enter: latex myfile (In an MS Windows system.. depending upon the style. Then. at the command line. I shall be using defaults for almost everything.tex. or by Start−→Run and entering cmd or command into the window. Among these are professional societies. \end{document}    Anything following % is ignored (used for comments). A Figure 1: The Structure of a L TEX Document. which the . report. Most publishers have their own style. notably the American Mathematical Society (amsmath style) and the Society for Industrial and Applied Mathematics (siam style). the intrinsic document styles are: article. Later. there are many options.tex. letter. Preamble (blank lines do not matter) Body This is the document environment. In the preamble.tex as a plain text source file that you create. as well as more advanced topics for customizing your document. like margin settings and other preamble specifications are covered. To keep things simple. % This is myfile. The only one you need to be concerned with now is the dvi file. Further.tex % notes to yourself can go here \documentclass[options]{style} optional specifications — e. its suffix is .

-2 ... Sometimes the error message tells you what went wrong. but they were not fatal errors. ). so you are advised to compile often. You got a fatal error message. Something went wrong and you need to ask for help. Among the non-fatal messages you will generally see are warnings like: Overfull \hbox .. You must find and correct it.... the first thing to do is save a backup by copying your source file to another subdirectory.tex (The -p is to keep the date and time of the source file. where the destination is either a: or some backup file name. (If you are familiar with .dvi. . You got no messages.. If all went well. or to a different name. In unix do this by entering: cp -p myfile. use copy myfile. Underfull \hbox .tex destination.) Change -1 to another qualifier each time (e. such as missing a brace (characters { and }. You got messages.g.tex myfile-1. so you have a collection of backups. Overfull \vbox ..2 1 OVERVIEW \documentclass{article} \begin{document} Hello world. or some command was not recognized due to being misspelled. Many times the message is not very informative. If you are running under DOS. Underfull \vbox . That way you will know that what you did in the last few minutes contains the error. Case 2. Case 3. which you will come to know and love). Do not worry about these. One of three things will have occurred: Case 1.. This is called debugging your source. \end{document} A Figure 2: Your First L TEX Source File latex program (called a “compiler”) names myfile. .

you can run a program called ghostview. xdvi. vi or vim. press Enter at the end of a line instead of letting the word processor do it for you). nothing more need be said. If you use a word processor. in which case you might ask someone for help. and you must continue to specify the suffix as . you need to learn how to create. (There is more to do if you are working remotely. You can print it by any number of ways. including the unix command.) If you are using DOS. Under unix.tex . it will use .doc as its suffix). xdvi does not have a print option. a free software system by Christian Schenk. or you can use Notepad. You should even put in hard return characters (i. emacs. and you might want to obtain myfile.ps. enter EDIT myfile.) This will result in the creation of the postscript file myfile. edit. you can view the result with the dvi viewer. If you are in a unix environment. You will need to find out more about viewing and printing postscript files that suit your particular needs.3 DOS.tex (otherwise. In unix this could be pico. dvips. your installation might already have file output as the default.ps for a variety of reasons. Wordpad or MS Word. and never use tabs. In MS Word. To view or print a postscript file under DOS. when saving the file. you begin by entering a plain text editor. This is done with the program. The same conversion program can be run under DOS (and comes with MiKTeX). you must take absolutely no advantage of its formatting. be sure to specify plain text. At the command line enter: xdvi myfile and it will come on your screen. in which case the -o is not needed. including posting it on the web. the viewer that comes with MiKTeX [13]. If you want to check that the file is really plain text. lpr myfile. rather than just print. At the command line enter: dvips myfile -o (The -o tells the system you want the output to go to a file. if not. In DOS you can use EDIT at the command line. however. is called YAP. Summarizing. so you first need to convert the dvi file to postscript.e.ps. and save plain text files. At the DOS command line you enter: YAP myfile You will see various options for viewing and printing..) Next you want to view the result.

but do not be reluctant to read the last chapter.pdf Figure 3: Command Sequence from Source to Postscript or PDF Now change your document to specify a font size of 12pt (default is 10pt) by changing your first line to the following: \documentclass[12pt]{article} The “pt” (abbreviation for “point”) is one of the units of measurement.” indicated in Figure 1. Congratulations! create/edit compile latex myfile. (There are some free text editors on the web. I also cannot elaborate just yet on “using packages. other units used in many parts of L T X are in (inches).) The orientation here is by function.ps myfile. use your favorite search engine to find them. Execute these commands for the source file shown in Figure 2. Under unix or DOS use dvips to convert the dvi file to a postscript (ps) file. beginning with how to write text. which is a printer measure equal to the width of M in the current font).dvi dvipdf print/post myfile. (One of A the strengths of L TEX is the community of people who provide packages for everyone to use at no cost. . and em (like the letter m. The result should be one line of output: Hello World. like notation. cm (centimeters). which can be printed.4 1 OVERVIEW at the DOS command line and see how the file appears. and I shall introduce specific packages throughout this book. and if all is well. except to say that they are used to fulfill some function. about 1 A E 72 in. enter latex myfile. enter your dvi viewer. It tells you how to define your own commands and how to separate them into an input file that simplifies changing things.) Once you have your source ready to compile. You can alternatively (or also) convert the dvi file to a pdf file by the command dvipdf myfile These steps are given in Figure 3.tex view/print convert dvips myfile. A This book is designed for quick entry into using L TEX.

Note the automatic numbering. \documentclass[12pt]{article} % We have defined the document to be an article using 12 point font. (The power of L TEX will be evident when we get to mathematical expressions.) First. \begin{document} % Begin document "environment". % Blank lines mean nothing here. I can move to a new line anytime. lots Skipping four lines is the same as skipping one line --.5 2 Text We begin by illustrating the most common text formatting. and how extra spaces and blank lines have no effect.it starts a new paragraph. but even some text. and I can put in of blanks with no effect. Figure 4 is the source that produces the result in Figure 5. \section{This is a Section} \subsection{This is a subsection} This is the body of the subsection. A will demonstrate the superior quality of the L TEX results. consider how to make sections and subsections in article style. much like you A would want in a word processor. especially tables. \subsection{Here is another subsection} \section{Here is another section} \end{document} Figure 4: An Introductory Document Source (Result in Figure 5) . showing how sections and subsections are defined. in the preamble.

.1 This is a Section This is a subsection This is the body of the subsection. Figure 6: Positioning Paragraphs Source (Result in Figure 7) . \end{flushright} I am back to normal justification. The added space you see between the above paragraphs is due to entering those environments. a concept you need to understand about L TEX. Note that these are A environments. and I can put in lots of blanks with no effect. 1. Text remains centered as long as we remain in this environment. You will have occasion to use all four paragraph positions. .2 Here is another subsection 2 Here is another section Figure 5: An Introductory Document Result (Source in Figure 4) 2. flush left. \end{environment} \begin{center} The text is centered because I have entered the center environment.6 2 TEXT 1 1. I can move to a new line anytime. and have begun the flushleft environment. and justified (the default). Skipping four lines is the same as skipping one line — it starts a new paragraph. The general form of an environment uses the following syntax: \begin{environment} . \end{center} \begin{flushleft} Now we are out of the centering environment. but in the flushright environment. \end{flushleft} \begin{flushright} This is another paragraph.1 Fonts and Paragraphs Figure 6 shows the source to produce different paragraph positions: centered. flush right.

2. I am back to normal justification. \begin{center} This line is centered. \centerline{This line is centered. Text remains centered as long as we remain in this environment. but in the flushright environment. shown in Figures 8 and 9. This is another paragraph. The added space you see between the above paragraphs is due to entering those environments. you can use the \centerline command.} This continues after centering. Now we are out of the centering environment.1 Fonts and Paragraphs 7 The text is centered because I have entered the center environment. You will have occasion to use all four paragraph positions. \end{center} Figure 8: Centering Source (Result in Figure 9) . This continues after centering. This precedes centerline. and have begun the flushleft environment. Figure 7: Positioning Paragraphs Result (Source in Figure 6) Instead of the center environment. This precedes center environment. they differ in that the environment skips a line before and after the paragraph.

sl. you will have particular use for the italic font. sc. roman. \textbf{\textit{bolditalic}} ⇒ bolditalic. This is \textsc{small caps}. sans serif. as it is used when introducing a new term. Table 1 lists the fonts that are intrinsic in a basic latex installation. This is \textrm{roman}. This is \texttt{typewriter}. ⇒ ⇒ ⇒ ⇒ ⇒ ⇒ ⇒ How This This This This This This This it is is is is is is is appears boldface. This is \textsf{sans serif}. typewriter.8 This precedes center environment. A \textit{group} is defined on a set of elements \dots ⇒A group is defined on a set of elements . . For example. This precedes centerline. Table 1: Intrinsic Font Styles What you write This is \textbf{boldface}. slanted. italic. Some combinations of font styles can be produced. This is \textit{italic}. This is \textsl{slanted}. produces: This paragraph is not indented. For example. The general form is \textfont{text}. rm. The argument of \textbf is \textit{bolditalic}. . This line is centered. (The ⇒ symbol can be read as “produces. small caps. . tt}.) In technical writing.”) Note the use of the \dots command. This continues after centering. (More fonts are available in packages. Figure 9: Centering Result (Source in Figure 8) 2 TEXT You can also suppress indentation of the first line of a paragraph with the \noindent command. sf. where font is one of {bf. This continues after centering. usually free of charge. which produces the ellipsis. This line is centered. it. Here is an example: \noindent This paragraph is not indented. as seen in Table 1.

Here is an example using the quote environment. and the braces format is useful for short phrases. Figure 11: Some Font Sizes Result (Source in Figure 10) The use of braces to enclose a font size specification is like an environment. You can make it really tiny. Figure 10: Some Font Sizes Source (Result in Figure 11) You can make the text large or You can also make it . Optionally. . \end{size}. Notice how the paragraph spacing changes to accommodate the variation in font size.2.) It is straightforward to underline text — just write \underline{text}. You might want to make something small or smaller or even smaller still. The environment syntax is useful when you want to keep the size for a large block of text. p. You can also make it {\huge huge}. Font size can also be varied at will. We can give frame some room around the edges by using \fbox instead.\end{large} produces the same result as {\large This is large. . These size variations can be combined with font styles. but learning it must wait until we enter math mode. Now consider ways to indent a block of text. 71. you must put \usepackage[T1]{fontenc} in your preamble to obtain: \textbf{\textsc{bold small caps}} ⇒ bold small caps. In particular. which was generated by putting \begin{quote} before the text and \end{quote} after it: huge larger or even larger still. For example. We can also frame text just by writing \frame{text}.) To overline is as straightforward. Figures 10 and 11 give the source and result for common variations.}. (There is no intrinsic environment for font styles. (More on framing in §6. You might want to make something {\small small} or {\footnotesize smaller} or even {\scriptsize smaller still}. You can make the text {\large large} or {\Large larger} or even {\LARGE larger still}. we can explicitly use the environment syntax: \begin{size} . . \begin{large} This is large. such as using {\Large\textbf{heading}} for some heading.1 Fonts and Paragraphs 9 A Not all combinations are in the basic L TEX 2ε installation. You can make it really {\tiny tiny}.

except the first line of each new paragraph is indented. to make the line flush right. • The dash that appears before each name is created by three minus signs. any more than they play. by skipping a line and entering \hfill (which means horizontal fill). notably by Dedekind cuts. ---. The more minus signs you use.e. generally one short paragraph (as above). and three are for punctuation — i. not " ". or a sequence of one line quotes. Knuth The quotes are by two pioneers of algorithms. The quotation environment is used for long quotations.” — Alan M. to create this more stylistic quotation punctuation. in ways and for reasons that are only very indirectly reflected in his overt activity. Their names appear on the right.) Here is an example that was created by putting \begin{quotation} before the text and \end{quotation} after it. Alan M. this can be overridden by the \noindent command. The indentation is the same as the quote. The convention is that one dash is for hyphenation. Turing “There are reports that many executives make their decisions by flipping a coin or by throwing darts.” — Donald E. The quote environment is intended for short quotes. for example in situations where a fixed decision made each time would cause the algorithm to run more slowly. having more than one paragraph (separated by blank lines). use --preceding “i. I used ‘‘ ’’.. Turing and Donald E. etc.10 2 TEXT The construction of the real number system.” . but we know what they indicate: a great deal of information processing goes on far beneath the surface of man’s purposive behavior. two are for ranges. was motivated by the need to fix calculus. after their quote. “Computers do not dream. Here are some other things to notice about this example: • There are left and right quotation marks. separated by blank lines. Knuth.e. the longer the dash. such as page numbers. It is also rumored that some college professors prepare their grades on such a basis. Sometimes it is important to make a completely ‘unbiased’ decision. which ran into trouble due to insufficient rigor in dealing with limits. We are far from certain what dreams are good for. (Just as in the regular text. this ability is occasionally useful in computer algorithms.

Figure 13: Skipping Line Spaces Result (Source in Figure 12) The verse environment indents oppositely: lines after the first. Figure 12: Skipping Line Spaces Source (Result in Figure 13) This is a first line. men who are thus ignorant are unable to perceive their own ignorance and so do not seek a remedy. }. — Roger Bacon The italics were specified in the usual way. each line is a stanza in the verse. Figures 12 and 13 illustrate three levels of skipping: small. The space you see just above is a medium skip. \medskip \smallskip The space you see just above is a medium skip. medium and big. The space you see above is a big skip. since he who is ignorant of it cannot know the other sciences or the things of this world. Later. we shall look at a way to have a much finer range of vertical spacing.1 Fonts and Paragraphs 11 • There is extra space between the two quotations. This is just an ordinary line space.2. . This is just an ordinary line space. . And what is worse. . The space you now see just above is a small skip. This is a first line. by enclosing Bacon’s verse with \textit{. The space you now see just above is a small skip. (Designed for poetry. \bigskip The space you see above is a big skip. The following was generated by putting \begin{verse} before the text and \end{verse} after it: Neglect of mathematics works injury to all knowledge. This was done with the \bigskip command.

This is a power of L TEX and one reason why it has become standard in writing mathematical papers. report or article.) Bacon’s name appears flush right. A LT Two new things appear in the example: the use of \LaTeX to produce Without the EX. This is achieved by not skipping a line after the verse: \begin{verse} \textit{Neglect of mathematics . Bibliography. following \LaTeX. L TEX provides a means to make tables with the tabular environment. make paragraphs. here is the use of a description list environment A to itemize steps involved in learning L TEX. This has progressed a great deal in the past few years thanks to many people who have provided packages free of charge. A Making Tables. but this time it is on the last line of the verse. in particular with BibTE X.) . and one might want a punctuation mark. To illustrate. even with a space after \LaTeX in the source. rather than a new line. A Mathematics. this form of indentation makes it clear.Roger Bacon \end{verse} 2. which requires no space.12 2 TEXT and if a stanza runs long. distinguished by what appears at the beginning of each item: number. (The reason is that a space (or some delimiter) is needed after \LaTeX (or any keyword) in order to distinguish it completely. and the use of ~ (called “tilde”) to enter a space. I will show you how to do virtually any mathematical expression in line with the text. whose source is indicated by Figure 14. } \hfill --.. and its versatility puts it far ahead of word processors. and list items are enough to prepare a basic document without mathematics or tables (like a resume). Knowing how to create a bibliography. Basic Document Preparation. A tilde.. or your description (perhaps nothing). Graphics. again from the \hfill command. Knowing how to setup the latex source file. or in math display mode. vary fonts. like a comma. bullet.2 Lists There are three intrinsic list environments. the result would be L TEXprovides. There are a great many things to learn beyond the simple introducA tion when using L TEX to prepare a thesis. Other.

. which prints bullets. Blank lines before an item have no effect.. \end{description} 13 Figure 14: Description List Environment The text within the square brackets is an option.. Figures 15 and 16 illustrate the itemize list environment. Next.} \end{description} produces the following result: This is how one item in a description list environment looks with no optional text at the beginning. and the lines extend all the way to the right margin. but any type of list can be nested within any other type. the description list is one way to have text indented the opposite of a normal paragraph: the first line is at the left and subsequent lines are indented.. \item [Making Tables. Figure 16: Itemize List Environment Result (Source in Figure 15) ..2. \begin{description} \item \textsf{This is how one item in a description list environment looks with no optional text at the beginning. A blank line within an item does create a new paragraph. \item [Mathematics..] This is the power of \LaTeX~ and one . \item [Graphics.. the first line goes almost to the left margin.. Unlike the verse environment. If present.. • This is item 2 and we shall limit to just this few.] Knowing how to create a bibliography .] \LaTeX~ provides a means . – A second (nested) itemized list changes the bullet and indents another level.] This has progressed a great deal in the . using the indentation of the itemize environment..2 Lists \begin{description} \item [Basic Document Preparation.. With no option. • This is item 1 and our task has just begun. as in this example. \item [Other. For example.] There are a great many things to learn ..] Knowing how to setup . Note the indentation of each item and the spacing between items. it is printed in boldface. \item [Bibliography. You see the nesting of two itemize lists.

I describe the enumerate list environment. \begin{enumerate} \item This is item 1. \begin{enumerate} \item Back to item 1. With nested enumeration. A blank line within an item does create a new paragraph. \end{itemize} \end{itemize} Figure 15: Itemize List Environment Source (Result in Figure 16) Finally. the numbering changes at each level. \begin{itemize} \item A second (nested) itemized list changes the bullet and indents another level.14 \begin{itemize} \item This is item 1 and our task has just begun. before an item have no effect. 2 TEXT Blank lines \item This is item 2 and we shall limit to just this few. where the default numbering is with Arabic numerals. \item Two is new. Figures 17 and 18 illustrate with three levels of nesting. using the indentation of the itemize environment. and we are having fun. and it’s time to number anew. \item This is item 2. but we are not yet done. \begin{enumerate} \item One again! \item Two (b) or knot 2b? \end{enumerate} \end{enumerate} \end{enumerate} Figure 17: Enumerate List Environment Source (Result in Figure 18) .

center or right.3 Making Tables which has the following A table is made with the tabular environment. abbreviated by just one character: l. How it appears left 1 center 2 right 3 What you write \begin{tabular}{lcr} left & center & right \\ 1 & 2 & 3 \end{tabular} Figure 19: A 2 × 3 Table We can draw a horizontal line before any new row by specifying \hline. and it’s time to number anew. This is item 1. The way to vary horizontal line drawing is by using . Each column specification can be left. . i. but we are not yet done. This is item 2. Figure 19 shows an example of a 2 × 3 table. each row ends with two backslashes. enter \\ \hline (the \\ is simply part of the syntax and does not add an extra row to the table).2. In the body of the table. The column specifications can have | on either side to indicate a vertical line. \\. respectively. syntax: \begin{tabular}{column specs} options first row spec \\ . To draw a line after the last row. each column is separated by &. last row spec [\\ options] \end{tabular} As indicated. One again! ii. c or r. We could draw lines that span some rows and/or columns. Two (b) or knot 2b? Figure 18: Enumerate List Environment Result (Source in Figure 17) 2. (b) Two is new. (a) Back to item 1.3 Making Tables 15 1. . Figure 20 illustrates a combined use of these options. 2. The way to vary vertical line drawing is with the column specifications: put | only where you want a vertical line. and we are having fun.

16 How it appears -110 210 -120.12 220. -130 230 What you write

2 TEXT

\begin{tabular}{|l|c|r|} \hline -110 & 120 & -130 \\ \hline 210 & -220 & 230 \\ \hline \end{tabular}

Figure 20: A 2 × 3 Table with Horizontal and Vertical Lines \cline{first col -last col } instead of \hline. This is illustrated in Figure 21. How it appears Name Bob Sue Test 1 67 72 Test 2 72 67 What you write
\begin{tabular}{l|cc|} Name & Test 1 & Test 2 \\ \cline{1-1} Bob & 67 & 72 \\ Sue & 72 & 67 \\ \cline{2-3} \end{tabular}

Figure 21: A Table with Partially Spanning Horizontal and Vertical Lines We can have tables nested within tables. Figures 22 and 23 illustrate this, while showing more variation with lines and using various fonts. Here are some things to note: • The entire table uses sans serif font style. This is done by specifying \textsf{ before entering the tabular environment (and closing with } just after it). • Within the tables, fonts are varied: Roman is in the Roman font, specified by \textrm{Roman}, Greek is in italic, specified by \textit{Greek}, and upper case is in small caps, specified by \textsc{upper case}. • A new column specification is introduced: p{length}, where any unit of measure can be used as the length of the spacing. In this example .3 inches is specified. Note that this counts as a column, so you see && to separate the two tables, each being a column of the main table. • The \underline command is used to underline Table 1, which is column 1 of the main table, whereas \cline{3-3} is used to underline all of column 3 of the main table, headed by Table 2. There are times when we want to put a good bit of text into some columns

2.3 Making Tables
\textsf{ \begin{tabular}{lp{.3in}l} \\ \underline{Table 1} && Table 2 \\ \cline{3-3} \\ \begin{tabular}{|lc|} \hline Object & Symbols used \\ \hline variable & lower case \textrm{Roman} \\ parameter & \textit{Greek} \\ constant & \textsc{upper case} \textrm{Roman} \\ \end{tabular} && % Begin Table 2 \begin{tabular}{|rcc|} \hline * & 1 & 2 \\ \cline{2-2} & 3 & 4 \\ \cline{1-1}\cline{3-3} \end{tabular} \end{tabular} } % end sf

17

Figure 22: Nested Tables Source (Result in Figure 23) Table 1 Object variable parameter constant Symbols used lower case Roman Greek upper case Roman Table 2

*

1 3

2 4

Figure 23: Nested Tables Result (Source in Figure 22) of a table. Suppose, for example, we write the following:
\begin{tabular}{|l|l|} \hline This amount of text is too long to fit on one line of the page. & This is column 2. \\ \hline \end{tabular}

The result will be to run off the edge of the paper: This amount of text is too long to fit on one line of the page. This is column 2.

One solution is to insert new rows and break up the text manually:

18
\begin{tabular}{|l|l|} \hline This amount of text is too long to fit on one & This is column 2. \\ line of the page. & \\ \hline \end{tabular}

2 TEXT

This amount of text is too long to fit on one line of the page.

This is column 2.

Instead, one can assign a fixed width to a column by specifying p{length}. For example,
\begin{tabular}{|p{2in}|l|} \hline This amount of text is too long to fit on one line of the page. & This is column 2. \\ \hline \end{tabular}

This amount of text is too long to fit on one line of the page.

This is column 2.

Another solution is to use the \parbox command (short for “paragraph box”). This has the form \parbox[option]{width}{text}, where the option is the placement: t = top and b = bottom (default is center). Here are two examples:
\begin{tabular}{|l|l|} \hline \parbox{2in}{This amount of text is too long to fit on one line of the page.} & This is column 2. \\ \hline \end{tabular}

This amount of text is too long to fit on one line of the page.

This is column 2.

\begin{tabular}{|l|l|} \hline \parbox[t]{2in}{This amount of text is too long to fit on one line of the page.} & This is column 2. \\ \hline \end{tabular}

This amount of text is too long to fit on one line of the page.

This is column 2.

They differ only in the placement of the paragraph box, the latter being at the top to align it with column 2 in the manner shown. When making a column or parbox small, the spacing can become unsightly due to being justified. This is overcome with the flushleft environment. Figures 24 and 25 illustrate this, and note that it contains other commands that can be in any paragraph.

can be determined by some length parameter. but we want the headers to be centered. The first argument is the number of columns to span (starting where \multicolumn is specified).. If we want some heading to span several columns.3 Making Tables \begin{center} \begin{tabular}{ll} \parbox[t]{3in}{\begin{flushleft} This is column 1. and I might want to display something: How sweet it is. rather than a fixed constant. we want the columns to be left justified.}} \end{flushleft} } & \parbox[t]{1in}{\begin{flushleft}\textsf{This is column 2. or without. Suppose. r. The second argument is any valid column specification. This is column 2. The \multicolumn command can also serve to override some column specification.. This must be in the range 1 to however many columns remain from the current position. the third argument is the text. such as the width of a paragraph box.2. c. |. see exercise 9 at the end of this chapter and consider \parbox{. on either side. this is done by the command. This is not the same as How sweet it is. which I have put in sans serif font. The line in the source begins with & to skip column 1. for example.} \end{flushleft} } \end{tabular} \end{center} 19 Figure 24: \parbox Source (Result in Figure 25) This is column 1. The first is used to center ‘Test number’ over columns 2 and 3. \multicolumn{number }{col spec}{entry}. a vertical line specification. For example. Finally. with. Figure 25: \parbox Result (Source in Figure 24) Any measurement. which I have put in sans serif font. such as l.}}\medskip This is not the same as \medskip\fbox{\centerline{How sweet it is. Figures 26 and 27 illustrate these uses of \multicolumn.3\linewidth}. then the . and I might want to display something: \medskip\centerline{\fbox{How sweet it is.

instead of the tabular environment.4 Special Characters We have already seen that some characters are special. In particular. The second use simply centers the ‘Student’ header. The last use of \multicolumn centers ‘Taken in class’ over columns 2 and 3. the vertical line at the end is missing because c was specified instead of c|.5 \\ Charleetah & 72 & 67 & 70. written as \{ \} to obtain { }. Of particular importance are the braces.5 Figure 27: Multicolumn Result (Source in Figure 26) Tables that are too long to fit on one page could be broken manually.5 70. which has most of the same options. \ delimits every L TEX command. 2. The Appendix contains complete tables of these special characters (including those I do not cover explicitly in the chapters).) In the preamble. like {\large. it appears with whatever font is active. . }. How do we print such characters? One way is with the symbol. Table 2 (next page) illustrates this with commonly . enabling comments. \begin{center} \begin{tabular}{l|cc|c} & \multicolumn{2}{|c|}{Test number} \\ \multicolumn{1}{c|}{Student} & 1 & 2 & Average \\ \hline Bill & 67 & 72 & 70.20 2 TEXT \multicolumn specifies 2 columns. centered with vertical lines before and after.) When using a keyword to specify a special character. and % ends a line. A but the longtable package enables automatic page breaks by the L TEX compiler. in that they delimit A something. like \%. like \textbackslash. Unlike the first use. Other times a keyword.5 \\ \hline & \multicolumn{2}{c}{Taken in class} \\ \cline{2-3} \end{tabular} \end{center} Figure 26: Multicolumn Source (Result in Figure 27) Test number 1 2 67 72 72 67 Taken in class Student Bill Charleetah Average 70. specify \usepackage{longtable}. . (You obtain the package from CTAN [4]. Then. specify the longtable environment. itself. is needed. (Recall that the braces by themselves create a local environment.

illustrated in Figure 28. It uses any other character to delimit a string. \end{description} [This is not option] for item. [ ]. but I omitted a discussion of position options that are specified by brackets. . One example is the description list environment. \end{description} \begin{description} \item {[This is not option]} for item. For example. . we can write \verb@{}%$#\@ to generate the string {}%$#\ (delimited by @). The verbatim environment uses the usual syntax: \begin{verbatim} . (Another is in the tabular environment (page 15). except when they are used to delimit an option in the syntax. What you write \begin{description} \item [This is option] for item.2.4 Special Characters 21 used special characters. which is printed in typewriter font style. The brackets. \verb does not use braces to delimit its argument. \end{verbatim} . which can contain any character except itself.) How it appears This is option for item. Figure 28: Obtaining Brackets in a Description List Environment Table 2: Writing Special Characters Character (Roman) {} %$#&_ \ ^ ~ r [] How you write it \{ \} \% \$ \# \& \_ \textbackslash \textasciicircum \textasciitilde \textregistered {[ ]} Other fonts italic {} %$#&_ \ ^ ~ r [] large {} %$#&_ \ ^ ~ r [] Another way to print the unprintable is with the verbatim environment or the \verb command. Unlike all other commands. in which case they can be obtained by enclosing them in braces. are different because they can be entered directly.

?‘⇒¿ and \aa⇒å. these are not sufficient. the first sentence continues normally.22 2 TEXT This is how the source code was created for the figures.5 Tabbing The tabbing environment provides an alternative to the tabular environment by letting you set your own column tabs. In addition. so that the position of the tab is not equivalent to that of a table’s column. Table 3 shows some common examples. \\ ends each row. A L TEX has a basic library of accents and special characters for writing in languages other than English. \"{b}\~{c}\^{d} ⇒ b˜d. even if it does not relate to some ¨cˆ language — for example. 20). Another class of special characters are letters with accents. . In some cases. such as Babel [1] (also see [5. some of which are shown in Appendix Table 29 — e. Table 4: The Tabbing Environment What you see Begin: set tab 1. An accent could be applied to any letter. a complete table is in the Appendix. Chapter 9]). Table 4 shows a simple case with two basic tabbing commands. particularly if the entire document is to be in a non-English language.g. write Poincar\’{e} to have Poincaré and G\"{o}del to have Gödel. set tab 2 skip to 1 then to 2 skip to 2 What you write \begin{tabbing} Begin: \=set tab 1\dots \=set tab 2\\ \>skip to 1 \>then to 2\\ \> \>skip to 2 \end{tabbing} Sometimes we do not want to have the longest portion of text first. yet it is needed to define the tab. and \> to move to a tab setting.. Table 3: Some Accents for Letters What you write \"{a} \‘{e} \’{i} \~{o} \^{u} ⇒ ⇒ ⇒ ⇒ ⇒ What you see ä è í õ û 2. like Figure 26 (p. For example. For that purpose there are some packages. . Table 5 illustrates how this is solved with the . but unlike the tabular environment. without extra spaces. \= to define a tab setting.

then it uses the name of the last field to set what follows. then specifies \kill instead of \\ to suppress (or “kill”) the printing of the line. and Word Breaks 23 \kill command. When text is justified (the default). this could result in an undesirable appearance. Table 5: The \kill Tabbing Command What you see 1-3 sting like a bee 4-6-8on’t be late d What you write \begin{tabbing} 1-3 \= sting like a bee \\ 4-6-8 \> don’t be late \\ \end{tabbing} \begin{tabbing} 4-6-8 \= don’t be late \kill 1-3 \> sting like a bee \\ 4-6-8 \> don’t be late \\ \end{tabbing} 1-3 sting like a bee 4-6-8 don’t be late Figures 29 and 30 illustrate the tabbing environment with fixed field widths.5in} \= \hspace*{2in} \= Last field: \= \kill \> Field 1 (following tab 1) \\ \> \> Field 2 on new line \> Last field \\ \> \> \> Last field on new line \end{tabbing} Figure 29: Tabbing Source (Result in Figure 30) Field 1 (following tab 1) Field 2 on new line Last field Last field on new line Figure 30: Tabbing Result (Source in Figure 29) 2.2. \begin{tabbing} \= \hspace*{. in order to set the tab correctly. The second tabbing puts the longer field first. and Word Breaks You can cause a new line by entering \linebreak. but the tab is set by the shorter string ‘1-3’. making ‘4-6-8’ extend past the tab. Page. the lines are in the order we want. Page. In the first tabbing.6 Line.6 Line. It first uses the \hspace* command for horizontal spacing. like the following: .

Sometimes. The \newpage command follows the analogy with \newline in forcing a page break precisely at the point it is specified.} ⇒This example extreme.) Then. This can be done by specifying \usepackage{hyphenat} in the preamble. rather than completing the line as \pagebreak does. Sometimes. preventing a line break. Here is an example that keeps line 1 on the same page as line 2. you want to suppress hyphenation. 10’. to suppress hyphenation. even if it means extending into the right margin. the first sentence of this paragraph has a hyphen. \textsf{Here is the extreme \newline example.~10.} ⇒Here is the extreme example. you obtain the following: A Word breaks are hyphenations that L TEX does for you. however. Sometimes. (You might have to download the package from CTAN [4]. There are two commands to force a page break: \pagebreak and \newpage. however. To prevent a line break where you want a blank. you want to suppress hyphenation. you want to suppress hyphenation. We would thus write figure~1 or p. For example. however. it is better style to keep certain ‘words’ together. such as ‘figure 1’ or ‘p.24 \textsf{This example is \linebreak extreme. The \newline command forces a new line without justifying it. . to suppress it. enter: \nohyphens{Word breaks are hyphenations that \LaTeX\ does for you. and follow the simple installation instructions.} Then. The \nopagebreak command disallows a page break immediately following the next blank line. {\samepage line 1 \nopagebreak line 2 } A Word breaks are hyphenations that L TEX does for you. Also. 2 TEXT is The \nolinebreak command works analogously. use the space character ~. The \samepage command prevents a page break within its scope. you specify the \nohyphens command.

. . . . The easiest way to control line spacing throughout your document is to specify usepackage{setspace} in your preamble. . This acts like renewing the \arraystretch setting and remains in effect until changed. . . . \vspace* and \vfill. at the very beginning of your document. . . . . . You could write \renewcommand{\baselinestretch}{1. Write a paragraph in article [and letter] style with the following properties: . . you use \vspace*{2in} to put a 2 inch margin at the top (\vspace would not insert the space). \hspace would not insert the 1 inch. .3 in here. The \hspace command has no effect at a line boundary. ⇒ I insert . . . . but \hspace* does. .2. This gives you three commands: \singlespacing \onehalfspacing \doublespacing Right after you specify one of these. The most versatile method to insert horizontal spaces is with \hspace and \hspace*.2}\small\normalsize the associated postscript result (ps file). In particular.7 Spacing We have already seen the use of ~ to insert one space and \hfill to put the remaining text flush right. . . . vertical spacing is controlled by \vspace. The vertical space is not added if this goes to the top of a new page. I insert . . . if you want to make your own title page. but the \hspace* inserts the space no matter what. Submit a printed copy of both the L TEX source (tex file) and 1. . . For example. . That is why you see the 1 inch space at the beginning of the second line. . . . . . . . that is what \vspace* does. . so vspace{\baselineskip} skips one line at the next new line. However. . . Two variations of \hfill are: \dotfill . . specify \renewcommand{\baselinestretch}{1. . . .3~in \hspace{. one difference is that you need to change font size before this change goes into effect. For example. but the \verb|\hspace*| \hspace*{1in} inserts the space no matter what. however. . . . . . . .2} to increase the spacing modestly (actual increase depends upon the font size).7 Spacing 25 2. You can. . Be sure your name is on each. . A Exercises. \hrulefill Analogously. These have one argument: the amount of space to be inserted.3in} here. The height of one line of normal text is in the keyword \baselineskip. . . . that spacing will commence. the previous sentence is written as: The \verb|\hspace| command has no effect at a line boundary. . . . . . . .

Write a paragraph in article style and make a cover page with the following properties (like the cover page of this document): • All lines are centered. 4. • The title appears first in very large letters. 3. Write two paragraphs in article [and letter] style with each of the following properties: (a) Default indentation on both paragraphs. (d) There is added space between paragraphs. • Course number and title appears next. • Date appears last. • Your web address appears fourth. Produce the following table: Colors Primary Secondary Red Green Blue Orange Yellow Purple 6.26 2 TEXT (a) Each font style in Table 1 is used as one letter in a word that has more than one letter. and it is preceded by extra space. Give the same list without numbers. (c) Each font style in Table 1 is used for two consecutive complete words. . (b) Each font style in Table 1 is used for one complete word. • Your e-mail address appears third. with extra space preceding it. Give an enumeration of at least three things you like about mathematics. but larger than normal size. with extra space preceding it. Produce the following table (including the accents and alignments). 2. 5. (c) Both paragraphs are indented. (b) No paragraph is indented. • Your name appears second in letters that are not as large as the title.

Create a 3-column text such that each column is a paragraph of arbitrary length using about 1 of the page width each.7 Spacing Mathematician Gabrielle Emilie Le Tonnelier de Breteuil Marquise du Châtelet Benjamin Banneker Sophie Germain Julius König Rózsa Péter 7.2. Produce the following table. 3 10. How can you have an entire table whose columns are of fixed width? 9. . Use the tabbing environment to produce the following: apples integral derivative grapefruit sum difference variables constants 11. 12. Payoffs ($) Player A 1 4 2 5 3 6 Player B 1 3 5 2 4 6 Birth Death 27 1706 1731 1776 1849 1905 1749 1806 1831 1913 1977 8. Produce the following: rate of mass accumulation in the compartment net rate of mass entering the = compartment by convection net rate of + mass entering by diffusion. Use the tabbing environment to produce what you see on page 30.

suppose your bibliography is in a file called mybiblio.pdf Figure 31: Adding bibtex to the Command Sequence 3. Once this is successful. instead of with §7.bib. The bibtex program that you apply to your source creates another file (which you need not examine). The bib file contains the bibliographic database.dvi dvipdf bibtex print/post myfile. So here we are.4. That is because I require students to produce an annotated bibliography early in the semester. and I want them to use BibTE X. from which a second latex compilation causes the bibliography to be created. called a bib file (plus one of the files created by the latex compiler.1 The bib File Main body For purpose of this introduction. which could extend beyond one document.tex view/print convert dvips myfile. about which you need not be concerned). until you do not have any “unresolved” bibliography citations.1 Bibliography with BibTE X Overview It might seem strange to have this section so early. but that name is arbitrary as long as it ends with .ps myfile. This added loop is illustrated in Figure 31. The execution looks like this (same under unix and DOS): latex myfile bibtex myfile latex myfile You might have to compile with latex more times.bib. create/edit compile latex myfile. We .2. BibTE X [11] was developed by Oren Ptashnik and is available free of charge. It reads a plain text file. you do not have to bibtex myfile again until you change your bib file or add a citation.2 3.28 3 BIBLIOGRAPHY WITH BIBTE X 3 3.

b. if there are two publications by the same authors in the same year. labels are case-sensitive. A • The label is arbitrary. Before listing each style (article is one style) and the fields they can or must have (author is one field).89 knuth:tex With two authors. .) In the first form. • Each bib entry must have a unique label. which you might consider: Form author . but do not use any L TEX special characters or blanks. (Linguistically. You must discover what style works best for you. MA". We cannot. here are a few things to note. } Most authors develop a style to labeling bibliographic entries. In the second form. with more than 2. address = "Reading.3. if there are two publications by the same authors in the same year. so it can be cited without ambiguity in the source file. so tex is not the same as TeX. have two entries with tex as their label. edition = "15th". The use of one keyword is somewhat simplistic and could cause problems with a great number of entries because the labels must be unique. . . the label is specified as tex and it must be followed by a comma. you can put both of their names. . after the year (no blank).year author :first_keyword_in_title Example knuth. some people add another keyword. title = "The {\TeX} Book". Here are two styles I have seen. . . Also. in formal writing follows this rule. author = "Donald E. field = "value". Knuth". In the example. which are the entries you want to include in its database. Knuth’s book [8] would be entered as follows: @article{tex. Each entry has the following form: @type {label . year = "1989". } For example. the use of the Latin et al.2 The bib File 29 begin with the most important part of your bib file. publisher = "Addison-Wesley Publishing Company". you can add et al. for example. . some people add a.

}. year. address. month. pages. article refers to an article from a journal or magazine. even mathematical symbols in math mode. chapter series. • There is a final } to close the entry — @type{ . publisher. first. address. chapter and/or pages. Required fields: author or editor. and fields are separated by commas (hence the comma after the terminal quote).) Fields that are neither required nor optional are ignored. title. . Required fields: author. edition. address. . incollection is a part of a book having its own title. year. pages. For example. • Fields do not have to be on separate lines. the volume and page numbers of an article are necessary to include even though they are optional to satisfy BibTE X syntax. number. volume or number. publisher. month. but without an explicit publisher. and it gives flexibility if you want to add a field or change the order. (What is “necessary” depends upon the standard one applies. If you put Knuth Donald. What are “optional fields” in BibTE X are not necessarily optional as far as having a complete bibliography citation. Here is a list of standard entry types with their required fields. address. series. A • The field value can be anything recognized by L TEX. title. Optional fields: volume or number. publisher. but it is more readable that way. month. howpublished. month. title. year. Remember to put each author’s name as first last or last. Required fields: author. book refers to a book with an explicit publisher. type. . but most journals require the volume of the journal and the page numbers for the cited article. Optional fields: volume. year. The last field does not require a comma at the end. the compiler will think the first name is Knuth and the last name is Donald. Optional fields: volume or number. inbook is a part of a book. type. Required fields: author or key. edition. series. Optional fields: author. year. printed document. edition. title. but it will not hurt anything. Optional fields: editor. such as a chapter or just some range of pages. month. even if they are valid fields in other types of entries. booktitle. journal. booklet refers to a bound.30 3 BIBLIOGRAPHY WITH BIBTE X • The order of the fields is arbitrary. title. Required fields: author or editor.

. edition. For example. Required fields: author or key (see note below). address. title. but it is best if you provide the missing field value. The bibliography will be sorted with the key. misc is when nothing else fits. Required fields: author or key (see note below). year. number. title. address. mastersthesis is a Master’s thesis. the note field is always an option. booktitle. year. publisher. unpublished is a document with an author and title.2 The bib File inproceedings is an article in a conference proceedings. title. Optional fields: type.. title. year. publisher. Required fields: author. Required fields: author. year. This lets you enter a note that will appear at the end of the citation. you must provide a key for sorting. phdthesis is a PhD thesis.3. month. school. If a document has no author. consider the following entry for A L TEX 2ε [10]. year.". year. which has no person as an author. title. used to order this entry relative to author . volume or number. month. Optional fields: type. note. Optional fields: editor. month. series. Optional fields: author. address. pages. month. organization. school. enter an unrecognized field. series. Required fields: author. title. Optional fields: type. such as comment = ". LaTeX2e. Optional fields: editor. Optional fields: month. To have a comment that is not printed.) Required fields: author. proceedings Required fields: title. year. Required fields: author. address. (this is ignored with no error message given). which vary by the type of entry. (Some note of explanation is required. address. howpublished. If a required field is missing when you compile. Possibly there will be some standard fixup. you will get an error message. Optional fields: author. but not formally published. year. address. 31 In addition to the optional fields listed. institution. month. organization. volume or number. month. even as a technical report. techreport is a report published by some institution. title. organization. month. manual is some technical documentation.

author = "Michel Goosens and Frank Mittelbach and Alexander Samarin". title = "The {\LaTeX} Companion". type = "World Wide Web site". key = "LaTeX2e". publisher = "Addison-Wesley Publishing Company". } The use of the braces in {\LaTeX} is to tell the bibtex program to take everything inside just as it is written (for the latex program to process). Some authors. otherwise. such as writing {F}ourier analysis to force the capital F. [5] in this document has the following BibTE X entry: @Book{companion. Now it is a major source of information. If it is a book. use the book type and specify: . we write G{\"{o}}del. The key will not be printed. 3. or simply G{\"o}del. MA". Otherwise. There is no universally accepted standard for how to reference web documents.32 names. In ordinary L TEX. For example. except the first letter of the first word). however. but here is one way.ps} (see~\cite{CTAN} for replacing CTAN)}". we separate them with and (no commas). address = "{CTAN\url{/macros/latex/doc/usrguide.2 Web citations When BibTE X was developed. we write G\"{o}del to produce Gödel. year = "1995--99". This applies to accents too. year = "1994". 3 BIBLIOGRAPHY WITH BIBTE X @manual{usrguide. That defeats one of the primary advantages of using A L TEX and BibTE X in the first place: we want to let the style files determine the final form. the bibtex program might try to process it itself and produce an A unintended result. The use of braces to force a particular result is necessary in other instances. institution = "Comprehensive {\TeX} Archive". title = "{\LaTeXe} for authors". the bibtex program will produce ‘fourier analysis’ (the plain style produces article titles in lower case. but this will not work in BibTE X. Instead. use this feature inappropriately by putting braces around everything. so we can switch styles and use the same source (tex and bib files). the World Wide Web did not exist. address = "Reading. } When there are multiple authors.2.

William Strunk’. . author = "William Strunk{. A The \url specification is not actually an intrinsic command in L TEX. and it would appear as ‘Jr. One A example is the L TEX 2ε reference [2].2 The bib File publisher = "World Wide Web". To have the \url command active in your document.edu/acis/bartleby/strunk/".columbia. Braams and David P. respectively. without the braces.’ is the first name of the author.3. Eventually. and with latex having no place to break. given by: @misc{latex2e. The default font it uses is tt. address = "url ". Also. ~ is in many urls. year = "1999". if needed. These difficulties are overcome by specifying: address = "\url{http://www. If the document is a technical report. Jr. you will see a line with lots of spaces (for justification). you will run into some difficulty with writing urls. note = "This is the web version of the classic book by Strunk and White~\cite{StrunkWhite}". For one thing. use that style but include the url as a note or in the address field. } The reference \cite{StrunkWhite} presumes there is the entry for the original publication. Its main use is to determine where the url can be broken in order to put it on two lines. publisher = "World Wide Web". Another feature of the url package is that \url prints special characters. but you can change this to TimesRoman or sans serif font with the specification: \urlstyle{rm} or \urlstyle{sf}. author = "Johannes L. and writing it will produce a space. a url could become very long. Otherwise. not the tilde. There are occasions when we want to reference an entire web site. Jr. like ~. address = "http://www. in particular. An unsightly line with spaces could also appear after the url.}". 33 Here is an example: @book{Strunk. Carlisle and Alan Jeffrey and Frank Mittelbach and Chris Rowley and Rainer Sch{\" o}pf". title = "Elements of Style". it is defined in a package. the comma would signal the bibtex program that ‘Jr. The use of the braces in the name is to be sure that the author appears as intended: William Strunk. put the following declaration into your preamble: \usepackage{url}.edu/acis/bartleby/strunk/}". the url could contain special characters. title = "{\LaTeXe} and the {LaTeX}3 Project". followed by the url.columbia.

title = "My MotherMy Father". we merely change the one string value and recompile. when we wrote literals. where I describe enhancements for having graphics in L TEX. we can define strings with the entry: @string{name = "string"} Then. However.” Besides consistency.) For example. title = mom # dad.34 3 BIBLIOGRAPHY WITH BIBTE X howpublished = "World Wide Web. editor = dad.html}". For example. (That is why we needed the quotes before.2. an advantage is that if some name changes. . 3.” To help be consistent and to save some work in the long run when we write many different documents and produce more bib files.latex-project/org/latex3. about which I shall say more when I describe ways to customize your document in §??. and you shall learn more about A packages in §6. } We have seen several packages so far.” other times “Kluwer Academic Publishers. One sometimes sees “Kluwer. the three field values are equivalent to: author = "My Mother". Then. We can concatenate strings and/or literals with #. suppose we define: @string{kluwer = "Kluwer Academic Publishers"} Then. suppose we write @string( mom = "My Mother" ) @string( dad = "My Father" ) author = mom. we can enter: publisher = kluwer. including abbreviations and names of publishers. editor = "My Father". \url{http://www. year = "1994". to produce the publisher value = “Kluwer Academic Publishers.3 Additional features One element of good style is to be consistent in your terms. we can refer to the string anywhere in the value of a field by excluding the quotes. this is the first use of \renewcommand.” and still other times “Kluwer Pub.

use the space character. as a literal: title = mom # "~" # dad. pages = "311--316". Kluwer Academic Press. Here is an example [11] that is useful for guiding the sorting of references in a special circumstance: . Byrnes. title = "Results on the Absolutely Convergent Series of Functions and of Distributions". To ensure a space. editor = "J. author = "N. crossref = "Byrnes:FAA-89".2 The bib File 35 Note the absence of a space between the string values in the title. The general form is @Preamble{ string } where string is any concatenation of literals and strings. Recent Advances in Fourier Analysis and its Applications: Proceedings of the NATO Advanced Study Institute.K.S. pages 311–316. Results on the absolutely convergent series of functions and of distributions. [2] BibTE X also recognizes a preamble in our bib files to enable us to define A some L TEX commands. editors. ~. } If these were the only references. For example. J. the result would appear as follows: [1] N. = mom # " My Father". Byrnes and Jennifer L. suppose we have the following entry (kluwer is a string. } Then. Art{\’e}miadis".3. Artémiadis. In Byrnes and Byrnes [2]. The same title as the above is obtained by any of the following: title title = "My Mother " # dad. year = 1990. Byrnes and Jennifer L. we can have the following entry: @InProceedings{Artemiadis:FAA-89-311. 1990.K.S. title = "Recent Advances in {F}ourier Analysis and its Applications: Proceedings of the {NATO} {A}dvanced {S}tudy {I}nstitute". Another useful feature of BibTE X is the crossref field for cross referencing. publisher = kluwer. Byrnes". the other values are literals): @Proceedings{Byrnes:FAA-89.

Here is how this can be used.3 Declaration and Citation At the end of your source file (where you want the bibliography to appear). and they are labeled with numbers. requiring one argument. Command \noopsort ignores the argument that it receives. alpha differs from plain by citing by labels. originally published 1971. There are other bibliography format styles. producing nothing (indicated by {}). we fool the bibtex program with the following specifications: Volume 1 year = "{\noopsort{a}}1973". unsrt differs from plain by sorting entries by the order in which they are cited. rather than numbers. The bib entries would have the years in the opposite order than we want because sorting is first by the authors. Here is a list of the most basic ones (included in every installation): plain is the most common because it formats entries according to accepted standards. but a second edition of volume 1 is printed in 1973. breaking ties with the year of publication. Entries are sorted by the alphabetical order of author names.36 3 BIBLIOGRAPHY WITH BIBTE X @Preamble{ "\newcommand{\noopsort}[1]{}" } The \newcommand is something I shall describe more fully in §??. it is used to define a command. The definition of \noopsort. rather than by the author names. then by year. mybiblio. To force the first volume to sort before the second. For now. among other things (to give a more compact bibliography). including some provided by publishers. however. This fools the bibtex program into thinking the years are a1973 and b1971. which are the same. The second command defines the format style of the bibliography to be plain. thus putting volume 1 first. Suppose there is a 2-volume work by the same authors. so just the years appear. does not actually print the letters. Volume 2 year = "{\noopsort{b}}1971". abbrv differs from plain by abbreviating names of journals. . \noopsort. before \end{document}. which comes with every installation of latex. put the following commands (in either order): \bibliography{mybiblio} \bibliographystyle{plain} The first declares the bibliography to be in the bib file.bib. 3.

. . 46] in this document. but we do not want to cite it in the text. There are times when we want to be sure a particular bib entry appears. If we have the same label in both bib files. the entries must be identical. If we want only some particular list of entries to appear. the L TEX command is \cite{label [. (In the option. 3. The reason is that we can maintain one large bib file and write many documents that use it. such as: \bibliography{mybiblio.4 Some Controls You might want to have the bibliography single-spaced. we will get a fatal error message.~46]{latex} produces [9. \cite{tex. For example. Figure 32 shows a complete source file for having all entries in mybiblio. separated by commas. even if it is not cited explicitly. otherwise.bib appear. In particular. and that is the entire document! \documentclass[12pt]{article} \begin{document} \nocite{*} \bibliographystyle{plain} \bibliography{mybiblio} \end{document} Figure 32: A Document to Print the Bibliographic Database We can specify more than one bib file. For example.) The rule is that only those bib entries that are cited appear in the final document.4 Some Controls 37 We shall use only the plain style here. 9] for this document. they will appear twice if both labels are used (or if we used \nocite{*}). such as \nocite{tex} to be sure Knuth’s TEX book appears. [8] is produced by specifying \cite{tex}. even if the main document is spaced differently. but know that many other styles have been written and are available free of charge. 46” in the citation. the ~ is used to ensure that there is a space but no line break when giving the page number as “p.3. You can put more than one citation. \cite[p. we use \nocite with their labels. This is done with the \nocite command. A To cite particular references. we specify \nocite{*}. If we have the same entry with different labels. ]}. p. You can insert some further citation information as an optional input argument to the \cite command. Repeated entry– telling us which label is repeated. delimited by [ ].another} The bibtex program will search them sequentially for any citation. Just before the \bibliography statement.latex} produces [8. if we want to have every entry in our bib file appear. For example. where label is what we put in our bib file entry.. put \renewcommand{\baselinestretch}{1} \selectfont .

M. (c) An entire web site. (c) A chapter in a book. (d) A technical report. Be sure your name is on each. volume II. such as the heading. Volume II of Rich [2]. you could specify \renewcommand{\baselinestretch}{1} \small to have the bibliography set in small font size (and single spaced).M. (b) A book on the web. which is available at CTAN [4]. 4. 1990. Produce a document that has only a bibliography composed of the following three entries (in the order shown). one for each of the following types: (a) A technical report on the web. U. Impossible Dreams. How to Square a Circle. and you can change the properties. 1990.38 3 BIBLIOGRAPHY WITH BIBTE X The \selectfont command is needed to activate the change in \baselinestretch. volume I. The default is Bibliography. A Exercises. Smart. at least one entry must have more than two authors. and V. Impossible Dreams. MacTaco. If you also want to change the font size (or style). Further. Produce a document with one paragraph that cites three bibliographic items. This will let you put different bibliographies throughout your document.M. the BibTE X data (bib file). I. 2. You can change the heading put by bibtex. Rich. Submit a printed copy of the L TEX source (tex file). Produce a document that lists your entire database. that will do. which consists of at least one entry for each of six different document types. (b) An entire book with at least three authors. editor. . Rich. 3. [1] I. but you might want to have it be References. chapter 1. Thus. second edition. and the associated postscript result (ps file). Produce a document with one paragraph that cites three bibliographic items. one for each of the following types: (a) An article in a journal. editor. Tu.F. MacTaco. Money. You do this by specifying (sometime before the bibliography): \renewcommand{\refname}{References} for article style \renewcommand{\bibname}{References} for book and report styles You can have more than one bibliography by using the bibunits package. [2] [3] I. 1999.R. 1. \selectfont is not needed.

How To Solve It. A where label is unique in the document. except do not use L TEX special characters or blanks. NJ. \thepage produces the page number.1 Counters. NJ. specify \value{counter }. A label is the identification of A a particular value. The next 100 reveal lack of understanding the first 100. This is a good book. Here I illustrate some of those that are in all document styles. Princeton. . without a label. theorem-proving — process. There are many editions. I shall introduce the figure and table environments. Similarly.. Halmos. Produce an annotated bibliography of the following form (note the indentations on left and right margins): [1] P. Van Nostrand. and there are modern descendants. The L TEX syntax for labeling a counter is \label{label }. . This is a seminal book that articulates the problem-solving — i. Then. and a reference is a citation to a label. Naive Set Theory. There are times when you just want to produce the counter value. 1945. figures. whose numbers I could write by \thesubsection\ and \thesection. You are looking at page 39. the following labels were defined (when the subsection and subsubsection were first written): . Labels. [2] G. in this book I defined: \section{Bibliography with \Bibtex} \label{sec:Bibliography} Now I can refer to §3 by \S\ref{sec:Bibliography}. Princeton University Press.R. The first 100 pages seem simple. Polya. such as . if you want to use the counter’s numerical value as an argument in a command.D. which I assign to my Ph. placed where the counter’s value is set. For example. To illustrate how I can reference other parts of this document. respectively. can be labelled and referenced in the same manner. For example.2 Intrinsic Counters A Anything to which L TEX assigns a number has a counter associated with it. which have intrinsic counters associated with them. This is done by \thecounter. The L TEX syntax for referencing a label is \ref{label }. and References Basic Concepts A counter is a numerical value that refers to something that is being numbered. like a chapter in a book. which I was able to print by writing \thepage. Princeton. students. sections. On the other hand. 4. just as the labels in the bib file entries.2 of section 4. 4 4. such as pages.e. The choice of label is arA bitrary. 1960. and equations.39 5. you are reading subsection 4. In the next section I describe intrinsic counters and illustrate how to label and reference them. Counters that depend upon the style.

I entered the label \label{exer:likeaboutmath} (page 26). is any string you want to use A that does not contain embedded blanks or special characters used by L TEX. and I used the prefix subsec here. page 60. (Recall from p. if present. . Some people use this same form but with different prefixes. I use the prefix eqn.2. 28 For any counter. (If you have a lot of labels and need to keep track of them by printing each label and citation in your drafts. gives the page number where its label is defined. was labelled \label{eqn:hessian}. .” or floats. LABELS. That is a matter of style. For that reason they are called “floating objects. The four choices . (If you put it outside the caption. [\caption{caption[\label{label }]}] \end{figure} \begin{table}[options] [\caption{caption[\label{label }]}] .3 Figures and Tables which have the same In this section I describe figure and table environments. \pageref{counter }. In my choice of label. 24 that ~ is used to have a space without a linebreak. syntax: \begin{figure}[options] [\caption{caption[\label{label }]}] . which I can now reference as exercise #4 by writing \#\ref{exer:likeaboutmath}. which is an element of good style.~\pageref{subsec:bibfile} ⇒ p.~\pageref{subsubsec:webcite} ⇒ p. it will not be understood.) Because figures and tables are not split. The label to reference a figure or table is put inside the caption.) Equation (6). [\caption{caption[\label{label }]}] \end{table} The caption. even though you will get no error message. I can refer to these as follows: \S\ref{subsec:bibfile} ⇒ §3. I used the structure prefix :name. You can choose any labeling convention that is meaningful to you.40 4 COUNTERS. such as subsubsec:webcite. AND REFERENCES \subsection{The bib File} \label{subsec:bibfile} \subsubsection{Web citations} \label{subsubsec:webcite} Then. . . such as ss for subsection and e for equation. For equations. 32 p. so in this sentence I wrote its number by \ref{eqn:hessian} (with parenthesis added) and its page number by \pageref{eqn:hessian}. where you put it is where it will appear.2 \S\ref{subsubsec:webcite} ⇒ §3. The choice of label.2 I can also refer to their page numbers: p. their exact location depends upon how much room there is. This helps me to distinguish labels for different things. can go at the top or bottom. In the exercise to list what you like about mathematics. The environment options define where the float is to be located. see the showkeys package at CTAN [4].) 4. as given in [9]. just as \ref{counter } gives its value.

3 Figures and Tables 41 are shown in Table 6. notably pictures. perhaps on a page by itself. which insists that the float be placed here (note the capital H and no other option specified). which has no text. if possible. Locate at the top of the next page. Locate on a separate page. . it is to be located at the top of the following page. except that it also prints all remaining floating objects. The table environment is not to be confused with the tabular environment. Table 6: Figure and Table Location Options Option h t b p Meaning Locate here (where the environment is declared). if this page does not have enough room). and we generally use the table environment to present information in tabular form. As a matter of style.4. We might specify [ht] and find the float in an unexpected place. For example. This is done with the \clearpage command. This does the same as \newpage. Locate at the bottom of the page (or the next page. It is also advisable to specify \usepackage{float} in the preamble. followed by a figure or table. The latter makes tables. This option is used in many places in this book. I did this to avoid confusion by having some float appear pages after it is cited and discussed. and they have separate counters. called a float page. One cause could be an accumulation of floats that should be cleared at some point before continuing. The placement of a float is sometimes a source of frustration. Table number : caption That’s it. only figures and tables. In this document most tables and figures are specified with [ht]. The figures and tables in this document appear as the form: Figure number : caption vs. which is why you sometimes see pages with some blank space in the lower portion. neither of these A conditions is necessary for their L TEX environments. it differs from a figure only in its name.” the place where it is specified in the source. One of the enhancements is the placement option: [H]. However. we generally use the figure environment to present what we usually think of as figures. Floats can be framed. using the \fbox command. If there is not enough room. which means they are to be placed “here. but the table environment does not have to contain a table. Figures 33 and 34 illustrate how to frame a figure with a thick border.

42 4 COUNTERS. The initial value of the counter is 0. (This is called a local setting. . 4. AND REFERENCES \begin{figure}[ht] \begin{center} \setlength{\fboxrule}{3pt} % make border lines thick \setlength{\fboxsep}{. \newcounter{mycounter}[section] defines mycounter to be within the section counter. 1. \newcounter{mycounter} defines a counter whose name is mycounter. . Also note how the caption is put at the top (see exercise 3). . Figure 34: Framed Figure with Caption at Bottom The parameter settings have returned to their default values. the frame in Figure 35 has thin lines and no extra padding around the border. they will be 1. but you can specify the type of numeral. instead of the printed values being 1. . s. when printed within section s. .4 Defining Your Own In the preamble you can define your own counter with the \newcounter command: \newcounter{name}[within] where name is the (unique) name of the counter (cannot be the same as one of the intrinsic counter names). more generally. For example.) Thus. . .1. shown in Table 7. The counter values are printed in Arabic numerals. . Further.1.2in} % increase distance to border \fbox{ This is a framed figure. You can also define the counter to be within another counter.2. .2. within section 1. upon leaving the figure environment. For example. Figure 35: Framed Figure with Caption at Top This is framed with default parameter values. the values of mycounter will be of the form s. LABELS. 2. } \end{center} \caption{Framed Figure with Caption at Bottom \label{fig:fboxbottom}} \end{figure} Figure 33: Framed Figure 34 Source This is a framed figure. This will cause the value of mycounter to be reset to 0 when entering a new section. . .

1. to increment mycounter by 1 and establish a label to its value at the place this is done. . consider the enumerate list environment. 3. \dots ⇒ (i). This can also be used to transfer the value of one counter to another. . A. \setcounter{mycounter}{5} sets the value of mycounter to 5. which also increments its value. b. For example. . whose counter is enumii. . This is done with the \refstepcounter command. This can be used for intrinsic counters too. . IV. The default numeral type is arabic. III. so we must change them back if we want to restore the defaults. Counter values can be set to some absolute value with the \setcounter command. These changes remain in effect (called a global setting). \setcounter{mycounter}{\value{page}} sets the value of mycounter to the current page number (value of the intrinsic counter. . c. d. . The second level. (ii). but you can change the appearance to be any of those listed in Table 7 by applying the \renewcommand to \thecounter. page). For example. . such as illustrated in Figures 36 and 37. We can change these to be any type we want. For example. \setcounter{mycounter}{0} \renewcommand{\themycounter}{\roman{mycounter}} \stepcounter{mycounter} (\themycounter). write \refstepcounter{mycounter} \label{mylabel} Then. I. . . 4. we want to be able to label it for future reference. \addtocounter{mycounter}{1} adds 1 to the value of mycounter. . . iii. . D. where the types of numerals for the four levels are: arabic.4 Defining Your Own 43 Table 7: Numerals to Print Counters What you see a. If we just want to increment the counter by 1. .} (the “appearance” parameter is \labelenumii). II. had its label changed to what is specified in the source: \renewcommand{\labelenumii}{\theenumii. iv. C. shown in Table 8. ii. For example. When using a counter for some non-intrinsic sequence. we can specify \stepcounter{mycounter}. For example. roman and Alph. What you write \alph{mycounter} \Alph{mycounter} \arabic{mycounter} \roman{mycounter} \Roman{mycounter} Counter values can be incremented with the \addtocounter command. . . . alph. \stepcounter{mycounter} (\themycounter). . 2. For example.4. . we can use \ref{mylabel} and \pageref{mylabel} wherever we like. B. i.

Terms and Concepts A. Groups and fields B. AND REFERENCES \renewcommand{\theenumi}{\Roman{enumi}} \renewcommand{\theenumii}{\Alph{enumii}} \renewcommand{\labelenumii}{\theenumii. A Exercises. and use the \ref or \pageref command to reference each of the following. Write a document with at least two pages and two sections. (a) Reference §2 by a label that you assign to section 2 (make whatever . LABELS.44 4 COUNTERS. Introduction II.} \begin{enumerate} \item Introduction \item Terms and Concepts \begin{enumerate} \item Groups and fields \item Picnics and frolic \end{enumerate} \end{enumerate} % changes numeral type % changes appearance Figure 36: Alternative enumerate Symbols Source (Result in Figure 37) I. Be sure your name is on each. Submit a printed copy of the L TEX source (tex file) and printed copy 1. Picnics and frolic Figure 37: Alternative enumerate Symbols Result (Source in Figure 36) Table 8: Default Settings for enumerate Counters Counter enumi enumii enumiii enumiv What changes numeral label numeral label numeral label numeral label Command \renewcommand{\theenumi}{\arabic{enumi}} \renewcommand{\labelenumi}{(\theenumi)} \renewcommand{\theenumii}{\alph{enumii}} \renewcommand{\labelenumii}{(\theenumii)} \renewcommand{\theenumiii}{\roman{enumiii}} \renewcommand{\labelenumiii}{(\theenumiii)} \renewcommand{\theenumiv}{\Alph{enumiv}} \renewcommand{\labelenumiv}{(\theenumiv)} of the associated postscript result (ps file). Put an enumerated list of items near the beginning of your document.

like this: A consequence of Einstein’s postulates is that \[E = mc^2. 5. The $ delimiter keeps the mathematical expression in the text.45 label name you like).2 . .\] ⇒ A consequence of Einstein’s postulates is that E = mc2 . . signified by delimiters $.1 . Also reference the page that they appear. (c) Reference item #2 of your enumerated list. 2. Include two tables and figures in your document. Produce two numbered lists such that the second starts its numbering where the first leaves off. item 2 5 Math Mode One can write mathematical expressions by entering math mode. . . . .1 . 2. ⇒ A consequence of Einstein’s postulates is that E = mc2 . List 1. begin list 2. . . Produce lists using the enumerate environment with the following appearance: 1. Produce Figure 35. item 1 2. List 1. . like this: A consequence of Einstein’s postulates is that $E = mc^2$. . 1. . For example. 3. item 1 4. . . List 2. $ or \[. . 4. 2. The other form is math display mode. 1. 2. . \]. produce the following: 1. . List 2. . (b) Somewhere near the end of your document reference the page number of the first section. . 3. .2 . and reference them by label. . item 2 Now we are out of list 1 . .

Table 9 shows other common operations in math mode. but this is not universal notation. xa+b is written as $x_{a+b}$ and xa is written as $x^{a^2}$. The order of subscripts and superscripts does not matter: x_{a+b}^{c+d} ⇒ xc+d a+b x^{c+d}_{a+b} ⇒ xc+d a+b Table 10 shows some set notation.46 5 MATH MODE Table 9: Some Mathematical Operations Operation Symbol subscript _ superscript ^ multiply \times divide \div Example How it appears What you write x_3 x3 3 x x^3 a×b a\times b a÷b a\div b 5. produced by $\sim A$. (Each of the tables in this section applies only to math mode. respectively. Preceding any symbol by \not puts the line through the symbol. For example. some authors use Ac or A . produced by $\overline{A}$. produced by $A^c$ and $A^\prime$. as in the following examples: A\not\subseteq B x\not\in A\cup B A\setminus B\not\supset B ⇒ ⇒ ⇒ A⊆B x∈A∪B A\B ⊃B Table 10: Set Notation What it is empty set intersection union set minus element in subset (proper) subset or equal superset (proper) superset or equal How it appears ∅ ∩ ∪ \ ∈ ⊂ ⊆ ⊃ ⊇ What you write \emptyset \cap \cup \setminus \in \subset \subseteq \supset \supseteq . and some use A.) The braces enclose an expression that can be used to define a more com2 plex operand. The complement of A often appears as ∼ A. ^.1 Mathematical Symbols The example also illustrates the use of the superscript operator.

For example. to produce α−β =∆−δ . the calligraphic fonts remain in effect: $\cal P = A + B$ ⇒ P = A + B. Note that \boldmath is surrounded by the braces. otherwise. The calligraphic alphabet looks like this (and it is available only in math mode): ABCDEFGHIJ KLMN OPQRST UVWX YZ. this applies only to letters. normal. cal. sf. tt} (analogous to the \textfont command. does not apply to math mode because math mode has its own. rm. Unlike \boldmath. \boldmath$A\supset B$ text $B\cup C$ ⇒ A ⊃ B text B ∪ C {\boldmath$A\supset B$} text $B\cup C$ ⇒ A ⊃ B text B ∪ C Within math mode. For example. it. \mathfont{expression}. For example. causing unintended results when applied to other symbols. math fonts would remain bold. we can control the font style of letters with the command. and they are specified by spelling them as keywords. {\boldmath $x^n+y^n=z^n$} ⇒ xn + y n = z n . Write ${\cal P} = A + B$ to produce P = A + B. as shown in Table 11. The following illustrates this. separate from text mode. {\Large $(x\div y) + z$} ⇒ (x÷y)+z . and returns to normal style in the second case. Font style. ˜ {\boldmath$\tilde A\times\vec{1}\otimes\overline{2}$} ⇒ A × 1 ⊗ 2 ˜ 1 $\mathbf{\tilde A\times\vec{1}\otimes\overline{2} }$ ⇒ A × ˜ ⊗ 2 Table 11 illustrates the outcome of each font for this expression: \mathfont{\tilde A\times\vec{1}\otimes\overline{2}} Table 11: The \mathfont Commands Font Style boldface calligraphic italic normal roman sans serif typewriter Command \mathbf \mathcal \mathit \mathnormal \mathrm \mathsf \mathtt Example Result ˜ 1 Aט⊗2 ˜ A×∞⊗∈ ˜ ×1 ⊗2 A ˜ ˜ A×⊗ ˜ 1 Aט⊗2 ˜ ט⊗2 A 1 ~×~⊗2 A 1 The calligraphic style applies only to capital letters. Greek letters are defined only in math mode.5. where font is one of: {bf. p. For example. digits and accents.1 Mathematical Symbols 47 You can control the size of the font by using the usual specification before entering math mode. You can make math fonts boldface by specifying \boldmath before entering math mode. however. 8). but not to special mathematical symbols. without the braces. even when leaving and re-entering. where B ∪ C is boldface in the first case.

between the integrand and dx. In the preamble specify \usepackage{bm}.) The \mathbf does not make Greek letters boldface. p. as 4 . “Making Greek letters is as easy as π (or Π)” (written $\pi$ or $\Pi$). The general form is \frac{numerator }{denominator }. Figures 40 and 41 illustrate this. In the case of the integrals. which uses the \sqrt and \prod functions: \[ \sqrt{\frac{\prod_{n=1}^N \left( \sum_{i\in I_n} x_i^n\right)} {\sqrt[3]{\sum_{i\in I_\infty} x_i}} } \] Figure 38: Variable Sizes Source (Result in Figure 39) N n=1 3 i∈In i∈I∞ xn i xi Figure 39: Variable Sizes Result (Source in Figure 38) Notice that even though it is written in math display mode. We could use \boldmath to achieve this. but there is a package that not only provides the boldface font. This inserts a thin space (compare the results by writing the expression with and without A the In L TEX. 43] says.2 Fractions and Variable Size Functionality To make fractions. symbols whose size you would want to adapt to expressions are gen\. We can make this appear x+y larger. 5. we use the \frac command: $\frac{x+y}{4}$. but you can force either of the two styles with the \displaystyle and \textstyle commands.\delta \]. but also produces proper spacing. Note how the sizes of the fractions adjust automatically. . Here is a more complex equation in math display mode: x2 + yα A= .\beta = \Delta . where the numerator and denominator can be any expression. and I present more examples below. note the use of \. As Lamport [9. (Not every Greek letter is included — see Appendix Table 36. then \bm{\beta} ⇒β. we could write $(x+y)/4$ to make (x + y)/4. Figures 38 and 39 illustrate this with another example. Some mathematical symbols adjust their size to fit the expression. by preceding the math mode with \large. \]. but if we want x+y 4 .). erally designed to do so.48 5 MATH MODE write \[ \alpha . 1 + x2η +1 written as \[ A = \frac{x^2+y_\alpha}{1+\frac{\eta}{x^2+1}}. the indices on the A sums and product appear as they would in line. L TEX compilers make judgments about the layout. Table 12 shows some of the most common of these.

as though it were in display mode.dx\right] parentheses () x 1+y {} xi i braces brackets ∞ 0 [] f (x) dx \[ \sqrt{\frac{\displaystyle \prod_{n=1}^N \left( \sum_{i\in I_n} x_i^n\right)} {\sqrt[3]{\displaystyle\sum_{i\in I_\infty} x_i}} } \] Figure 40: \displaystyle Source (Result in Figure 41) In text mode you can force the display style of placing these subscripts and superscripts on functions. math display mode does not always use displaystyle. using \textstyle and \displaystyle to override the default form for the mode.5. Figure 42 gives more examples to compare in line text and display mode.dx \left( \right) \left(\frac{x}{1+y} \right) \left\{ \right\} \left\{\sum_i x_i \right\} \left[ \right] \left[\int_0^\infty f(x)\. in particular. The “default” is not always predictable. as well as sizing the expression. .2 Fractions and Variable Size Functionality 49 Table 12: Variable Size Mathematical Operation Symbols Operation sum How it appears n What you write \sum xi i=1 \sum_{i=1}^n x_i \int integral b f (x) dx a \int_a^b f(x)\.

and a half space by \. write \[ (x\in A\Rightarrow x\in B) \Leftrightarrow (A\subseteq B). 0] = {x aj bi ≡ bi x ≤ 0} aj ∀y {x : x y} ⊂ A . Here is how each looks: \forall x \exists y ⇒ ∀x∃y \forall x\. shown in Appendix Table 35.\{x: x\not\prec y\} \not\subset {\cal A} ⇒ ⇒ ⇒ (−∞. including negative spacing. write \[ \forall x\exists y\ni [P(x)\wedge Q(y)]. Here are some examples: (-\infty. \exists y ⇒ ∀x ∃y \forall x\.50 5 MATH MODE N xn i n=1 3 i∈In xi i∈I∞ Figure 41: \displaystyle Result (Source in Figure 40) What to write in text mode \frac{x}{2} \displaystyle\frac{x}{2} \max_{x\in X} \displaystyle\max_{x\in X} Appearance x 2 What to write in display mode \textstyle\frac{x}{2} \frac{x}{2} \textstyle\max_{x\in X} \max_{x\in X} x 2 maxx∈X max x∈X Figure 42: Examples to Compare Text and Display Modes Table 13 shows symbols used in logical expressions.. to have (x ∈ A ⇒ x ∈ B) ⇔ (A ⊆ B). so we might want to add some spacing between terms. In math mode a full space is obtained by specifying \.0] = \{x\ni x \le 0\} a_j\prec b_i \equiv b_i \succ a_j \forall y\. \] The quantifiers in this last example seem a bit crowded. For example.) Table 14 shows some relations for ordered sets (besides those on the keyboard: < = >). \exists y ⇒ ∀x ∃y (There are other spacing commands. \] To have ∀x∃y [P (x) ∧ Q(y)].

. xi < 0 for all i = 1. Compare each of the following: xi < 0f oralli = 1. . written as $x_i < 0 for all i=1.3 Arrays and Equations The array environment is to math mode what tabular environment is to text mode. xi < 0 for all i = 1. . . . . 5. . and more. .\dots$ written as $x_i < 0 \mbox{ for all } i=1. . but we can also do the reverse with the \mbox command.\dots$ The first line points out that blanks mean nothing in math mode. It has the form: .3 Arrays and Equations 51 Table 13: Some Symbols in Logic Logical Term existential quantifier universal quantifier negation disjunction conjunction implication equivalence such that How it appears What you write ∃ \exists ∀ \forall ¬ \neg ∨ \vee ∧ \wedge → \rightarrow ⇒ \Rightarrow ⇔ \Leftrightarrow ≡ \equiv \ni Table 14: Order Relations Relation less than or equal greater than or equal not equal precedes precedes or equals succeeds succeeds or equals How it appears ≤ ≥ = What you write \le \ge \ne \prec \preceq \succ \succeq We have seen how to embed math mode into text. and all letters are in the math form of italic (not quite the same as the italic in text mode).5. The use of \mbox is particularly convenient in math display mode.\dots$ written as $x_i < 0$ for all $i=1. which I shall illustrate in the next section.

.7 (= x+y) ⇒ 7. .234567$ & $ 1 $ \\ $y$ & $-9.2 \\ y &=& 2.5 z = 7. using the math mode designation for each body entry: $. last row spec [\\ options] \end{array} 5 MATH MODE The column specifications and options are the same as in the tabular environment.5 \\ z &=& 7. otherwise. 7.7 \. Variable Current Value Limit x 1.2 \\ \hline \end{array} \] or \begin{center} \begin{tabular}{ccc} Variable & Current Value & Limit \\ \hline $x$ & $ 1. $.2 y = 2.87 −12.234567 1 y −9. . specifies a space.234567 & 1 \\ y & -9.52 \begin{array}{column specs}options first row spec \\ . . so it can be generated in either of two ways: with the tabular environment. (= x+y) \end{array} \] The \.87 $ & $-12.7(= x + y). but the body is in math mode.2 This can be generated by either of the following two ways: \[\begin{array}{ccc} \mbox{Variable} & \mbox{Current Value} & \mbox{Limit} \\ \hline x & 1.7 (= x + y) The above was produced by the following use of math display mode (which is always centered): \[ \begin{array}{lcl} x &=& 5. .87 & -12.2 $ \\ \hline \end{tabular} \end{center} You can align a series of equations to appear this way: x = 5. using \mbox for each header entry. The following table has text headers and math body. or with the array environment.

as above. there are times when we need to use more than one line for an ‘equation. which is the same as eqnarray.3 Arrays and Equations 53 Another environment is eqnarray. (Note that \ref gives just the number. let us change our mind easily as to whether or not to include equation numbers by simply adding or removing the * from the environment specification.’ in which case we need to suppress the numbering of all but one of the rows. . Figures 43 and 44 give an example. but each row is numbered: x = y = y z (1) (2) (Another difference is that the eqnarray environment uses displaystyle. however. The \nonumber command causes no number to be assigned to the first part of the second equation.) We use the eqnarray environment directly (without entering math display mode). but without the equation numbers. Further.5. This is like a 3-column array with specifications {lcl}. There is no apparent advantage to this since the same result can be produced by the ordinary array environment. It does.) The relation need not literally be an equation. anything could be used for the middle column. specialized to this column specification. parentheses are added. \begin{eqnarray} x &\mbox{is equal to}& y \\ y & \preceq & \frac{a+b+c+d}{\Psi} + \frac{e+f+g+h}{\Phi} + \nonumber \\ & & I+K+J+L \end{eqnarray} Figure 43: eqnarray Environment Source (Result in Figure 44) x is equal to y a+b+c+d e+f +g+h + + y Ψ Φ I +K +J +L (3) (4) Figure 44: eqnarray Environment Result (Source in Figure 43) There is also an eqnarray* environment. so the above is produced by the following: \begin{eqnarray} x &=& y \label{eqn:xy} \\ y &=& z \label{eqn:yz} \end{eqnarray} The \label statements are to illustrate that we can reference these by writing (\ref{eqn:xy}) to produce (1) and (\ref{eqn:yz}) to produce (2).

2 & 1. . numbered equation. which suppresses the equation numbering.0 & -2.3 21. Notice how the vertical line was drawn by the column specification. and the horizontal line separating the blocks is obtained by specifying \hline before the second row of the outer array. there is the equation environment. To illustrate.3 \\ 21. {c|c}. \end{equation*} Figure 45: Matrix Equation Source (Result in Figure 46)  Ax = 1.0 22.1 \\ \end{array} \right] \left( \begin{array}{cc} x_1 \\ x_2 \\ x_3 \end{array} \right). note how x is specified. Also. as illustrated in Figures 47 and 48.0 −2.2 1.54 5 MATH MODE For a single.1  x1  x2  . there is the equation* environment. Analogous to eqnarray*.1 & 1.1 1.0 & 22. Table 15: Transpose of a Vector What you write x’ x^t x^T x^{\mathsf{T}} x^{\mbox{\tiny $T$}} ⇒ ⇒ ⇒ ⇒ ⇒ How it appears x xt xT xT xT \begin{equation*} Ax^\prime = \left[ \begin{array}{rrr} 1. Table 15 shows other ways to denote the transpose of a vector. Figures 45 and 46 show how to present a matrix equation. x3 Figure 46: Matrix Equation Result (Source in Figure 45) Array environments can be nested. This poses no particular advantage over specifying eqnarray and merely entering one row (except that column separators (&) are not used in the equation environment).

such as writing $x = \fbox{y} + z$ to produce x = y + z.3 Arrays and Equations \[ \left[ \begin{array}{c|c} \begin{array}{ccc} A_{11} & A_{12} & A_{13} \\ A_{21} & A_{22} & A_{23} \end{array} & 0 \\ \hline 0 & \begin{array}{cc} B_{11} & B_{12} \\ B_{21} & B_{22} \end{array} \end{array} \right] \] 55 Figure 47: Nested Arrays Source (Result in Figure 48)  A11 A12 A13  A21 A22 A23   0  0 B11 B12 B21 B22    Figure 48: Nested Arrays Result (Source in Figure 47) We can enclose mathematical expressions in a box. rather than absolute measurements for spacing. causing an undesirable clash. Note how the line height does not adjust to the frame.dx &=& \displaystyle\frac{1}{\tau} \\ \\ &=& \displaystyle\oint_a^{b+c} \Psi(x)\.dx \end{array} $} ∞ 0 xe−τ x dx = 1 τ b+c ⇒ = a Ψ(x) dx We can use \fbox within math mode. which you might change. (In the longrun. sometimes used for emphasis. it is better to use parameters. \fbox{$ \begin{array}{lcl} \displaystyle\int_0^\infty xe^{-\tau x}\.2\baselineskip} after x = y + z causes extra vertical space equal to 20% of the value of \baselineskip. because the former takes into account the font size. which is the height of one line of normal text.) . In particular. For example. This could be overcome by putting a vertical space command just after the expression. like \baselineskip. putting \vspace{.5.

56
Now consider the following conditional   −1 0 f (x) =  1
A produced by the following L TEX code:

5 MATH MODE
assignment: if if if x < 0; x = 0; x > 0.

\[ f(x) = \left\{ \begin{array}{rll} -1 & \mbox{if} & x < 0; \\ 0 & \mbox{if} & x = 0; \\ 1 & \mbox{if} & x > 0. \end{array}\right. \] Note the use of \right. after the array. This is because \left and \right must balance — i.e., there must be an equal number of each. It is not necessary that the left symbol be related to the right one — i.e., \left\{ does not require \right\} to balance; any right symbol will do. The period is not printed in this case, used specifically for this purpose of balance. We have seen the use of \left and \right for brackets around a matrix. Now A the use of the \left L TEX command for conditional assignment raises related uses of the underbrace and overbrace. Figures 49 and 50 illustrate these, along with \overline, \underline, \widehat and \widetilde.
\[ \begin{array}{cc} \mbox{This sum has} \\ \mbox{an overbrace} \\ \overbrace{\overline{i\dots j} + \underline{k\cdots l}} & \underbrace{\widehat{xy} - \widetilde{ab}} \\ & \mbox{This difference} \\ & \mbox{has an underbrace} \end{array} \]

Figure 49: Horizontal Braces Source (Result in Figure 50)

This sum has an overbrace i...j + k···l xy − ab This difference has an underbrace Figure 50: Horizontal Braces Result (Source in Figure 49)
We often need to mix mathematical notation and text. We could use the tabular environment and specify in-line math mode where needed (with $), or we

5.3 Arrays and Equations

57

could use the array environment and use either the \mbox or \parbox (see p. 18) to enter the text. There are, however, some nuances to understand. Figures 51 and 52 show the problem with using \flushleft to make the text within the parbox flush left. (Try it with the default justify and you will see that the spacing gives a poor appearance.) The problem is that \flushleft skips a line, which ruins the alignment (even though [t] is specified). The solution is to use the \raggedright command, as shown in figures 53 and 54. In addition, the \raisebox command is used to lower the small matrix, giving it some space below the horizontal line.
\renewcommand{\arraystretch}{1.2} \begin{center} \begin{small} \begin{tabular}{lll} Matrix & Definition & Example \\ \hline \parbox[t]{.9in}{Covariance} & \parbox[t]{2in}{\flushleft $A_{ij} = E[(X_i-\mu_i)(X_j-\mu_j)]$, where $\{X_i\}$ are random variables, and $E[\cdot]$ is the expected value operator with $\mu_i=E(X_i)$. } & \parbox[t]{1.4in}{\scriptsize $\left[\begin{array}{rrrrr} \frac{1}{2}&0 \\ 0&\frac{1}{2} \end{array}\right]$ \flushleft for $X_2=X_1^2$ and $Pr[X_1=x]$ \\ $= \left\{ \begin{array}{lll} \fourth &\mbox{for}& x=-1~ \vspace{.05in} \\ \half &\mbox{for}& x=~~0~ \vspace{.05in}\\ \fourth &\mbox{for}& x=~~1. \end{array}\right.$ } \vspace{.1in} \\ \hline \end{tabular} \end{small} \end{flushleft} \renewcommand{\arraystretch}{1}

Figure 51: \flushleft in parbox Source (Result in Figure 52)
Matrix Covariance Aij = E[(Xi − µi )(Xj − µj )], where {Xi } are random variables, and E[·] is the expected value operator with µi = E(Xi ). Definition Example
1 2

0
1 2

0

2 for X2 = X1 and P r[X1 = x]   1 for x = −1   4  1 = for x = 0  2   1  for x = 1. 4

Figure 52: \flushleft in parbox Result (Source in Figure 51)

58

5 MATH MODE
\begin{center} \begin{small} \begin{tabular}{lll} Matrix & Definition & Example \\ \hline \parbox[t]{.9in}{Covariance} & \parbox[t]{2in}{\raggedright $A_{ij} = E[(X_i-\mu_i)(X_j-\mu_j)]$, where $\{X_i\}$ are random variables, and $E[\cdot]$ is the expected value operator with $\mu_i=E(X_i)$. } & \parbox[t]{1.4in}{\scriptsize\raisebox{-.1in}{ $\left[\begin{array}{rrrrr} \frac{1}{2}&0 \\ 0&\frac{1}{2} \end{array}\right]$} \flushleft for $X_2=X_1^2$ and $Pr[X_1=x]$ \\ $= \left\{ \begin{array}{lll} \fourth &\mbox{for}& x=-1~ \vspace{.05in} \\ \half &\mbox{for}& x=~~0~ \vspace{.05in}\\ \fourth &\mbox{for}& x=~~1. \end{array}\right.$ } \vspace{.1in} \\ \hline \end{tabular} \end{small} \end{center}

Figure 53: \raggedright in parbox Source (Result in Figure 54)
Matrix Covariance Definition Aij = E[(Xi − µi )(Xj − µj )], where {Xi } are random variables, and E[·] is the expected value operator with µi = E(Xi ). Example
1 2

0
1 2

0

2 for X2 = X1 and P r[X1 = x]   1  4 for x = −1   1 = for x = 0  2   1  for x = 1.

4

Figure 54: \raggedright in parbox Result (Source in Figure 53)

5.4

Special Functions and Alphabets

Math mode recognizes a collection of special functions. Table 16 shows some common ones. These special functions are used to make the source clearer, rather than using \mbox to achieve the same result. Among the special functions are the complete set of trigonometric functions. For example, we write $\tan\theta = \frac{\sin\theta}{\cos\theta}$ to prosin duce: tan θ = cos θ . Appendix Table 40 (p. 115) has a much longer list of special θ functions, as also a package of AMS symbols, of the examples shownyour preamble There is well as the arrows used in some which you declare in in Table 17. with \usepackage{amssymb}. This gives the following alphabet with the mathbb font: $\mathbb{ABCDEFGHIJKLMNOPQRSTUVWXYZ}$ ⇒ ABCDEFGHIJKLMNOPQRSTUVWXYZ

for which you specify \usepackage{mathrsfs} in the preamble. Table 18 shows how \mathbb can be used for specifying other numerical spaces.4 Special Functions and Alphabets 59 Table 16: Some Common Mathematical Functions Function limit lim inf log maximum tangent How it appears lim lim inf log max tan What you write \lim \liminf \log \max \tan Table 17: Examples of Mathematical Functions How it appears textstyle displaystyle limn→∞ xn lim inf n↓0 log xn maxx∈X f (x) tan(θ+π) ln x n→∞ n↓0 What you write \lim_{n\rightarrow\infty}x_n \liminf_{n\downarrow 0}\log\. . Table 18: Notation Using mathbb Fonts from amssymb Package What you write† \mathbb{R} \mathbb{C} \mathbb{Z} \mathbb{Q} † In math mode. ⇒ ⇒ ⇒ ⇒ How it appears R C Z Q What it means Real values Complex values Integer values Rational values Another alphabet is \mathscr. the real line is sometimes denoted by R. This gives the following alphabet: $\mathscr{ABCDEFGHIJKLMNOPQRSTUVWXYZ}$ ⇒A BC DE F G H I J K L M N OPQRS T U V W X Y Z In particular. $\Re$.x_n \max_{x\in X}f(x) \frac{\tan(\theta + \pi)}{\ln\. L is often used to denote the Laplace transform or the Lagrangian.5. which is A the L TEX special symbol. rather than .x} lim xn lim inf log xn max f (x) x∈X tan(θ + π) ln x For example.

dx}\right| \. ∂. The partial derivative symbol. . . which is an upside down delta (introduced by A Hamilton in 1853). so you can write \partial f(x)/\partial x to produce ∂f (x)/∂x. which are both variablesize symbols. \end{array} \] There seems to be some crowding in this direct specification. . A I leave it as an exercise to show the L TEX code that produced equation (5). ∂xi ∂xj (6) There are two integral signs: \int⇒ and \oint⇒ . In L TEX it is produced by \nabla. is written \partial. df (x)/dx. or.) 5. \] .dx} {\oint_{X(v)} e^{\lambda f(x)}\.5 Derivatives and Integrals We can express a total derivative. we can use (x) the \frac command to produce dfdx . which is also used by some authors. and its mathematical definition is the vector of first partial derivatives: f (x) = (∂f (x)/∂x1 . The usual notation for the gradient of a function is the nabla. by writing $df(x)/dx$. and {\large$\frac{\partial f(x)}{\partial x}$} to produce ∂x . eλf (x) dx a λ→∞ X(v) This was obtained by the following code: \[ \int_a^b \lim_{\lambda\rightarrow\infty} \left| \frac{\oint_{X(v)} xe^{\lambda f(x)}\. For example.dv .\Phi(v)\. (also called “del”). note how the outer integral is large in the following expression: b xeλf (x) dx X(v) lim Φ(v) dv. ∂f (x)/∂xn ). The Hessian is the matrix of second partial derivatives: 2 ∂f (x) (5) f (x) = ∂ 2 f (x) . Compare with the following and see if you can produce it: 2 f (x) = ∂ 2 f (x) . ∂xi ∂xj This was produced by the following code: \[ \begin{array}{lll} \nabla^2f(x) &=& \left[ \displaystyle \frac{\partial^2f(x)} {\partial x_i \partial x_j} \right]. (Compare with \mathcal. denoted by the symbol . . H.60 5 MATH MODE and H is sometimes used to denote the Hamiltonian.

The domains of integration.x_n)\. are better with A the following. like “Theorem.1. xn ) dx1 · · · dxn write \[ \int_0^\infty \int_0^{x_n} \int_0^{x_{n-1}} \cdots \int_0^{x_2} H(x_1. This is so fundamental that L TEX has the facility to define a special environment that includes a keyword. The rules create theorems.6 Theorems and Definitions (Note the use of the thin space. 223]): \usepackage{amsmath} in the preamble (see The L TE (u v − v u) · dS = S τ (u · v−v · u) dτ. .. z ∈ Z++ .6 Theorems and Definitions The foundations of mathematics are axioms and rules of inference. y.5. and we have the counter value: \thetheorem=5. the theorem was produced by the following L TEX code: \begin{theorem} For $n > 2$. there is no solution to $x^n + y^n = z^n$ for \newline $x.) Definite multiple integrals are no problem. \.dx_1\cdots dx_n \] However. but is also the name of the associated counter. . but by specifying A X Companion [5.1” appears.y. To have ∞ 0 0 xn 0 xn−1 x2 61 ··· 0 H(x1 . p. Consider the following example: Theorem 5. there is no solution to xn + y n = z n for x. Notice how “Theorem 5.” and a name. \end{theorem} . . This was defined in the preamble by: \newtheorem{theorem}{Theorem}[section] A Then. . which are statements whose truths are established relative to the A underlying logic. which is not produced by standard L TEX 2ε . and the spacing of the integral signs.\dots. consider the following: (u v − v u) · dS = S τ (u · v−v · u) dτ. Note how the domains are centered on the multiple integrals and the spacing of the integral signs. all text is in italic.z\in \LZ_{++}$. which is not only the name of the environment. 5.1 For n > 2.

Then. rather than Theorem 1. It was defined in the preamble as follows: \newtheorem{corollary}{Corollary}[theorem] Note that this is within the theorem counter. The environment created by \newtheorem puts the text in italics. so it must be different from all other environment and counter names. here is a corollary environment: Corollary 5. In this document. but this is generally not desirable for a definition. Consider the following example: Definition 5. and a unique name for the environment. \end{axiom} The label allows us to refer to the Axiom of Choice as ‘Axiom 1 on page 62’ by writing Axiom~\ref{axm:choice} on page~\pageref{axm:choice}. This was created by the following code: \begin{axiom} \label{axm:choice} From any (infinite) family of sets a new set can be created that contains exactly one element from each set in the family. \end{corollary} The following creates an axiom environment that is not within any other counter. so you see Theorem 5.62 5 MATH MODE Other theorem-like environments can be defined to have the same properties.1. which can be intrinsic or some other counter defined by the \newcounter command (p.1 The circumference of a sphere is the circumference of any great circle on the sphere. I defined the theorem environment to be numbered within the section. which is valid by having been defined by its own \newtheorem. The within option defines the counter to be within some other. 42) or by some other \newtheorem. This requires both a keyword. Here is the syntax: \newtheorem{name}{keyword }[within] The name defines the environment name. and it defines a counter. To further illustrate. like theorem. the above corollary was written as: \begin{corollary} The sum of cubes cannot be a cube.1. like Theorem. also used as a counter. This was created by first entering (in the preamble): .1 The sum of cubes cannot be a cube. \newtheorem{axiom}{Axiom} The “Axiom of Choice” can then be stated thusly: Axiom 1 From any (infinite) family of sets a new set can be created that contains exactly one element from each set in the family.

\end{defn} Compare this with the following: 63 Definition 5.5.1 The circumference of a sphere is the circumference of any great circle on the sphere. delimiter size control commands. $\bigg\langleE=mc^2\bigg\rangle$ ⇒ E = mc2 . in the text: \begin{defn} The circumference of a sphere is the circumference of any great circle on the sphere. we can also enlarge these delimiters ourselves. For example.7 Refinements Mathematical delimiters. $\big(E=mc^2\big)$ ⇒ E = mc2 . must be varied to enclose some expressions. . in the text: \begin{mydefn} The \textit{circumference of a sphere} is the circumference of any great circle on the sphere. . . however.7 Refinements \newtheorem{defn}{Definition}[section] Then. like parentheses and braces. {\Large[}$E=mc^2${\Large]} ⇒[E = mc2 ]. 5. the theorem package enables a wide range of variations over the font style (among other things). and \Bigg. especially the thicknesses. \Big. (large↔big. The use of text font environments comes close to the corresponding math size. . This is more evident with the square and angular brackets: $\Big[E=mc^2\Big]$ ⇒ E = mc2 . but they are different. One way is with a size command — for example. \bigg. There are. This was created by first entering (in the preamble): {\theorembodyfont{\rmfamily} \newtheorem{mydefn}{Definition}[section]} Then. Huge↔Bigg). Whereas \left and \right commands adjust the size of a mathematical delimiter to fit the enclosed expression. \end{mydefn} For more customization. {\large(}$E=mc^2${\large)} ⇒ (E = mc2 ). . which apply to a single character: \big.

. in particular.) This is not equivalent to preceding the array specification with a text size environment. While the letters inside the matrix are apc proximately the smallmatrix size. The same result with equation numbers is obtained by the gather environment. An alb ternative is with the amsmath smallmatrix environment: a d is obtained c by $\left(\begin{smallmatrix} a&b \\ c&d \end{smallmatrix}\right)$.64 5 MATH MODE {\LARGE$\langle$}$E=mc^2${\LARGE$\rangle$} ⇒ E = mc2 . in math mode. \\. The amsmath package has a command to put dots across any number of columns in an array. and n is the number of columns it spans. \begin{gather*} (a+b)^2 = a^2 + 2ab + b^2 \\ {\cal L} \oplus M^\varepsilon . where spacing determines the spacing between the dots. Its syntax is \hdotsfor[spacing]{n}. Figures 55 and 56 illustrate this. we could produce a b c d by spec- ifying $\left(\begin{array}{cc} a&b \\ c&d \end{array}\right)$. 61 for obtaining better multiple integrals). The gather and gather* environments allow the new line specification. the spacing and parentheses are not the same.V = H_0 \\ A(x) = \{y: \phi(y) = \cup_{a\in \cal A} \Psi(x)\} \end{gather*} Figure 55: gather* Environment Source (Result in Figure 56) (a + b)2 = a2 + 2ab + b2 L ⊕ M ε − V = H0 A(x) = {y : φ(y) = ∪a∈A Ψ(x)} Figure 56: gather* Environment Result (Source in Figure 55) When writing a matrix within text. b \scriptsize produces a d . For example. respectively. The remaining refinements use the amsmath package (introduced on p. (Note that there are no column specifications. except the equations are not aligned. They behave like the eqnarray and eqnarray* environments.

...7 Refinements \left|\begin{array}{ccccc} 1 & 2 & 3 & 4 & 5 \\ \hdotsfor{3} \\ & \hdotsfor{3} \\ \hdotsfor[2]{5} \\ \hdotsfor[... For example.. Figures 57 and 58 illustrate this... .. (Specify \usepackage{amscd} in the preamble.. the \overset and \underset amsmath commands enable us to put any characters over or under any character..... .. With more generality......... which makes it easy to draw commutative diagrams.....) The horizontal arrows are specified by @>>> (left to right) with any expression placed above or below. $\overset{a}{X}$ $\underset{b}{Y}$ $\overset{a}{\underset{b}{Z}}$ ⇒ ⇒ ⇒ X Y b a Z b a This can be used to stack subscripts: $\displaystyle{\sum_{ \stackrel{ \mbox{\scriptsize$i\in I$} } {j\in J} } } A_{ij} = \underset{j\in J} { \underset{i\in I}{\sum} } A_{ij} $ ⇒ i∈I j∈J Aij = i∈I j∈J Aij Nesting the \underset command can be unwieldy......5]{5} \\ \end{array}\right| 65 1 2 3 4 5 . The \stackrel command lets us put characters over a relation: For def example... . but you can use the smallmatrix environment or the \substack command: $\displaystyle\sum_{\substack{i\in I\\j\in J\\k\in K}} A_{ij} = \underset{\begin{smallmatrix} i\in I\\j\in J\\k\in K \end{smallmatrix}} {\sum} A_{ij}$ ⇒ i∈I j∈J k∈K Aij = i∈I j∈J k∈K Aij Another package in the ams family is amscd... The ... n^+\stackrel{\mathrm{def}{=}n+1 ⇒ n+ = n + 1.5....

where P is a permutation matrix. A symmetric rearrangement of a matrix has the following form: R = P t M P. Punctuate math display mode. and phrases are appropriately punctuated. people make some common errors. Many of these are described in The L TEX Companion [5.66 5 MATH MODE vertical arrows are specified by @VVV (down) or @AAA (up). 2. For example. \[ \begin{CD} A @>a>> \alpha @VV\beta V D @>>d> \end{CD} \] B @>>> \gamma @AAA E @>e>> C \\ @VVV\delta \\ F Figure 57: Commutative Diagram Source (Result in Figure 58) A −−→ B −−→ C −− −−      α β γ δ D −−→ E −−→ F −− −− d e a Figure 58: Commutative Diagram Result (Source in Figure 57) There are many more refinements. As you read articles notice that those that are among the most confusing are when the authors used a term that is not defined until pages later. and you can see an online catalog of packages at CTAN [4]. . For example. The general guide is to treat a mathematical expression linguistically. Define before use. 1. which is incorrect to omit. Here are some of the most common elements of grammar to consider. . 5. . and more packages to make things A nicer. note the colon before the display and the comma at its end. we might see “The distinguishing property of an abelian group is the commutivity . The expression usually needs a comma or period. clauses are separated by commas.8 Grammar When writing mathematical expressions. Chapter 8]. ” But a group had not yet been defined. with an expression placed to its left or right. All possible horizontal and vertical placements are illustrated. In English this means that every sentence has a subject and predicate.

(Lookup special symbols in the Appendix. B. . If . ‘If A. 2. The expression. √ (a) x2 = B 2 − 4AC implies x = ± B 2 − 4AC.’ The first form is preferred if A and C are simple expressions.’ should be written as ‘A if. (b) If ∆Fn+1 = Fn . The form. Produce each of the following formulas in line with text (construct your own sentences that contain them. and the comma is used to clarify where the antecedent (A) ends and the consequent (C) begins. . and it is important that the reader be told of this. instead of Φ(uk . if we write Φ = au + bv. Produce the following in math display mode with the array environment and/or with the eqnarray* environment. v) = au + bv. when there is no risk of confusion. then C. and only if. Φ(u.) 1. Be sure your name is on each. 6. this is not correct.’ A Exercises.5. If either A or C are compound. An object has only one definition. 0 otherwise.8 Grammar 67 3. Equivalence needs commas. . but you might want to take control over locating figures. and include proper punctuation).’ seems like it ought to be all right. 3. then . (c) x+ = x if x ≥ 0. we cannot later refer to Φ(u. then tell the reader something like. however. Suppose A and C are expressions. a A figure appears after its first reference. C. ∆2 Fn = Fn+2 − 2Fn+1 + Fn = 2Fn − Fn+1 . Then. (a) ln ex = x (b) sin{θ + 2π} = sin θ . is not correct. For example.’ or ‘Suppose A. ‘A if and only if B. Submit a printed copy of the L TEX source (tex file) and of the associated postscript result (ps file). 4. C. “We shall use Φk . . Produce each of the following in math display mode.” The overriding principle is clarity. For example. We can write either ‘If A. Sometimes we define the complete object. it follows that ∆2 Fn+1 = ∆Fn . In English. L TEX does this automatically. the second form is clearer. Reference object is located after the reference. v). v k ). 5.

(c) (−x 3 + a)2 dx. (b) α1 + β 2 − x2 + y 3 8. Produce each of the following expressions: (a) x = y mod n = x − y = kn for some k = 0.3 + a1 a2 a3 b1 b2 b3 = α −β γ −δ λ θ . 7.2 −1. 1. xT )T = (0. 1. 5. 1 . (a) A = {S ∈ S : S ∈ S} (b)  |F × P| ≤ π a11 a12 a21 a22 CB BC     ? def  (c)   9. 6. Produce the following equation in math display mode.1 −1. C T x)T . . . . √ A−1 (b) (0. (a) q ∗ (G) ≥ max ∆(G). xT ) √ 3 3 V = 2 a 0 A B 0 C = 0 xT C = (0. Produce each of the following formulas in math display mode (with punctuation): 2m(GA ) if G = ∅. Produce the expression in the Preface.2 2.68 (c) yn = (d) f (x) = n=0 n−1 i=i0 ∞ 5 MATH MODE xi ⇒ yn+1 − yn = xn − xi0 f (n) (0) xn n! ∂ (e) ∂x x2 a f (y)dy = 2xf (x2 ) (f) MFe(H2 O)6 = 6MH2 O + MFe 4. Produce each of the following in line with text (that you compose) and in math display mode.1 2.3 −2. Produce equations (5) and (6).

Produce the following: Definition 1 A matrix is singular if its determinate is zero. (d) Non-negative rational n-vectors: Qn . + 12. Definition 2 A matrix is non-singular if it is not singular. This implies it has an inverse. ∂f (x) ∂xj x=¯ . Produce the following: 13. (a) A key is how to add velocities the formula is (u + v) 1 + uv c2 where c is the velocity of light. What is grammatically wrong with each of the following segments. (b) A result of these assumptions is the following equation E = mc2 Einstein first noticed this equivalence between energy (E) and mass (m). Theorem 1 Every non-square matrix has an inverse. Combine your knowledge of derivatives. Therefore. (c) Complex n-vectors: Cn . ∂f (x)/∂xj } if xj = bj 11. conditional assignment (with array environment). Produce the following symbols: (a) Extended reals: R∞ . ∂f (x)/∂xj } if xj = aj     + ∂f (x)/∂xj if aj < xj < bj f (x)j =      min{0. (b) Strictly positive integers: Z++ . 14.5. and mathematical symbols to produce the following (called the truncated gradient):   max{0. x . the matrix is non-singular.8 Grammar 69 10. Proof: The determinate of a non-square matrix cannot be zero because it is not defined.

Note the row and column labels outside the matrix.  1 2  3 4 a 1 1 0 0 c 0 0 1 1 d 0 1 0 1 e 1 0 0 1     A= 17. where A is an m × n matrix and b is an m-vector. then n < 3. Row pointers: A= 11 12 21 22 ← rows in 1 ← rows in 2 (this arrow is closer to matrix) 11 12 21 22 ↑ ↑ columns columns in 1 in 2 18. Now suppose y(ω) is specified and we want to find x. (e) Theorem If x. and define y = Ax − ωb. Adding velocity vectors: u + v.70 5 MATH MODE (c) Let x be an n-vector and ω a scalar. z ∈ Z+ and xn + y n = z n . The remaining exercises are more difficult. You are to produce the mathematical expressions shown in math display mode. Column pointers: A= . u u+v v Figure 1. The following is tricky to get the evaluation expression. (d) Now we consider adding velocities. Figure 1 (above) shows how to add velocities simply as vectors. y. (1/2)1/4 16. 15. d f (x + t dt + 1 2 to be f (x)) t= 1 2 b 0 1 1 0 =− 2v + 1. t = the right size and location.

Use a package to import some standard graphics file. it lets us stack short phrases. and r = right. we can do this simply with \fbox and a long arrow in math mode. given as 2 cm for each box. but I do not provide a complete list of the relevant packages A (see CTAN [4] and The L TEX Companion [5]). \framebox also has two optional arguments to control the length of the box and the position of the text within it. For example. . However. 6. \framebox[2cm][l]{left}$\longrightarrow$\framebox[2cm][c]{center}$% $\longrightarrow$\framebox[2cm][r]{right} ⇒ left −→ center −→ right The % at the end of the first line is to avoid having a blank between the center box and the $\longrightarrow$ that follows it.1 Picture Environment If all we want is a series of boxes and arrows. as follows: \fbox{left}$\longrightarrow$\fbox{center}$\longrightarrow$\fbox{right} ⇒ left −→ center −→ right The \framebox command can be used instead of \fbox to produce the same result. notably the picture environment. c = center. The second optional argument is the position of the inscribed text: l = left. By itself. I illustrate each. 2. Use a graphics package to draw within the document. We can make the contents of a box obey all paragraph controls in text mode by the \parbox command. Use standard L TEX 2ε commands. 3. Row and column pointers: A= 11 12 21 22 ↑ ↑ columns columns in 1 in 2 ← rows in 1 ← rows in 2 6 Graphics A Graphics may be part of a L TEX document by one of three ways: A 1.71 19. The first optional argument of this \framebox command is the width of the box.

but there is a need for more versatility.72 6 GRAPHICS top like middle (note how the paragraph spacing adjusts). To begin. like ovals and diagonal arrows.b c   Z  Z Z ~  . respectively. Combined with bottom \framebox. and more control over positioning. whose source is shown in Figure 62. Figure 61 shows a more elaborate chart. \begin{center} \parbox{2cm}{ \framebox[2cm]{top} \\ \centerline{$\downarrow$} \\ \framebox[2cm]{middle} \\ \centerline{$\downarrow$} \\ \framebox[2cm]{bottom} } \end{center} Figure 59: Vertical Diagram Source (Result in Figure 60) top ↓ middle ↓ bottom Figure 60: Vertical Diagram Result (Source in Figure 59) The box created by \parbox has its center aligned with the text. v top left - center Z ?  1m a . along with other box commands. but it has an optional argument to align its top or bottom with the text. A basis for this is the picture environment. we can create vertical diagrams easily. These commands can be combined. which was created by the picture environment. This is done by specifying \parbox[t]{width}{text} or \parbox[b]{width}{text}. Going through its parts will serve to explain the various commands. as illustrated in Figures 59 and 60.

. This means that when . and I use the \setlength command to set the units of measurement to be 1 inch.oval  bottom right Figure 61: Variety of Objects in Picture Environment The first command begins a center environment.

and the default for the picture environment is 1 pt.-. In picture mode it enables control over not only the width.-.3)[tl]{top left} } \put( 1.3){center} } \put(-1.2.1}} ⇒ v The next three commands put three different kinds of boxes.-.0) \put( 0.-.-.5).-3){.35.3975}} \end{picture} \end{center} \vspace{1in} 73 Figure 62: Source for Figure 61 I specify some length = 5.1}} \put( 0.3}} \put(-.-1){\fbox{$\begin{array}{c}a\\b\\c\end{array}$}} \put(-. we enter the picture environment stating that the point of entry is the origin.7.85.0) is in this picture.3)[br]{bottom right} } \put(-. y) coordinates are relative to where the position is when the picture environment is entered.55}} \put(. it could p be a column in a table defined within the tabular environment.0){\circle*{. This could be at the left margin. The complete syntax is: \put(x.5.895){\line(1. y = −. -1){\circle{.0){.-.) The filled circle shows where (0. and the stuff is a filled circle with diameter .6.-1){.-1){\vector(1. (There is an alternative way to begin the picture environment.1}(1..5){\dashbox{. just as the smiley face appears here (see Exercise 1)..-1.-1.1 Picture Environment \begin{center} \setlength{\unitlength}{1in} \begin{picture}(0.05){oval} \put(0.0){.25)} \put(. I am specifying 5 inches. and this extends the position options to a second character: t = top. b = bottom. but also the height.65.32. The first is similar to \framebox in text mode.. as in beginning a paragraph with \noindent. but its syntax is different.4}} \put( .2. The .55}} \put( 0. indicated by the coordinates (0. or ph it could be in the middle of a sentence. The first \put in Figure 62 specifies the position at the origin.35){\vector(1.2}} \put(-. which is exclusively for the picture environment.1){.. The parameter that determines this is \unitlength.5){\dashbox{.7.7.5}} \put(-.e.5 inches below the origin (i.-1){\vector(-1.55.5){\framebox(. which is not described here.1 inches (centered at the origin): \put(0.01}(.5){\vector(4.65.0){.3.-. -1){\oval(.y){stuff }. where stuff can be text or some picture object. Every picture command begins with \put.0).-1){\vector(1. The (x. Then.0){. each beginning at . 0){\circle*{.5){\vector(0..43}} \put(1.05){1} \put( 1.

−1). The new point is determined by moving .height ? % width After the \oval specification.01 inches. Both \vector and \line have the same syntax: \line(∆x. has dimensions . This is the same as I used in text mode. Now we come to the \circle specification. Otherwise. If ∆y = 0. and some trial and error was needed to establish its position. The box length is set to 1.65. −1) (from \put). but that is not where we want to put the inscribed text to be centered. the line is vertical.65. the actual ∆y change in x is still len.5 measures the entire width: '  & $ 6 . the specifications are width = .3 inches. Now the code begins to draw the vectors. located at coordinates (−. if ∆x = 0. except here I use it to frame an array. and len is the amount of change to the right or left of the original point (it does not matter what the magnitude of ∆x is. which is centered. The “1” inside the circle required another \put.7 inches and height = . only its sign matters).∆y){len} If ∆x = 0. the position is centered because that is the default. I use the \fbox command. The oval.∆y){len} \vector(∆x. defined as usual in math mode: the array has three rows and one column. The next dashed box has the dash length set equal to .5 × . and the text is at the bottom right because of the optional specification.2 inches.1 inches. itself. which are lines with arrow heads. having the same dimensions as the framed box.74 6 GRAPHICS general form of the \framebox command in the picture environment is as follows: \framebox(width. followed by putting text that required some trial and error to locate. This is undoubtedly confusing. we cannot include the centering of text within the circle command. and len is the amount of change above or below the original point (it does not matter what the magnitude of ∆y is.height)[posn]{text} In the example shown in Figure 62.25 (inches). resulting in fewer dashes to compose the box. The same applies to the \oval specification. Unlike the box family. with the length of the dash set to . only its sign matters). so consider Figure 63. where . the line is horizontal. The next \put puts a dashed box. [br]. We know the center of the circle is at (−.2 inches. and the slope of the line is ∆x . with diameter = .

we want (xt . yt ) = (x0 + 1. we could have problems with approximating the results. Suppose. len = |yt − y0 |.3 to obtain the correct x-coordinate. We could setup a leastsquares estimation problem. Then. the calculation is simple: set ∆x = 0. Either way. then searching for a nearest slope approximation. The actual length of ∆y 2 the line segment is len 1 + ∆x . the new y-coordinate is y0 + len ∆x . we would set ∆x = 15 .1 Picture Environment 75 ∆y from (x0 . If xt = x0 . which I calculated to be from the “top left” box to the “center” box. and we want our destination point to be (xt . ∆y = −1 otherwise. If we we set len = 1. The ∆y closest we could come is 4 . Suppose our original point is (x0 .3. yt ).65. is not necessarily the best overall approximation. but trial and error in selecting the parameters tends to be just as efficient. ∆y must be integer-valued and within −6 to 6. y0 ) along the line with slope ∆x until the new x-coordinate is ∆y x0 + len. y0 + ∆y ∆y y0 + len ∆x Original point y0 x0 ∆y slope = ∆x New point x0 + len x0 + ∆x Figure 63: Line Parameters As if this unnatural definition of the line segment were not enough. but the restrictions do not permit this. y0 + 1. 5 Fixing len = xt − x0 . how should we set the slope parameters? 13 Ideally. y0 ).5). The first \vector command in Figure 62 starts at (−. −1).6. we have some work to do. for example. If xt = x0 . and 1 if yt > y0 . there is an important restriction: ∆x. .

3}. Thus. We begin the same way. −5).3) = . Finally. −1 − (−. specified as {. It required these computations to determine the complete picture command: \put(-. −. which is a vertical arrow from the same box to the circle below it. The left end point is at the y-coordinate of the center of the circle. y0 ).3.-1){. We want the coordinate of the end of the arrow (called the head ) to be flush to the left side of the “center” box. yc − yt + r) = (−.5 + 1 . so ∆x = 0 − (−. y1 ) = (xc + r. we need to determine the length. In this case.. as specified with \vector(1.-. We must add the radius.35){\vector(1.0){. y0 ) = (−1.2}.5). −1). y0 ) and h = height. The length is determined by where we want the arrowhead: at the top of the circle.3. the position of the arrowhead is (xh . drawn left to right.5). by computing the coordinates of the tail and head.5) and h = . and we set len = |∆y| = .55. yh ) = (xt . −.3. which is what is specified: top left ⇐ \vector(0. as specified. and its width is . Starting at y = −. The arrow is to be horizontal. yt ) = (−1 + 1 2 .35). which is its center. yc ) = (−. so we use two \vector commands to draw one arrow left to right. y) = (x0 + 1 h.7.-.4.3. The circles y coordinate is −. but drawn right to left.35). then an arrow at the same end points. −.35. so that is where we \put the first arrow. The uncertainty is the width of the box. given by \vector(0. we know only that the center of the box was put at (0. The initial position is calculated simply as the midpoint of the bottom edge of the box: (x. this involves more calculations because the arrow is not simply horizontal or vertical.76  66 GRAPHICS width - height ? box begins here → c In Figure 61 the “top left” box starts at (−1.65. That box begins at (0.0). so the coordinate where the arrow begins (called its tail ) is (−.3. −. . where the 2 box starts at (x0 . the vertical position changes by moving up half of the height. −1). and this needs some trial and error.5) = (−. −1) = (−.3.3.1) = (−. ∆x = 0 and ∆y < 0. −.3}}. but we ? .-1). Since the arrow is downward. The head is to be flush with the left edge of the \fbox.65.2) since .3) = (−. which is 1 (.1.2 is 2 the diameter specified by \circle{.65.65 + . so ∆x = 1 and ∆y = 0. and its x-coordinate is to the right by the length of the radius: (x1 .5) + . Now consider the next \vector. That 2 accounts for the initial position given by \put(-. so we obtain the coordinates of the arrow’s tail: (xt .3.65. −. (x0 .4). Further.4} 1m The next arrow is double-headed. so the right edge of that box is at x = −.

• There is no direct way to control the size or style of the arrow heads. we specify \line(1. the circles and ovals require separate \put commands. I shall cover these in the next section with a powerful package called PSTricks. With just a few iterations.0){. to a set of points. and its slope is (−1. so the midpoint of the bottom edge is at (1. There is a better way! 6. ∆y) = (1. Given this slope.6.2 PSTricks 77 do not know the width of the box.395) = . the best choice of len can be found as the average of the deviations: len = 1 (. p.3975} to obtain the line shown in Figure 61. • Some calculations and some trial and error are needed to align objects and lines.6. the end points were determined to be from (1. In this case.55. The closest slope we can have is with (∆x. −. which can take some trial and error to position. requiring re-calculations and more effort for the new positions. These can make using the picture environment time consuming and rather unpleasant. The reverse arrow begins at (0. but here are some things to note: • Only boxes can have inscribed text.2. and is provided free of charge.895) to (1. Now the true slope of the line we want is .4 do not allow this. due to not having the corner of the oval coordinates. but the restrictions .) In the preamble specify \usepackage{pst-all} for the entire system.6. and there is very limited control over line thicknesses. 117) gives the commands in the picture environment. 2 Thus. so len = ∆x = . −. However.5) and has a width of 1. which is what is specified. 1). • Moving a portion of the picture can be tedious.2 PSTricks PSTricks [15] was written by Timothy Van Zandt. but the latter was computed by knowing that the “bottom right” box starts at (1. There are packages to extend the picture environment. which is why we have \vector(-1. called Bezier approximations.395 . The former was found by trial and error. The last vector also required trial and error. but you can obtain it at CTAN [4].55}. −1). (You can use parts. and we can plot curves.2. 0).1){. Table 46 (in the Appendix.3975. in which case you specify the parts you use instead of pst-all — see [15] for loading individual portions.5). (It is not standard with MiKTeX. −. the end point was determined to be x = 0.5). −.4 + .) .

written by John D. • Arrow heads are adjustable. the default unit of measurement is 1 cm. identifying any of a great variety of arrowheads simply. • Lines and arrows have the same command. and the default fill color is white. . but MetaPost is more open-ended in its design. Hobby. All of the pst commands have options to override default settings for relevant parameters. for short) is designed to overcome difficulties with using the picture environment. which makes it potentially more versatile. The commands. some of which were listed above. typically available free of charge. • Circles and ovals. There are many packages [4]. Table 19 gives . thereby eliminating the need for calculation or trial and error. can be set with the \psset command: \psset{parameter = value[. Another widely distributed picture-drawing system is MetaPost [6. Some require converting to postscript and viewing the ps file. • Drawing curves is simple. but unlike the \put command in the picture environment. ]}. this is not the only way to put objects. The defaults. can have inscribed text. PSTricks (pst. themselves.fillcolor=gray} A fundamental command in pst is \rput. For example. themselves. and Bezier approximations of four points are available. 7]. This is especially true of commands that involve rotations. • Only one command is needed to put lines through a sequence of points. Many A of these are described in The L TEX Companion [5]. in addition to boxes. including plots of points that can come from a data file. that do many of the things done by PSTricks (and some additional things). can specify where to put them. Here are some of the features of PSTricks that I shall illustrate. also provided free of charge. but it does with MetaPost. It is more difficult to learn than PSTricks. especially on varying the types of file outputs (PSTricks is tied to postscript). • Shapes are highly variable. pdflatex (not covered here. but we can change them by specifying: \psset{unit=1in. see [4]) does not work with PSTricks. . • Objects can be named (as nodes) and lines and arrows can be drawn between them by naming the tail and head. and slopes need not be calculated.78 6 GRAPHICS One thing you need to know is that not all of the pst results can be seen with a dvi viewer. In particular..

<-> double arrow’ <. This is meant to be an introduction. (xn . y) with radius = r. along path given by coordinates. Thus. For example.0)(5.6. to produce a solid circle with radius .0)(10. y0 ) . y) with horizontal radius = rx and vertical radius = ry Draws line or arrow.-3)(6. The User’s Guide [15] is freely available and clearly written. -> forward arrow. shown in Table 20.-2) Draws rectangle with a corner at (x0 .2 PSTricks 79 some of the common commands to draw objects and lines. . we can specify relevant options as [parameter = value]. we write \pscircle[fillstyle=solid](0. and objects that could be made solid. Table 19: Some Basic Drawing Commands in PSTricks psframe(x0 . yn ) \pspolygon(0.0) (0. y)(rx . The origin is determined by where you are when issuing a pst command. .0)(5. Draws ellipse centered at (x.1)(10. Draws circle centered at (x. pscircle(x. . use the fillstyle parameter. we do want the \rput command in order to put text into various objects.2) psline{a}(x0 .0) \psline{<->} (0.backward arrow.-2)(1.0){.0)(10. yn ) \psline{-}(0. showing the ease and versatility of PSTricks. (xn .1} (having already set fillcolor=gray). A parameter used by these commands is the distance .0){2} psellipse(x.-2) pspolygon(x0 .1 cm. so many features are not presented here. determined by a: . yn ) to (x0 . For those examples. y0 )(x1 . . For each command. Draws closed polygon with given coordinates. . y0 ) .-3) In using these commands. All no environment is entered. The idea of a box is to have some shape enclose text. same as \psline{-}. PSTricks extends the rectangle in \framebox by having a variety of shapes. the unit of measurement was set to 1 mm. y1 ). like boxes and circles. then I shall show some additional shapes and commands. ry ) \psellipse(3. y){r} \pscircle(5. (there are more!). y1 ) \psframe(0.0) \psline{|-*} (0. y0 ). I can put that circle right here: commands use the linewidth parameter to control the thickness of the lines used in the drawing. . centered at the origin. I shall illustrate the commands in Table 19 first. .no arrow. y0 ) and opposite corner at (x1 . except figure is closed by drawing line from (xn .

and they can be scaled by specifying one of the following: \scalebox{size}{stuff} \scalebox{width.fillcolor=white}. linecolor. Boxes need not be enclosed (like \makebox). we obtain this oval by writing: . we obtain \psovalbox{this oval}. . . (As usual. .height}{stuff} Here are some examples: scales stuff keeping the same aspect ratio scales the width and height individually . For example. and fillcolor. . called framesep=len.4]{framebox} Adds shadow to psframebox \psshadowbox{shadow added} Draws double frame \psdblframebox{double frame} Draws circle around stuff \pscirclebox[linewidth=2pt]{circle} Draws oval around stuff \psovalbox[linestyle=dotted]{oval} These commands can be used in the text.) The pst figures are drawn after specifying \psset{unit=1mm.80 6 GRAPHICS between the border and the text inside. Table 20: Boxes in PSTricks psframebox{stuff} framebox framebox psshadowbox{stuff} shadow added psdblframebox{stuff} double frame pscirclebox{stuff} circle psovalbox{stuff} oval Draws rectangle but could have rounded corners \psframebox{framebox} \psframebox[framearc=. linestyle. where the default value of len is 3 pt. other parameters include linewidth.

These can be connected by \psline. In PSTricks. Here is how: \rotateleft{Left}\rotatedown{Down} \rotateright{Right} Left Right Down One application is given by the following: Who is the founder of TEX? Answer: Donald E. The name is set to A. and the text Node A is put there (with no frame). To avoid the tedious calculations in locating the coordinates of the tail and head.6. including variations of arrowhead shape. with a great variety of styles. the named objects are called nodes. Knuth So far I have described a variety of shapes. with the \rnode command.5}{\pscirclebox{ \begin{tabular}{c} Halving \\ the \\ circle \end{tabular} } } \scalebox{2}{\psframebox{ \textsl{Doubling} }} 81 Doubling Tall Wide \scalebox{1 3}{Tall} \scalebox{3 1}{Wide} There are times when we want to rotate stuff. Consider the following example: Node A Node B Node C Who is the founder of \TeX? \rotatedown{Answer: Donald E. Knuth} The source code is shown in Figure 64. The syntax for \rnode is: \rnode{name}{stuff } The next two commands put nodes named B and C. After entering the centering environment and setting the default units of measurement. each enclosed with a frame. The \ncline command has the same arrow options as \psline. the \rput command puts a node. but with the following syntax: \ncline{a}{name of node A}{name of node B } .2 PSTricks Halving the circle \scalebox{. the objects being joined can be referenced by name. by themselves and as enclosures for boxes.

linearc=. 0){1tail} \ncline{->}{1tail}{1} % tailess arc into (1) % Arcs (with labels) \ncline{->}{1}{2} \aput{:U}{1/2} % \aput puts label above arc \ncline{->}{2}{3} \aput{:U}{2/3} \ncline{->}{2}{4} \bput{:U}{2/4} % \bput puts label below arc \ncarc{->}{3}{4} \Aput{\small 3$\rightarrow$4} % \Aput keeps \ncarc{->}{4}{3} \Aput{\small 4$\rightarrow$3} % label horizontal \ncloop[angleB=180. The separation is exaggerated to 5 pt in the arrow from node C to node A. nodesepB=n.5. The [nodesepA=3pt] option gives 3 pt separation between the end of the line and node A.2]{<-}{4}{4} \Bput[5pt]{loop} \end{center} \vspace{1cm} Figure 65: Graph Source (Result in Figure 66) .-1){\rnode{B}{\psframebox{Node B}}} \rput( 2. 1){3}{3} \cnodeput[doubleline=true]( 2. which is what we want when the nodes are enclosed boxes. (Try adding one line at a time and observe each effect. In general.2]{->}{3}{3} \Bput[5pt]{loop} % \Bput keeps label horizontal and 5pt is % the space added between label and arc \ncloop[angleA=180. 0){\rnode{A}{Node A}} \rput(-2. which is not what we want. respectively.arm=. node separation can be specified for either end point. the line would touch Node A text. or for both end points. by specifying nodesepA=n. (nodesepA and nodesepB are keywords and have nothing to do with the names we assign to our nodes.-1){4}{4} \pnode(-3. Its source. or nodesep=n. 0){1}{1} \cnodeput(0.-1){\rnode{C}{\psovalbox{Node C}}} \ncline[nodesepA=3pt]{A}{B} \ncline[nodesepA=5pt]{<-}{A}{C} \ncline{<->}{B}{C} \end{center} \vspace{.) \begin{center} \psset{unit=1cm} % Nodes \cnodeput(-2.loopsize=.2. The default value is nodesep=0pt. is shown in figure 65.arm=.5in} Figure 64: PSTricks Source for Connecting Nodes Figure 66 shows a graph that could represent any number of things.82 6 GRAPHICS The first \ncline in Figure 64 draws a plain line from node A to node B.loopsize=.5. using PSTricks.2. like B and C. 0){2}{2} \cnodeput( 2. Otherwise.linearc=.) \begin{center} \psset{unit=1cm} \rput( 0.

-1) \psccurve(0. For example.4) \parabola{<->}(4.-1) . perhaps approximately. The examples that follow use the following pst settings: \psset{unit=. y1 ) is the (unique) point having dy/dx = 0. − 4a + c) The following shows two commands: pscurve and psccurve. given points.showpoints=false](1.2 PSTricks 83 3 1 loop 2/4 2/3 1/2 4→3 3→4 2 4 Figure 66: Graph Result (Source in Figure 65) Now I describe curves that go through.1)(1. whose command syntax is: \parabola{a}(x0 .1)(1.showpoints=true} (The showpoints=true setting is what causes the points to be included in the picture you see. y1 ).0)(1. 2 \pscurve{(->}(0. where (x0 .1)(2.0)(1.3) ⇒ Question: What is the pst command to draw the parabola given by y = ax2 + bx + c. and (x1 .) We begin with the parabola. y0 )(x1 .-1)(-1.3)(2. the latter being a closed curve that joins the last point with the first. \parabola* specifies filling the parabola. \psgrid[subgriddiv=1.gridlabels=7pt](-1.0) \parabola*[fillcolor=black. where a = 0? b b Answer: \parabola(0.-1)(-1.6. c)(− 2a .1)(-1.griddots=10. y0 ) is one point on the parabola.0)(4.1)(-1.5cm.

showpoints=false} \fileplot[plotstyle=dots]{mydata.0)(1.2)(19.0) \rput(45. Octave c .11)(29.Oy=0. Mathematica c . and S-PLUS c .10){\textsf{B}} \psline(40.6)(10. braces. the data file is read and its points plotted with the \fileplot command. The data file just needs pairs of coordinates.Dx=10.8)(39.8)(39. (The offset of 50 was used in establishing the origin in the plot.1)(3.dy=5. or nothing around each pair.5)(50. Maple c .0) \rput(25.3)(2.dat} \psaxes[Ox=50. (Setting showpoints=false .0)(40. The command syntax is: \psbezier[parameters]{a}(x0 . The following histogram was plotted by the source code in Figure 67. 8){\textsf{F}} \psline(11.) 15 C 10 F 5 0 50 60 D 70 80 90 100 B A score \psset{unit=2mm.11)(29.0)(1. y3 ) \psbezier(0.2)(19.ticks=y]{<->}(60.4) We can read data from a file. perhaps produced by mathematical software like gnuplot c .0) \rput(35. y0 )(x1 .0) \rput(14.5)(50. MATLAB c .0)(11. which can be separated by a comma or just blank and can have parenthesis.dx=10.0) \rput( 5.0)(20. The data file had y = number of students with test score = x + 50.6)(10. 7){\textsf{A}} Figure 67: Source Code for Drawing Histogram of Test Scores After setting the units of measurement to 2 mm. 4){\textsf{D}} \psline(20.13){\textsf{C}} \psline(30. y1 )(x2 .-2){\large score} \psline(1.84 6 GRAPHICS The Bezier curve joins two end points and comes as close as possible to two intermediate points.17) \rput[r](60.Dy=5. which I shall explain. y2 )(x3 .0)(30.

0).5cm]{->}(4.65) [65.0)(-1. This is suppressed for the x-axis in Figure 67 by specifying the option. y1 ) is the Southeast corner. Here are some examples: 2 1 0 0 1 2 3 2 1 −1 0 1 2 Note that ticks are uniformly spaced on each axes. dx=dy=0. itself. and (x2 .) The data file is plain text and has the following entries: % This is mydata. y0 ) is the origin.0)(2. As in \psline. the origin is assumed to be at (0.100] The plot. y0 ) is absent. respectively.dat} ⇒ Next. (x1 .2 PSTricks 85 suppresses plotting the points in the \psline commands.3) ⇒ \psaxes[unit=. (The default values. y2 ) where (x0 . specified by plotstyle=dots.1.2.1) ⇒ .90) [90. such as plotstyle=line. The other parameter settings are described in Table 21. ticks=y. y0 )(x1 . if (x0 .6.dat 5 2 9 4 15 2 18 1 22 6 27 4 30 2 31 1 35 4 39 1 40 2 45 2 50 1 % % % % % F D C B A = = = = = [0. Here is one of the alternatives: \fileplot[dotstyle=+.5cm]{->}(0. and there are 11 dot styles.1.) \psaxes[unit=. cause the spacing to be equal (approximately) by using Dx÷\psxunit and Dy÷\psyunit.0)(8. y1 ) is absent. There are other plot styles. y1 )(x2 .70) [70. y2 ) is the Northwest corner.plotstyle=dots]{mydata. If (x1 . it is assumed to be equal to the origin. is just the points.80) [80. axes are superimposed with the \psaxes command: \psaxes[params]{a}(x0 .

including bitmap (xbm). (If you can get a ps file. by specifying \psset{unit=1mm}): We shall stop here. is given by Keith Reckdahl [12]. There are free conversion systems on MS Windows.edu/octave/. at http://www. A basic plotting system for functions and data. Thus. when I superimpose the commands \fileplot. or there is a unix conversion utility. There are also commercial systems. It comes with MiKTeX and basic unix installations. is gnuplot.-2){\large score} puts “score” in large font. Octave extends the capabilities of gnuplot and is also available free of charge. This is available free of charge at FTP://ftp.) Many systems that let us draw figures. xfig is an excellent system to draw figures. (He also goes deeper into customizing placements of pictures in figures. this is scaled (simply. provided free of charge. The unix systems xv and Image Magick can do this for a large variety of graphic file formats. which converts Windows Metafiles (wmf) to eps.3 Importing pictures A The way to import a picture into L TEX is to convert it to encapsulated postscript (eps). but this does not exhaust the PSTricks commands. rput[r](60.dartmouth. Carlisle. and the export options include the eps file format. Once the file is in eps format. Leaving off the “score. Mathematica c . Another way to obtain an eps file is with conversion. There are two packages that provide . and S-PLUS c .che.wisc. See [15] for lots more. −2). An exceptionally clear description of this.edu/pub/gnuplot/. gif and jpeg files. we can import it using the Graphics Bundle [3]. To fit the picture and the code next to it. \psaxes and \rput. written by David P. MATLAB c . for both unix and DOS that produces eps files. which converts jpeg files to eps. \ps2epsi. which can produce eps files of plots.” Figure 68 shows the sequence of how each \psline and \rput adds to the picture. and those that plot mathematical functions or data. including many examples. have an option to export an eps file.86 6 GRAPHICS Table 21: Parameters for Horizontal Vertical Default Ox=n Oy=n 0 Dx=n Dy=n 1 dx=n dy=n 0 \psaxes Meaning Label at origin Label increment Label spacing The next command. and emftoeps. we obtain the data plot. flush right (indicated by [r]) at the coordinates (60. 6. notably jpeg2ps. including historical context. like Maple c .) On unix. The remaining commands draw the histogram boxes and put the letter grade above each box in sans serif font. you could use \psfig.

simply specify \includegraphcs[options]{filename}.0)(1.13){\textsf{C}} \psline(11.ticks=y]{<->}(60.6)(10.0)(40.dy=5.4){\textsf{D}} 50 F \psline(1.6.10){\textsf{B}} 50 F D 60 70 C 80 90 100 \psline(20.Oy=0.2)(19. the other is graphicx. One is called graphics.0) \rput(45.8)(39.0)(11.6)(10. For example. \fileplot[plotstyle=dots]{mydata. as specified in the preamble by \usepackage{graphicx}.17) 60 70 80 90 100 Figure 68: Sequence of PSTricks Commands to Draw Histogram To include an eps file.dat} 15 10 5 0 15 10 5 0 15 10 5 0 15 10 5 0 15 10 5 0 15 10 5 0 50 F D 60 70 80 90 100 50 F D 60 70 C B A \psline(40.Dx=10.8){\textsf{F}} 50 60 70 80 90 100 \psaxes[Ox=50.0) \rput(25.2)(19.0) \rput(5.0) \rput(35.Dy=5.8)(39.3 Importing pictures 87 essentially the same capabilities but with different syntax. Figure 69 shows a figure that was imported with the following statement: . Here I use graphicx.5)(50.0)(20.11)(29. dx=10.7){\textsf{A}} 80 90 100 50 F D 60 70 C B 80 90 100 \psline(30.5)(50.0) \rput(14.11)(29.0)(30.

file.eps} \end{center} 1 0.eps}\end{center} 1 0.8 0.8 −1 −8 −6 −4 −2 0 2 4 6 8 Figure 69: Applying \includegraphics to Import an eps File In this case I specified the half the size it was produced print sin -deps after plotting Figure 70 shows the same eps follows: option.6 0. we might want to specify width=\textwidth. by Michael C.6 −0. and let it fill the entire width of the page.2 −0.6 −0.4 0.8 −1 −8 −6 −4 −2 0 2 4 6 8 Figure 70: Specifying Dimensions in \includegraphics For a very large picture.8 0.4 0.2 0 −0. Grant and . The height specification (!) says to maintain the aspect ratio.4 −0.height=1in]{sin. If you find yourself importing eps files but would like to make some A changes in L TEX.2 −0.5]{sin.height=!. but with the width and height set as \begin{center} \includegraphics[width=2in. read about the PSfrag package. which prints the figure (in this case by MATLAB.6 0.88 6 GRAPHICS \begin{center}\includegraphics[scale=. scale=.4 −0. by specifying the sin function over the indicated grid).5.2 0 −0.

. Here are examples: Double your fun \scalebox{2}{Double your fun} \resizebox{1in}{!}{\fbox{Open wide}} \reflectbox{Reflect on this} \rotatebox[origin=c]{90}{Landscape} \rotatebox[origin=rt]{45} {\psframebox{ \begin{tabular}{c} Was\\Pythagoras\\a square? \end{tabular} } } Open wide siht no tcefleR Landscape a Py W a sq ag s or ua a re s ? th These operations are available because the programs that perform them are used in the \includegraphics command. It has two basic operations: (1) edit A some string or position in the figure (i. or almost any stuff. Importing graphics is only one of the functions of graphicx. and sizing of an arbitrary box. Although it is feasible to perform the operation after importing a graphic. the eps file). it is more efficient to specify that option in the \includegraphics. Here are some examples: . whose documentation is at CTAN [4]. It can also perform scaling. rotation. pictures. The documentation gives examples.3 Importing pictures 89 David Carlisle.e. with eps files produced by MATLAB and xfig.6. which comes with a basic installation (including MiKTeX). and (2) translate L TEX commands that you put in the figure in the first place. The box could contain text.

Submit a printed copy of the L TEX source (tex file) and printed   ©   vm 4' vm 3 3. Draw the following graph with the picture environment. .width=!. 2.origin=c]{protractor.25\textwidth.eps} \includegraphics[height=.eps} \includegraphics[width=. Use PSTricks to draw Figure 3 (p. height=!]{protractor. angle=90. Use the picture environment to draw the smiley face on page 73. where \thicklines is specified and \unitlength = 1mm vm 1   E vm 2     c A Exercises.90 6 GRAPHICS \includegraphics {protractor.5in.eps} copy of the associated postscript result (ps file). Be sure your name is on each. 4). 1.

6. (They were drawn here with PSTricks. but this section did not describe all that is needed.3 Importing pictures 4. Make a figure in some system that lets you save it as an eps file (or use some conversion program).) (a) Graphic view of Pythagorean Theorem: square of hypotenuse square of leg 2 hypotenuse leg 1 2 square of leg 1 . Use PSTricks or the picture environment to draw the following. Use PSTricks or the picture environment to draw the following. include it in your document. Use whatever means you prefer (or that your instructor requires) to include each of the following figures in your document. 91 rhombus 5. Then. 7. so you must obtain the PSTricks User’s Guide [15]. β α 6.

(35. 30) 50) (45. 20) (25. 60) 3 5 (c) The sin function: y = sin π x 2 x (d) Bernoulli family tree: Nikolaus (1623–1708) Jacob I (1654–1705) Nikolaus (1662–1716) Jahann I (1667–1748) Nikolaus I (1687–1759) Nikolaus II (1695–1726) Daniel (1700–1782) Johann II (1710–1790) Johann III (1746–1807) Jocob II (1759–1789) .92 (b) Network with arc data: 2 4 6 GRAPHICS (15. 10)130) 50) (15. 40) (45. (35.

the following does this while specifying 12pt font as another option: . which can be defined anyplace before the \maketitle.) Specifying \date is optional (\maketitle puts in the current date if the date is not defined). Typically these are put into the preamble. Multiple authors are separated by \and. titlepage must be specified as an option in the \documentclass command. You might also need to initialize the page counter with \setcounter{page}{0}.1 Making Special Parts Cover Page The easiest way to make a cover page is with the \maketitle command. The cover page is by itself and is not numbered. If you see a page number on your cover page. For example. it depends upon your management style. or right after \begin{document}.93 7 7. suppress this by adding \thispagestyle{empty}. \title{The \LaTeX\ Companion} \author{Michel Goosens \and Frank Mittelbach \and Alexander Samarin} \date{1994} \maketitle Figure 71: Title Page Source (Result in Figure 72) A The L TEX Companion Michel Goosens Frank Mittelbach Alexander Samarin 1994 Figure 72: Title Page Result (Source in Figure 71) Since articles often have this information on the first page of the article (rather than a separate page). right after \maketitle. This is usually done just following \begin{document}. such as in the example shown in Figures 71 and 72. followed immediately by \maketitle. (The jagged edges in Figure 72 mean that there is more space between the title and the top of the paper. The necessary parameters are \author and \title.

affiliations. or if the names were very long. A The L TEX Companion Michel Goosens Geneva. \maketitle puts the third author on a separate line. . Germany \and Alexander Samarin \\ Geneva Switzerland} As illustrated in Figure 73. Figure 73 shows how the authors appear when the \author definition in Figure 71 is changed to the following: \author{Michel Goosens \\ Geneva. The \thanks command does this by creating a footnote. using \\ to create new lines. using different footnote marks for each one. Germany Alexander Samarin Geneva. For example. All three authors would be put on separate lines if the address information were extended further.94 7 MAKING SPECIAL PARTS \documentclass[12pt. Switzerland \and Frank Mittelbach \\ Mainz. Figures 74 and 75 illustrate this along with some variation in the date. This is because the added width of author information makes it too long to fit on one line. Switzerland Frank Mittelbach Mainz. and other information about each author can be added. Switzerland 1994 Figure 73: Adding Addresses to Authors There are times when we want to acknowledge support for one or more of the authors.titlepage]{article} Addresses.

1 Cover Page \title{Pieces of $\pi$\thanks{Renamed. 510 bc) 1 Renamed. 2 Supported by the army.R. \title{Doing \author{I. \and U. This uses the \footnotemark command. Ionia } \date{210 {\sc bc} (revision of earlier version. Prior to that. } \LaTeX} Rich \thanks{Smart University} Grand \thanks{Street-Smart} All \addtocounter{footnote}{-2} \footnotemark Figure 76: Authors with same footnote (Result in Figure 77) . 510 {\sc bc})} 95 Figure 74: Footnotes in the Cover Page Source (Result in Figure 75) Pieces of π 1 Archimedes2 Syracuse. which puts the footnote mark without any new text.M. Sicily Pythagoras Samos.C. the value of the footnote counter is set back to the first footnote mark. Sicily \and Pythagoras \\ Samos. \and I.7.}} \index{\texttt{$\backslash$thanks}} \index{footnote} \author{Archimedes\thanks{Supported by the army.}\\ Syracuse. Figure 75: Footnotes in the Cover Page Result (Source in Figure 74) Figure 76 shows how to use only one footnote for authors having the same affiliation. Ionia 210 bc (revision of earlier version.

specify titlepage as an option in \documentclass (even if you do not intend to use \maketitle). the abstract is placed far from the top of the paper.C. which is not shown in Figure 79.M. itself. Grand2 I. To have it. in boldface and centered. Rich1 U. is one paragraph and is printed without indentation. denoted $\pi$. This environment is defined to produce an abstract on a separate page (placed wherever you put the environment specification).96 7 MAKING SPECIAL PARTS A Doing L TEX I. The abstract.2 Abstract The abstract environment is in all document styles. All1 1 Smart University 2 Street-Smart Figure 77: Authors with same footnote (Source in Figure 76) 7.) \begin{abstract} This shows that the ratio of the circumference to the diameter of any circle is the same constant value. except article. We further prove that this constant is bounded by $\frac{223}{71} < \pi < \frac{22}{7}$. \end{abstract} Figure 78: Making an Abstract Source (Result in Figure 79) . (Like the cover page. with the header: Abstract. Figures 78 and 79 illustrate this.R.

it is placed wherever you put the command. We further prove that this constant is bounded by 223 < π < 22 . followed by the list of tables. which is on page v. To include other front matter. I use the \addcontentsline to add it to the table of contents.3 Other Front Matter The \tableofcontents command makes a table of contents. Then. Each of these are put on a new page. For example. Then. like sections and A subsections. which should be right after the cover page. That is why you see the Table of Contents on page i (first numbered page. Just above each declaration. respectively. 71 7 Figure 79: Making an Abstract Result (Source in Figure 78) 7. I declare \listoffigures. indicated by the toc specification. just after the cover). \newpage \pagenumbering{roman} \pagestyle{myheadings} \tableofcontents \newpage \addcontentsline{toc}{section}{List of Figures} \listoffigures \newpage \addcontentsline{toc}{section}{List of Tables} \listoftables \newpage Figure 80: Some Front Matter Specifications for This Document The \pagenumber specification causes the page numbers for the front matter to be put into Roman numerals. . The table of contents generally includes numbered parts.7. The section parameter tells the latex program to format it like a section — flush left. the table of contents in this document was obtained with the specifications given in Figure 80. denoted π. The page numbering is reset when we finish the front matter by specifying \newpage \pagenumbering{arabic} \pagestyle{headings} This switches to the Arabic numerals and initializes the page counter.3 Other Front Matter 97 Abstract This shows that the ratio of the circumference to the diameter of any circle is the same constant value. L TEX provides the \addcontentsline command. you can include lists of figures and tables with the \listoffigures and \listoftables commands.

For now. . we put the bibliography (see §3 and §8. \end{abstract} Alternatively. we might want something to look like a section (and automatically added to the table of contents). enter at the command line: makeindex myfile Then. but with no number (and the section counter remains unchanged). \end{appendix}. where the * suppresses the numbering. Put \printindex just before \end{document}. The only change we require is another header name.98 7 MAKING SPECIAL PARTS The same format as the abstract can be used for other front matter that we want to format the same way. consider the following example that illustrates how to have an Acknowledgments page: \renewcommand\abstractname{Acknowledgments} \begin{abstract} I thank my family and friends for all of their support. compile again. but we do not want it to have a section number. I also thank the contributors to the Comprehensive \TeX\ Archive Network (CTAN). This could be done with the appendix environment: \begin{appendix} . In the text. We might first want to have appendices that follow the main text. For example. 2. To have entries in the index. . the index in this book contains several occurrences of ‘flushleft’. This is analogous to the use of bibtex (p.6). we have three things to put into our source file: 1. you use \index{entry}. 28). Put \makeindex at the end of the preamble. This is done by re-defining the \abstractname parameter used by the abstract environment. 7. 3. This could also be desirable in a long report. with all references resolved. After a successful compilation. This is achieved by the \section* command.4 Back Matter After the main part of the document is finished. \section*{Preface} puts “Preface” in the same style as any section. and is illustrated in Figure 81. §?? has more to say about using this command to customize many things. For example. To make an index. Put \usepackage{makeidx} in the preamble. The last portion in the back of any book is its index. . The \renewcommand enables us to do this.

5 Footnotes The \thanks command is one way to have a footnote on the cover page. 1 2 first second Figure 83: Setting a Footnote Result (Source in Figure 82) \thefootnote gives the counter for footnotes. 7. More generally. create/edit compile latex myfile. here is my second\footnote{second}.5 Footnotes 99 at each occurrence. \index{package!makeidx} puts the entry ‘makeidx’ under the entry ’package’. The distance between the line that underlies part of the last line of text to the footnote is \footnotesep. Figure 82: Setting a Footnote Source (Result in Figure 83) Here is my first footnote1 . nomencl makes a list of nomenclature.ps myfile.pdf Figure 81: Adding makeindex to the Command Sequence There are packages to make other back matter: acronym makes a list of acronyms.tex view/print convert dvips myfile. as shown in figures 84 and 85. which can be controlled by \setlength. I specify \index{flushleft}.dvi dvipdf makeindex print/post myfile. For example. Figures 82 and 83 illustrate. . and you can change from numbers to letters by \renewcommand{\thefootnote}{\alph{footnote}}. You can also change to common footnote symbols by specifying \renewcommand{\thefootnote}{\fnsymbol{footnote}}.7. use !. and gloss makes a glossary. and was shown on page 95. here is my second2 . the \footnote command can be used anywhere. To put the entry as a subordinate. Here is my first footnote\footnote{first}.

and whatever is whatever you want the command to do. A Exercises. which enable you to define your own commands and change parameter values of existing commands.1 Your Own Abbreviations and Commands The command that gives us the ability to make our own has the following form: \newcommand{\name}[n]{whatever }. 2. Main Results. Submit a printed copy of both the L TEX source (tex file) and 1. but also the abstract. 8. and Conclusions. Make the main body have at least three sections: Introduction. Write an article with a title page and abstract. 3.100 8 TAKING CONTROL \renewcommand{\thefootnote}{\fnsymbol{footnote}} Here is my first footnote\footnote{first}. Figure 84: Setting a Footnote Source (Result in Figure 85) Here is my first footnote∗ . It is still in the context of an introduction. choosing only a few of the things you can change. Be sure your name is on each. Here are two examples simply to abbreviate commands with long names: \newcommand{\ul}{\underline} \newcommand{\mc}{\multicolumn} . A key to these changes are the \newcommand and \renewcommand commands. acknowledgments and references. Extend exercise 1 to have acknowledgments and references (using BibTE X). 8 Taking Control This section introduces you to fundamentals of customizing your document. Combine exercises 1 and 2 and add a table of contents showing not only all sections and subsections. where n is the number of arguments. here is my second† . here is my second\footnote{second}. ∗ † first second Figure 85: Setting a Footnote Result (Source in Figure 84) the associated postscript result (ps file).

tex (note the . so we must be careful not to override them with ours. then add to the preamble: \newcommand{\usenul}{\mynul} and specify \usenul in the document. The preamble can become very long as we add our commands. we use the \input command to have the latex compiler read . spanning 3 columns and centered. we can specify \Gs{subscript}. we can write \nul(A) to obtain N (A) (and we can be in text or math mode when we write this). it is replaced by $G_{i+j}$ to put it into math mode first. and obtain the correct result.7)} \end{picture} }} (\mbox is used to ensure text mode). no matter which mode we are in. For example. \Gs{i+j} is replaced by G_{i+j} to produce Gi+j .8. We can choose one and define \newcommand{\nul}{\ensuremath{\mathcal{N}}}. like mynul. Consider the following example: \newcommand{\Box}{\mbox{\begin{picture}(0. in the preamble I specified: \newcommand{\chkbox}{$\Box^\surd\. but we want them to work in any mode. you will get a fatal error message since there is already a \fbox command. √ \chkbox ⇒ Some commands are specifically for math mode. Thus. simply change the one line to: \newcommand{\usenul}{\nul} (where the publisher’s command name is \nul). A way to do this is to choose a different name. Suppose we have a key term. Now \Box ⇒ and. For example. otherwise. like the null space of a matrix. For example. The second lets us write \mc{3}{c}{stuff} to enter a multicolumn. Then.1 Your Own Abbreviations and Commands 101 The first lets us write \ul{something} to underline something. A related use is when the command requires some lines of code. Some publishers have their own notation. consider \newcommand{\Gs}[1]{G_{#1}}. and there are still more symbols people use. in either a tabular or an array environment. we can use it in other new commands. so it is useful to put them in a separate file. If we are already in math mode.tex suffix)..0) \put( 2. Another reason to have our own commands is for consistency.. if \Gs{i+j} is specified in text mode. if you specify \newcommand{\fbox}. Then. some write nul A. The latex compiler will not let you use a name that is already being used. particularly of notation.$} Then. Some authors write N (A). having defined the \Box command.. If you need to use the publisher’s. say mydefs. This is achieved by the command: \ensuremath{math stuff }.0){\framebox(7.

If you have no chapters. In particular. We saw an example of this in changing the counters for enumeration lists (p. . Making the document class an article will also solve the problem. since the section is the first level division. Making the first level division chapters will overcome the numbering problem. in this document the “Table of Contents” was obtained by specifying the following in the preamble: \renewcommand{\contentsname}{Table of Contents}. The general form is \renewcommand{\thecounter }{something} Another example is to change section numbering in a report document style.) Different source files could simply input this same file. 43).1. but the format of chapters is different. so the numbering will be chapter. Table 22 shows the common names we might want to change. such as entering the document into a database using .2 Your Own Names. 8. Table 22: Intrinsic Name Parameters What it is Abstract Appendix Chapter Contents Index List of Figures List of Tables Part References How it is called (keyword) \abstractname \appendixname \chaptername \contentsname \indexname \listfigurename \listtablename \partname \refname for article style \bibname for book and report styles You might want to change the numbering of some intrinsic counter.102 8 TAKING CONTROL it wherever it is placed. similar to a book.subsection]. the preamble of this document contains the command: \input{mydefs} (The suffix . The first level of division is assumed to be a chapter. Titles and Numbers There are times when we prefer some name other than the default.tex is assumed. it will number the first section as 0. This might not be appropriate due to other considerations.section[. For example. so duplication of work is avoided. .

Thus. not as an article. \makeatother. is to make the @ character a letter. For example. 8.4 Your Own Margins and Spacing The default margins and spacing are set with purposeful values. .1mm} \end{description}\end{flushleft} } % end proof Then. equal to about 2 spaces). \end{proof} produces: Proof: First. restores @ to its special meaning (\@ is for certain spacing. \begin{proof} First.3 Your Own Environments The \newenvironment command enables us to define our own environments. suppose \dots \linebreak Thus. the theorem follows. and the \renewenvironment command enables us to revise an existing environment. and you will usually not need to change them. \makeatletter. the theorem follows. . the following creates a proof environment: \newenvironment{proof} {\begin{flushleft} \begin{description} \item \textit{\textbf{Proof:}}~ } % begin proof {\hfill\rule{2. they can be changed by setting certain parameters in the preamble. begin is what is executed upon entering the environment. suppose . and end is what is executed upon leaving the environment. n = number of arguments (omit [0] for n = 0). where you want it to be counted as a report.1mm}{2. 8. The succeeding command. The way to do this is as follows: \makeatletter \renewcommand\thesection{\@arabic\c@section} \makeatother The preceding command. however.8.3 Your Own Environments 103 BibTE X. When you do. The margins . They have the same syntax: newenvironment{name}[n]{begin}{end } renewenvironment{name}[n]{begin}{end } where name is the name of the environment.

we raise the body 1 inch by specifying \topmargin=-1in in the preamble. .38744pt \textwidth 360.396 in 8.5 in end body stuff 113. We might want some global settings to make repeated use of these unnecessary.025 pt 1.0pt \voffset 0. the current settings (shown 2 in Table 23) break down the horizontal parts as follows: 72.0pt 25. Margin settings can be negative.96999pt \paperwidth 614. The body expands to the right unless we also change \oddsidemargin.27 in. The geometry package provides easy specifications for page layout.14. respectively.0pt Meaning space between bottom of body and top of footer space between bottom of header and top of body height of header horizontal offset to add to indentation of body extra space added at left (applies only to odd numbered pages if the style is two-sided. 1 pt = 72. This might be accompanied by increasing the text length. Table 24 lists some you can set with the \setlength command. for example.) For example.0pt 12.0pt † Printed using \theparameter. Table 23: Margin Parameters Parameter \footskip \headsep \headheight \hoffset \oddsidemargin Current Setting† 30. if we are using 8 1 × 11 paper.295pt \textheight 585. in which case there is also an \evensidemargin parameter) height of the paper width of the paper height of the body width of the body space added before the top of the header vertical offset to add to indentation of body \paperheight 794.295 pt 390 pt 5. For example. \textwidth=6in increases the text width to 6 inches. (See Table 25 for conversion factors.0pt \topmargin 21.0pt 29.0pt 0.27 pt 39 pt 1 in 54 in begin body stuff 6.104 8 TAKING CONTROL of the document are controlled by the parameters shown in Figure 86 and described in Table 23. in particular. showing their default values (used in this document). The \hspace* and \vspace* commands provide a great deal of control over horizontal and vertical spacing.564 in We can increase the text width by setting \textwidth=length in the preamble.

4 Your Own Margins and Spacing \pagewidth 105 1 in + \voffset \topmargin header \headsep \headheight \textheight \textwidth \pageheight body \oddsidemargin footer 1 in + \hoffset \footskip Figure 86: Document Margins .8.

Figure 87 shows two enumerate lists. and I have saved it by: \setlength{\mylength}{\itemsep}. This list specified \setlength{\itemsep}{12pt} right after \begin{itemize} 2.106 8 TAKING CONTROL Table 24: Spacing Parameters Parameter \itemsep \parindent \parsep \parskip Meaning space added to \parsep between items in a list. indentation at beginning of paragraph. This item is farther from the first one. 12) are spaced by default values. This list specified \setlength{\itemsep}{0pt} right after \begin{itemize} 2. For \itemsep this can occur just after the \begin statement. What you see next is with \setlength{\itemsep}{0pt}. This must be set before the \begin command.5pt plus 2.2 (p. • What you see next is with \setlength{\itemsep}{10pt}. Figure 87: Varying \itemsep to control item spacing in a list The left margin parameter is \leftmargini for a level-1 list (note the i at the end). . • See the above spacing between items. In the case of list parameters. varying by the value of \itemsep.0pt minus 1. This item is fairly close to the first one. • Next is back to normal by \setlength{\itemsep}{\mylength}. space between paragraphs. (Defaults are restored after leaving.0pt. Here is what happens when we change \itemsep: • The default value of \itemsep is 4. they must be set within the list environment. space between paragraphs in the same item of a list. • We are back to normal with \itemsep = 4. 1.0pt. Figure 88 gives an example.) For example. 1.0pt minus 1.5pt plus 2. the lists in §2.

so we want to increase its vertical spacing.3 line spacing: \renewcommand{\arraystretch}{1.4 Your Own Margins and Spacing {\setlength{\leftmargini}{2in} \begin{itemize} \setlength{\itemsep}{0in} \item Item 1 of list with 2in left margin \item Item 2 of list with 2in left margin \end{itemize} } 107 Figure 88: Changing the Left Margin of a List (Result in Figure 89) • Item 1 of list with 2in left margin • Item 2 of list with 2in left margin Figure 89: Changing the Left Margin of a List (Source in Figure 88) Figures 90 and 91 show the presentation of an array with a p-column to put horizontal space between the other two columns.3in}l} \.x_B = b_N + \frac{1}{2}\theta\delta b_N && \pi_N B = c_B + \frac{1}{2}\theta\delta c_B \\ B^*x_B > b_B + \frac{1}{2}\theta\delta b_B && \pi_N N < c_N + \frac{1}{2}\theta\delta c_N \end{array} \] Figure 90: Array with Fixed Width Column Source (Result in Figure 91) 1 B xB = bN + 2 θδbN B ∗ xB > bB + 1 θδbB 2 1 πN B = cB + 2 θδcB 1 πN N < cN + 2 θδcN Figure 91: Array with Fixed Width Column Result (Source in Figure 90) Here is 1.6 line spacing: \renewcommand{\arraystretch}{1.8.B\. \[ \begin{array}{lp{.3} B xB = bN + 1 θδbN 2 B ∗ xB > bB + 1 θδbB 2 1 πN B = cB + 2 θδcB 1 πN N < cN + 2 θδcN Here is 1. Note how congested it is.6} B xB = bN + 1 θδbN 2 B ∗ xB > bB + 1 θδbB 2 1 πN B = cB + 2 θδcB 1 πN N < cN + 2 θδcN .

followed by your second argument. 8. You will have control of the whole document. Suppose that you want no page numbering if there is only one page. }#2}}{} } This will test the condition if your first argument equals 1. there is more than one page. How much is 2+2? \printsol{\sol}{4} Figure 92: \ifthenelse Source (Results in Figure 93) \def\sol{0} . suppose your preamble contains \newcommand{ \printsol}[2]{ \ifthenelse{\equal{#1}{1}}% {\textsf{\textbf{Solution. If there is more than one page. First. 4 Figure 93: \ifthenelse Results (Source in Figure 92) Another example is to decide on page numbering. but it is easier to use the setspace package.5 Your Own Output Control You can write conditionals to do one thing or another with the ifthen package. and the first statement applies: \pagestyle{empty}. If not. . it will do nothing. just before \end{document}.108 8 TAKING CONTROL Back to default: \renewcommand{\arraystretch}{1} 1 B xB = bN + 2 θδbN ∗ x > b + 1 θδb B B B B 2 1 πN B = cB + 2 θδcB 1 πN N < cN + 2 θδcN Line spacing is controlled by \baselinestretch . put \label{lastpage}. or just some lines whose space you want to control. How much is 2+2? \def\sol{1} . it will print Solution. If so. and the second statement applies: \pagestyle{headings}. How much is 2+2? Solution. . You can do this as follows. . . . suppose you want to use headings in the header with page numbers. put the following: \ifthenelse{\equal{\pageref{lastpage}}{1}}{\pagestyle{empty}}% {\pagestyle{headings}} This will test if the value of \pageref{lastpage} is equal to 1. if not. there is only one page. If so. The syntax is \ifthenelse{condition}{action if true}{action if false} For example. Just after \begin{document}. This is illustrated in figures 92 and 93.

With you in control. The bib style file.) Each \bibitem is an entry. so each entry appears however you put it. MA. This is one reason it is usually better to use BibTE X. 1989. Knuth. specify the following: \begin{thebibliography}{n} \bibitem[what appears]{label (that you cite)} entry . Alternatively.8. \end{thebibliography} where n is the width of the widest label you want to allow. \bibitem{tex} Donald E. Instead of the BibTE X commands. Frank Mittelbach and Alexander Samarin. \end{thebibliography} These will appear in the document’s list of references even if they are not cited. which I formatted to agree with the plain style of BibTE X: \begin{thebibliography}{99} \bibitem{companion} Michel Goosens. \textit{The \LaTeX\ Companion}. \bibliography{mybiblio} and \bibliographystyle{plain}. . such as . 1994. . 15th edition. Some publishers give you no choice.. we obtain [2] in the text. .e. and there will be no sorting — the list of references will appear in the order you put them. Knuth. Thus. . Here is a complete example with two references. it generally helps the reader to know something about the citation. 1989]{tex} Donald E. Reading. and use thebibliography environment instead. and you can enter whatever you like. in which case \cite{tex} ⇒ [Knuth. even if there are inconsistencies in style. They can be cited in the same way described in §3: by \cite{companion} and \cite{tex}.6 Your Own Bibliography 109 8. MA. even though you lose control over what appears (i. Addison-Wesley Publishing Company. they will be numbers). The option is an alternative to having the references numbered. 1989] is preferred to [2] because it immediately gives the reader information about the document without having to flip to the bibliography section. \textit{The \TeX\ Book}. . When citing Knuth’s book. for example. but if you are writing a report and have control over the formatting. there is no format monitoring.6 Your Own Bibliography You can choose not to use BibTE X. You will have complete control over the formatting. with label the unique identifier used by the \cite command. (It works if you specify 99. [Knuth. respectively. Reading. 1989]. Addison-Wesley Publishing Company. as described for BibTE X in §3 (p. 28). we can exercise the option: \bibitem[Knuth.

Here are some packages that give you special symbols: chemsym.text. such as at the end of each chapter of a book.bst. 1989] instead of [8]. available from CTAN [4]. You will find other packages useful. the algorithm package enables an environment to write source code with standard language elements. applies in either case. These packages provide even more versatility in how the citations appear (see [4] or [5. Also. such as [Knuth. apalike and plainnat (from natbib). wasysym. c-pascal and listings). Closing Remarks A Now you know how to write a mathematical document in L TEX 2ε and you know there is much more you can learn to gain refinements. which you obtain from CTAN [4]. Chapter 13]). Alternative bst files are achicago (from the frankenstein package). Most installations come with more than the basic plain. there are many packages. 36). and xypic. and its three variations (given on p.. Figure 94 shows the preamble used for this document. SIunits. including those commonly found in automata theory.110 8 TAKING CONTROL plain. As you begin to use packages. qsymbols.tex (can be done in Google by clicking on Groups). If you want to have several bibliographic units in one document. and there are others with similar properties or for particular programming languages (viz. Besides what you can do yourself to elevate the quality of the results. it is necessary to become aware of updates. The graphtex package specializes in all sorts of graphs. . or you can join their newsgroup at comp. depending upon your technical area. You can visit CTAN from time to time. use the bibunits package. which give the author and year.

. which gives conversion among three common units of measurement. \mathscr) like multicolumn enables more control over newtheorem \url{.54 1 Table 26 is a guide to how most of the remaining tables are organized.45 in . Afterwards.3937 cm . Table 25: Conversions of Common Units of Measurement pt in cm pt 1 72..APPENDIX 111 \usepackage{amsmath} \usepackage{amssymb} \usepackage{amscd} \usepackage{bm} \usepackage{graphicx.pst-all} \usepackage{hyphenat} \usepackage{fancyvrb.} \renewcommand\contentsname{Table of Contents} % Change ‘Contents’ \renewcommand\url{\begingroup\urlstyle{sf}\Url} % put url in sf font \input{mydefs} % My commands and environments \makeindex % make myfile..27 28.moreverb} \usepackage{float} \usepackage[T1]{fontenc} \usepackage{ifthen} \usepackage{makeidx} \usepackage{mathrsfs} \usepackage{multirow} \usepackage{theorem} \usepackage{url} % % % % % % % % % % % % % % % formerly amstex ams symbols (\mathbb fonts) draws commutative diagrams bold math fonts (\mathbm) graphics enables control over hyphenation verbatim enable float [H] option . and Table 46 gives the commands for the picture environment.}} ifthenelse {condition}{true}{false} index more math symbols (viz.03515 2. It is designed like a reference manual for easy lookup.. ..to write \textbf{\textsc{. beginning with Table 25.01384 1 .idx (input to makeindex at command line) Figure 94: Most of the Preamble for this Book Appendix A This contains complete tables of font information and basic L TEX commands.. Table 45 gives special symbols that can be used in either text or math mode.

Triangles and Lines Variable Size Symbols Table 27: Commands/Environments for Text Font Appearance textbf tiny Large verbatim textit scriptsize LARGE textrm footnotesize huge textsc small Huge textsf normalsize underline texttt large verb Table 28: Commands/Environments for Controlling Text Position bigskip flushright medskip pagebreak smallskip vspace center hfill newpage quotation tabbing vspace* centerline hspace noindent quote tabular clearpage hspace* nolinebreak raisebox verse flushleft linebreak nopagebreak samepage vfill .112 APPENDIX Table 26: Reference Tables Table Text mode 27 28 29 30 31 32 Math mode 33 34 35 36 37 38 39 40 41 42 43 44 Contents Commands/Environments for Font Appearance Commands/Environments for Controlling Position Text Accents and Symbols Commands for Counters Commands/Environments to Organize Document Commands to Control Document Style Commands to Control Fonts in Math Mode Accents in Math Mode Spacing Commands in Math Mode Greek and Special Letters Frequently Used Mathematical Symbols Binary Operations Operators and Quantifiers Special Functions Relation Symbols Arrows Dots Circles.

APPENDIX 113 Table 29: Text Accents and Special Symbols á è î ö æ Æ Ł \’{a} \‘{e} \^{i} \"{o} \ae \AE \L u ˘ ñ ˝ H oo œ Œ ß \u{u} \~{n} \H{H} \t{oo} \oe \OE \ss ç d .{x} \={z} \v{v} \o \O !‘ Table 30: Commands for Counters addtocounter ref thecounter label refstepcounter value newcounter setcounter pageref stepcounter Table 31: Commands/Environments to Organize Document abstract appendix listoffigures maketitle subsection tableofcontents addcontentsline bibliography listoftables printindex subsubsection thanks addtocontents bibliographystyle makeindex section subsubsubsection thebibliography Table 32: Commands to Control Document Style markright renewcommand markboth setlength pagenumbering thispagestyle pagestyle Table 33: Commands to Control Fonts in Math Mode left cal mathit mathtt right boldmath (set displaystyle mathnormal mbox textstyle in text mode) mathbf mathcal mathrm mathsf overbrace overline underbrace underline . b ¯ ... å Å ¿ \c{c} \d{d} \b{b} \dots \aa \AA ?‘ x ˙ ¯ z v ˇ ø Ø ¡ \.

y x\quad y x\qquad y x\!y x\negmedspace y x\negthickspace y ⇒ ⇒ ⇒ ⇒ ⇒ ⇒ ⇒ ⇒ What you see xy xy xy x y x y xy x y x y no space thin space medium space space = 1em space = 2em negative thin space negative medium space negative thick space Table 36: Greek and Special Letters α β γ δ \alpha \beta \gamma \delta \epsilon ε \varepsilon ζ \zeta η \eta Γ \Gamma ∆ \Delta Θ \Theta ℵ \aleph A . to avoid the clash between the accent and dot.114 APPENDIX Table 34: Accents in Math Mode a \check{a} e ˇ ˘ x \dot{x} ˙ y ¨ ˆ \hat{\imath} ı xyz \widehat{xyz} \breve{e} ´ \acute{i} i \ddot{y} z \bar{z} ¯  \tilde{\jmath} ˜ abc \widetilde{abc} o ` v \grave{o} \vec{v} \hbar (Note that it is better style to use \imath. . Z {\mathcal \theta \vartheta \iota \kappa \lambda \mu \nu$ \xi \Lambda \Xi \Pi \ell A. rather than i..) Table 35: Spacing Commands in Math Mode What you write x y x\. rather than j..Z} θ ϑ ι κ λ µ ν ξ Λ Ξ Π o π ρ σ ς Σ Υ Φ o \pi \varpi \rho \varrho \sigma \varsigm \Sigma \Upsilon \Phi \Re τ υ φ ϕ χ ψ ω Ψ Ω \tau \upsilon \phi \varphi \chi \psi \omega \Psi \Omega \Im . . and \jmath.y x\.

= |= ∝ ∈ \equiv \doteq \models \propto \in \ni ⊂ ⊆ ⊃ ⊇ .APPENDIX 115 Table 37: Frequently Used Mathematical Symbols {superscript} {subscript} ^{} _{} \prime ∞ \infty ∅ \emptyset Table 38: Binary Operations ± × ÷ \ \ \pm \mp \times \div \setminus \backslash \biguplus ∩ ∧ ∨ \cap \sqcap \wedge \vee \bigcap \bigvee \bigwedge ∪ \cup \sqcup \uplus \oplus \bigcup \bigoplus \bigsqcup ⊗ \odot \otimes \oslash \ominus \bigodot \bigotimes ⊕ Table 39: Operators and Quantifiers ∀ \nabla \forall ∂ ∃ \partial \exists √ ¬ \surd \neg ℘ \wp Table 40: Special Functions arccos csc lg Pr arcsin det lim sec arctan dim liminf sin arg exp limsup sinh cos gcd ln sup cosh hom log tan cot inf max tanh coth ker min Table 41: Relation Symbols ≤ \leq \prec \preceq \ll \subset \subseteq \sqsubseteq ≥ \geq \succ \succeq \gg \supset \supseteq \sqsupseteq = ∼ ∼ = ≈ \neq \sim \simeq \cong \asymp \approx ≡ .

• .. \circ \cdots \vdots \frown \triangle \triangleright \bigtriangleup \bowtie \top \dashv \angle \mid . Triangles and Lines ◦ ··· . Circles. \bigcirc \ddots \bullet \smile \diamond \triangleleft \bigtriangledown \perp \bot \vdash \| \parallel ⊥ ⊥ ∠ | .116 APPENDIX Table 42: Arrows ← ⇐ → ⇒ ↔ ⇔ → ← \leftarrow \Leftarrow \rightarrow \Rightarrow \leftrightarrow \Leftrightarrow \mapsto \hookleftarrow \leftharpoonup \leftharpoondown \rightleftharpoons ←− ⇐= −→ =⇒ ←→ ⇐⇒ −→ → \longleftarrow \Longleftarrow \longrightarrow \Longrightarrow \longleftrightarrow \Longleftrightarrow \longmapsto \hookrightarrow \rightharpoonup \rightharpoondown ↑ ⇑ ↓ ⇓ \uparrow \Uparrow \downarrow \Downarrow \updownarrow \Updownarrow \nearrow \searrow \swarrow$ \nwarrow Table 43: Dots. . .

Some Tips

117

Table 44: Variable Size Symbols
\sum \prod \coprod
n d

\int . . √ . \overbrace{ . } \overline{ . } \sqrt{ . } \{ \} \lfloor \lceil \rangle [] . .

\oint \underbrace{ . } \underline{ . }

\frac{n}{d} ( | \backslash \langle )

These use \left and \right () | \ {} [ ] \rfloor \rceil

Table 45: Special Symbols in Both Text and Math Modes

‡ ...

\dag \ddag \ldots

§ ¶

\S \P

c £

\copyright \pounds

Table 46: Commands and Parameters in Picture Environment
put(x, y){stuff } multiput(x, y)(∆x, ∆y){number}{stuff } line(x, y){length} framebox(width, height)[p]{text} vector(x, y){length} dashbox{dashs ize}(width, height)[p]{text} circle{radius) makebox(width, height)[p]{text} circle*{radius) oval(width, height)[p] linethickness{dimension} p ∈ {l,r,t,b,lt,lb,rt,rb}. For oval, it is the portion selected; for boxes, p is where the text goes.

Some Tips
Here are some tips that apply in special situations. Structuring large documents. Theses, books, and other large documents are best managed with the \input command. The document might look like this:

118 Preamble \begin\{document} \input{Abstract} \input{Chapter1} \input{Chapter2} . . . \end\{document}
(a) Separate chapters in a document (input files are Abstract.tex, Chapter1.tex, . . . )

Some Tips Preamble \begin\{document} %\input{Abstract} \input{Chapter1} %\input{Chapter2} . . . \end\{document}
(b) Compiling just Chapter 1

This is also a good structure for co-authoring. Counting words. Under linux enter dvi2tty file.dvi | wc --words This takes file.dvi as input into the program dvi2tty, which converts it into a plain text file (without latex commands). That output file is passed to the command wc, which counts the words. If you need to apply this to a portion of the document, such as an abstract, you can put that portion into a skeleton latex file and compile it. Using the above structure, this is particularly easy to obtain the word or line count of any section. Commenting text. Besides putting % in column 1, you can logically delete a block of text by \if 0 ... \fi. This is a good way to manage a document when you might want to remove some text with the idea that you it might be restored, at least in part. \if 0 stuff here is ignored \fi (See §8.5 for more general control.) \newcommand vs. \def. It is generally better to use \newcommand because you will get an error if the command name is already defined, such as by some package. The syntax is \newcommand{\name}[#]{what to do}; for example, \newcommand{\note}[1]{\texttt{Note: #1}} defines a \note command that has one argument. When you specify \note{hello}, you will get Note: hello in the text. \def is to be used only when you want to allow an over-ride (without warning). Its syntax is \def\name{what} or \def\name#1{what} if there is one argument. To suppress printing all notes, defined by the above \note command, you can specify \def\note#1{ }. This says to do nothing when seeing \note{stuff}. You can have an \input file with many commands and environment definitions that you use repeatedly. It might have something like the following:

REFERENCES
\newcommand{\note}[1]{\texttt{note: #1}} % \def\note #1{ } % activate to suppress printing notes

119

A “better” alternative (avoiding \def) is to use
%\renewcommand{\note}[#1]{} % activate to suppress printing notes

The \note command is useful when collaborating, and you could define a separate one for each author, showing initials. There is also the \providecommand. This has the same syntax as \newcommand, but it does not issue an error if the command is already defined. It does nothing if the command exists; otherwise, it defines the command. Making clear. It is good practice to use \ell in math mode when you want the letter ell because it is easier to distinguish \ell from the number 1. Compare vs. l with 1 — also compare the source: $\ell$ vs. $l$ with $1$. Also consult the TE X Catalog Online [14].

References
[1] Johannes L. Braams. Babel, a multilingual style-option system for use A with L TEX’s standard document styles. TUGboat, 12(2):291–301, 1991. Available at CTAN [4]. [2] Johannes L. Braams, David P. Carlisle, Alan Jeffrey, Frank Mittelbach, A Chris Rowley, and Rainer Schöpf. L TEX 2ε and the LaTeX3 project. World Wide Web, http://www.latex-project.org/latex3.html, 1994–99. [3] David P. Carlisle. Packages in the ‘graphics’ bundle. World Wide Web, CTAN /macros/latex/required/graphics/ (see [4] for replacing CTAN), 1994–99. [4] Comprehensive TEX archive (CTAN). UK: ftp.tex.ac.uk/tex-archive/; Germany: ftp.dante.de/tex-archive/; USA: ftp.tug2.cs.umb.edu/texarchive/ . These are host sites, which contain a list of mirror sites.
A [5] Michel Goosens, Frank Mittelbach, and Alexander Samarin. The L TEX Companion. Addison-Wesley Publishing Company, Reading, MA, 1994.

[6] John D. Hobby. A User’s Manual for MetaPost. Computing Science Technical Report no. 162, AT&T Bell Laboratories, Murray Hill, New Jersey, 1992. Available at http://cm.bell-labs.com/who/hobby/ MetaPost.html/.

World Wide Web. CTAN /macros/latex/doc/usrguide. http://www. Addison-Wesley Publishing Company. Hobby. http: [14] Graham Williams.ps.2). A [12] Keith Reckdahl. Available at http://cm. 164. 15th edition. 1998–99 (version 1.html. 1993–98. Reading.gz. 1993. BibTE Xing. CTAN /info/epslatex.ctan. Knuth. AddisonWesley Publishing Company. New Jersey. World Wide Web site Version 2. L TEX: A Document Preparation System. [13] Christian Schenk.ps. [11] Oren Patashnik. Computing Science Technical Report no. 1995–97. [15] Timothy Van Zandt.bell-labs. The TEX Book. AT&T Bell Laboratories.0. World Wide Web. World Wide Web. MA. Drawing Graphs with MetaPost. //www. 1989. Murray Hill.120 REFERENCES [7] John D.de/. PSTricks: PostScript macros for Generic TeX. MiKTE X Local Guide. 1995–99. Using imported graphics in L TEX 2ε .ps (see [4] for replacing CTAN). . 1994).org/tex-archive/help/Catalogue/catalogue. Comprehensive TEX Archive. MA.edu/ depts/adn/infwww/ps/btxdoc.com/cs/cstr/164. CTAN.uic. Reading.tug.ps (see [4] for replacing CTAN). http://www. 1986 (also see 2nd edition.miktex. A [9] Leslie Lamport.org/applications/PSTricks/. Technical report. TE X Catalogue Online. 1988. [8] Donald E. A [10] L TEX 2ε for authors. http://www.

71. 10 \nolinebreak. 118 \displaystyle. 7. 42 \fbox. 25 \hspace. 25 \hline. 25 \imath. 93 \dotfill. 47 \mbox. 55 \baselinestretch. 23. 103 \newline. 74 \frame. 99 \frac. 24 \pagenumbering. 63 \Big. 42 \fboxsep. 101. 25 \dots. 72 \date. 93 \baselineskip. 101 \newcommand. 25 121 \hspace*. 39. 61. 117 \linebreak. 117 \itemsep. 9. 107 \author. 8 \dvi2tty. 42 \newenvironment. 65 \pagebreak. 51. 37 \clearpage. 23 \label. 118 \newcounter. 19. 55 \documentclass. 12. 28. 108 \bibliographystyle. 19 \cite. 98 \maketitle. 3 \ensuremath. 25. 114 \kill. 56 \overline. 106 \jmath. 97 \pageref. 11 \big. 1. 24 \newpage. 60 \framebox. 63 \bigskip. 57. 93 \mathbb. 108 \fboxrule. 99 \dvips. 62 \nocite. 36 \bigg. 56 \overset. 11 \multicolumn. 23 \line. 56. 48. 114 \input. 53 \left. 36 \bibliography. 41 \cline. 48. 41. 40. 74 \footnotesep. 55. 118 \dvipdf. 41 \newtheorem. 24 \oddsidemargin. 97 \makebox. 58 \medskip. 59 \mathfont. 93 \def. 47 \cdots. 24 \noindent. 104 \overbrace. 25. 72 \makeindex. 54 \hrulefill. 37 \nohyphens. 9 \hfill. 74 \listoffigures. 36. 101 \equal. 59 \mathscr. 24. 56. 15. 62 . 61 \centerline. 10. 4. 97 \listoftables. 63 \boldmath. 16 \dashbox. 43 \arraystretch.Index \Bigg. 24 \nopagebreak. 63 \addtocounter. 71. 100.

1. 43 \vector. 65 INDEX \unitlength. 94 \theenumi. 35 dash. 19. 62 \renewcommand. 25. 47 box. 4 emftoeps. 77 \value. 86 . 25. 95 \thecounter. 56 \\. 21. 56 counting words. 103 \right. 54 \printindex. 43 \setlength. 61 Bezier curve. 33 \usepackage. 28 body. 119 \renewenvironment. 93 \underbrace. 99 \smallskip. 97. 118 column specification. 78 titlepage. 1 amsmath. 24. 60 document styles. 119 \psset. 3. 117 \samepage. 48 \textwidth. 9. 1 conditional assignment. 72 \parindent. 98 \section. 43 \subsection. 3 dvips. 74 \verb. 43 \ref. 56 \underline. 98 \prod. 73 \url. 100. 42. 1 boldmath. 118 cover page. 60 \prime. 18. 25. 24 \section*. 11 \sqrt. 15 pdftex. 56. 56 \widetilde. 22 AMS. 16. 106 \parsep. 55 \widehat. 8 \thanks. 39. 108 \parbox. 16 command line. 72. 43 \thefootnote. 118 \vfill. 104 \textfont. 97 \textstyle. 5 \substack. 57 \raisebox. 59. 72. 106 \partial. 93 accents. 48 \providecommand. 5 \selectfont. 72 co-authoring. 58. 61. 93 cross referencing. 28 dvi viewer. 48 \stackrel. 10 debugging. 25 \vspace*. 57 \refstepcounter. 65 \stepcounter.122 \pagestyle. 39. 78 \raggedright. 98. 53. 43 \title. 43. 38 \setcounter. 20 comments. 1 comment. 56 \underset. 41. 84 bibtex program. 1 DOS. 65 \tableofcontents. 2 derivative. 34. 118 compile. 106 \parskip. 25 \vspace. 40.

32. 41 flushleft. 42 123 . 3 global setting. 82 Hamiltonian. 28. 41 floating object. 93 enumerate. 25 bibliography. 59 latex command. 54. 62 center. 58 non-English. 47. 43 graph. 1. 9 smallmatrix. 14 itemize. 18. 10 quote. 10 hyphenation. 29. 40. 14 eqnarray. 13 local setting. 9. 6 abstract. 3. 40 font size. 86 jpeg. 110 dvi. 11 file bib. 1. 8 boldmath. 12 description. 98 jpeg2ps. 109 theorem. 56 thebibliography. 16 sans serif. 20 picture. 16 typewriter. 57 flushright. 16 math mode. 6 corollary. 96. 61 verbatim. 62 description. 60 horizontal fill. 53 eqnarray*. 86 float. 36. 54. 9 longtable. 41. 41 tabular. 51. 47 Greek. 6 gather. 54 figure. 64 itemize. 1. 8. 47 italic. 8. 21. 19. 15. 38 list environment. 86 tex.INDEX environment. 24 index (making of). 64 gather*. 9 boldface. 40 float page. 47 calligraphic. 8 underlined. 22 Roman. 8 footnote. 95. 1 line spacing. 8 small caps. 6. 22 table. 78 eps. 34–36 bst (bib style). 32 wmf. 12. 86 ps (postscript). 99 fractions. 64. 4. 78. 72 quotation. 81 Laplace transform. 1. 98 array. 40. 53 equation. 59 landscape. 13 large. 86 Lagrangian. 98 appendix. 11. 8 slanted. 21 document. 65 tabbing. 57 axiom. 12 enumerate. 21 verse. 48 ghostview. 55 font style bold small caps.

110 page numbering. 52 mathbf. 9 frankenstein. 41 fontenc. 110 bm. 48. 2 MetaPost. 58 stacking. 103 \textspec . 16. 77. 20 makeidx. 65 amsmath. 20. 33 wasysym. 22 bibunits. 45. 12. 110 amscd. 110 theorem. 99 algorithm. 22 table. . 104 gloss. 58 units of measurement. 104 ~. 1. 47 matrix equation. 1. 25. 10 rotate. 108 paragraph positions. 46 stack. 88 pstricks. 1 spacing. 25 list. 24 horizontal. 2 Repeated entry. 21. . 2 warning. 54 message. 50. 25 line. 99 graphicx. 110 hyphenat.124 make index. 107 vertical. 58 babel. 87 graphtex. 54 trigonometric functions. 111 float. 86 nodes. 5 superscript. 72. 6 preamble. 59 moreverb. 110 fancyvrb. 8 acronym. 5 SIAM. 99 psfrag. 33 special function. 111 nomencl. 12. 64 amssymb. 11. . ix. 2 Overfull . 61. 78. 110 xypic. 110 longtable. 25 special character. 81 package. 15. 35. 78 MiKTeX. 77 qsymbols. 1. 4. 46 tabbing commands. 108 listings. 63 url. . 106 math mode. 79. 65 subsection. 15 ticks. 98 math display mode. . 103. 65 subscript. 110 quotation marks. 98 mathrsfs. 110 geometry. 108 showkeys. 21 in url. 110 setspace. 45. 110 chemsym. 85 transpose. 38. 48. 3. 11. 48 c-pascal. 24. 84. 81. 4. 40 SIunits. . 111 . 37 Underfull . 24 ifthen. 81 INDEX section. 29.

118 xdvi. 86 wc. 3. 118 word count.INDEX unix. 3 YAP. 3 125 . 28.

Sign up to vote on this title
UsefulNot useful