A Simpliﬁed Introduction to L
A
T
E
X
Harvey J. Greenberg
University of Colorado at Denver
Mathematics Department
PO Box 173364
Denver, CO 802173364
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
+ε
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 Kahnooth) 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 Lahtek or Laytek) 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 indepth 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:]{:tnc}
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−→MSDOS 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 nonfatal 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 myfile1.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}{lcr} \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}{lcc}
Name & Test 1 & Test 2 \\ \cline{11}
Bob & 67 & 72 \\
Sue & 72 & 67 \\ \cline{23}
\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{33} 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{33}
\\
\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{22}
& 3 & 4 \\ \cline{11}\cline{33}
\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}{ll} \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}{ll} \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}{ll} \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}{ll} \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}{lccc}
& \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{23}
\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 nonEnglish 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 ‘13’, making ‘468’ 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
13 sting like a bee
468 don’t be late
\begin{tabbing}
13 \= sting like a bee \\
468 \> don’t be late \\
\end{tabbing}
13 sting like a bee
468 don’t be late
\begin{tabbing}
468 \= don’t be late \kill
13 \> sting like a bee \\
468 \> 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 email 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 3column 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 = "AddisonWesley 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 casesensitive, 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 = "199599",
}
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 = "AddisonWesley 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.latexproject/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:FAA89,
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:FAA89311,
crossref = "Byrnes:FAA89",
author = "N.K. Art{\’e}miadis",
title = "Results on the Absolutely Convergent Series
of Functions and of Distributions",
pages = "311316",
}
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 2volume 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 singlespaced, 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 problemsolving — i.e.,
theoremproving — 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 nonintrinsic 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 reentering. 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$\tilde A\times\vec{1}\otimes\overline{2}$} ⇒
˜
A×
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:oi = 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 3column 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, {cc}, 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
\[ \left[ \begin{array}{cc}
\begin{array}{ccc} A_{11} & A_{12} & A_{13} \\
A_{21} & A_{22} & A_{23}
\end{array}
& 0 \\ \hline
0 & \begin{array}{cc} B_{11} & B_{12} \\ B_{21} & B_{22}
\end{array}
\end{array} \right]
\]
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 inline 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
\[ \int_0^\infty \int_0^{x_n} \int_0^{x_{n1}} \cdots
\int_0^{x_2} H(x_1,\dots,x_n)\,dx_1\cdots dx_n
\]
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 theoremlike 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 nvectors: C
n
.
(d) Nonnegative rational nvectors: 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 nonsingular if it is not singular.
Theorem 1 Every nonsquare matrix has an inverse.
Proof: The determinate of a nonsquare matrix cannot be zero
because it is not deﬁned. Therefore, the matrix is nonsingular.
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 +
n·
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 nvector 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
= −
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 rcoordinate is
r
0
+ cn. Then, the new ncoordinate 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 integervalued 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 rcoordinate, 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 doubleheaded, 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 ncoordinate of
the center of the circle, and its rcoordinate 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 recalculations
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{pstall} for the entire system. (You
can use parts, in which case you specify the parts you use instead of pstall
— 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 picturedrawing 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 openended 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
SPLUS
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 raxis 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 SPLUS
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{StreetSmart}
\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
StreetSmart
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 redeﬁning 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 twosided, 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 level1 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 pcolumn 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}, AddisonWesley
Publishing Company, Reading, MA, 1994.
\bibitem{tex} Donald E. Knuth, \textit{The \TeX\ Book},
AddisonWesley 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.,
cpascal 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,pstall} % 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 coauthoring.
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 override (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 styleoption 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.latexproject.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/texarchive/;
Germany: ftp.dante.de/texarchive/; 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. AddisonWesley 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.belllabs.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.belllabs.com/cs/cstr/164.ps.gz.
[8] Donald E. Knuth. The T
E
X Book. AddisonWesley 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/texarchive/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
coauthoring, 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
nonEnglish, 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
cpascal, 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
. Adding bibtex to the Command Sequence . . . . . . Positioning Paragraphs Source (Result in Figure 7) . . . . . A Document to Print the Bibliographic Database . . . . . . . . . . . Tabbing Source (Result in Figure 30) . . . . . . . . . . . . . . . . . . . . . . . . . . A 2 × 3 Table with Horizontal and Vertical Lines . . . . . . . Description List Environment . . . . . . . A Table with Partially Spanning Horizontal and Vertical Lines Nested Tables Source (Result in Figure 23) . . \parbox Result (Source in Figure 24) .
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
. . . . . . .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. Variable Sizes Result (Source in Figure 38) . . . . . . . . . . . . Framed Figure 34 Source . . . . . . . . . . . . . . . . . . \displaystyle Result (Source in Figure 40) . . . . . . . . . . . . Some Font Sizes Source (Result in Figure 11) . . . . . . . . . . Multicolumn Result (Source in Figure 26) . . . . . . . . . . . . . . . . . . . . A 2 × 3 Table . . . . . . . A Your First L TEX Source File . . . . . . . . . . Framed Figure with Caption at Bottom . Multicolumn Source (Result in Figure 27) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Skipping Line Spaces Result (Source in Figure 12) . . . . . \displaystyle Source (Result in Figure 41) . . . . . . . . . . . . . . . . . . Some Font Sizes Result (Source in Figure 10) . An Introductory Document Result (Source in Figure 4) . . . . . . . . . . . \parbox Source (Result in Figure 25) . . . . . . . . . Framed Figure with Caption at Top . . . . . . . . Enumerate List Environment Result (Source in Figure 17) . . . . . . . . . . . . . . . . . . . . . . Enumerate List Environment Source (Result in Figure 18) . . . . . . . . . Centering Source (Result in Figure 9) . Obtaining Brackets in a Description List Environment . . . . . . . . . . . . . . . Positioning Paragraphs Result (Source in Figure 6) . . . . . Nested Tables Result (Source in Figure 22) . . . . . . . . . . Tabbing Result (Source in Figure 29) . Itemize List Environment Source (Result in Figure 16) . . . . Alternative enumerate Symbols Source (Result in Figure 37) Alternative enumerate Symbols Result (Source in Figure 36) Variable Sizes Source (Result in Figure 39) . . . . . . . . Itemize List Environment Result (Source in Figure 15) . . . Skipping Line Spaces Source (Result in Figure 13) . . . . . . Centering Result (Source in Figure 8) . . . . . . . . . . Command Sequence from Source to Postscript or PDF . . . . . . . . An Introductory Document Source (Result in Figure 5) . . . . .
. . . . . . . . . . 55 . 94 . PSTricks Source for Connecting Nodes . . . . . . . . . . . . 73 . . . . . . . Making an Abstract Result (Source in Figure 78) . . . . 54 . . 50 . . Some Front Matter Speciﬁcations for This Document . . Sequence of PSTricks Commands to Draw Histogram . . . . . . . . . . . . . Matrix Equation Result (Source in Figure 45) . . . . . . . gather* Environment Result (Source in Figure 55) . . . . . . . . . . . 82 . . . . . . . . . 93 . . 75 . 64 . . eqnarray Environment Result (Source in Figure 43) . 58 . 95 . Variety of Objects in Picture Environment . Setting a Footnote Source (Result in Figure 85) . . . . . 58 . . . Matrix Equation Source (Result in Figure 46) . . . . . . . . . Title Page Source (Result in Figure 72) . . . . . Title Page Result (Source in Figure 71) . 99 . Adding Addresses to Authors . . . . . . . 99 . 99 . . . . . 100
. . . . . . . . . Commutative Diagram Source (Result in Figure 58) . . 57 . . 53 . 83 . . . . . . . . . . . 56 . . \flushleft in parbox Result (Source in Figure 51) . . . . 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) . . 66 . . . . . \flushleft in parbox Source (Result in Figure 52) . . . . 84 . . . 55 . . . . . . . . . 88 . . . Line Parameters . . . . . . . . . Graph Result (Source in Figure 65) . . . . . . . . . Commutative Diagram Result (Source in Figure 57) . . . . . . . . . . . 56 . . . . Setting a Footnote Source (Result in Figure 83) . . Adding makeindex to the Command Sequence . . Vertical Diagram Result (Source in Figure 59) . . . . . Applying \includegraphics to Import an eps File . Graph Source (Result in Figure 66) . Setting a Footnote Result (Source in Figure 82) . . . . . \raggedright in parbox Result (Source in Figure 53) . . . . 72 . 82 . . . . . . . . . . . . \raggedright in parbox Source (Result in Figure 54) . 93 . 64 . . . . . 95 . . Specifying Dimensions in \includegraphics . . . 72 . 53 . . eqnarray Environment Source (Result in Figure 44) . 96 . . . . . Nested Arrays Result (Source in Figure 47) . . . Source Code for Drawing Histogram of Test Scores . . . . . . 88 . . . . . . . . . . . . . gather* Environment Source (Result in Figure 56) . 87 . Making an Abstract Source (Result in Figure 79) . . . . . . . . Authors with same footnote (Source in Figure 76) . Source for Figure 61 . 54 . . . . . . 95 . 96 . . . . . . . . . . . 97 . 72 . . . . . . . . . 97 . . . . . Nested Arrays Source (Result in Figure 48) . . . . . . . 66 .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 . . Vertical Diagram Source (Result in Figure 60) . . . 57 . . Horizontal Braces Result (Source in Figure 49) . Horizontal Braces Source (Result in Figure 50) . . . .
. . . . Document Margins . . . . . . . . . . . Most of the Preamble for this Book . . . . . . . . . Changing the Left Margin of a List (Source in Figure 88) . . . . . . 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) . . . . . . . . . . Varying \itemsep to control item spacing in a list . . . . . . . . . . . . . . . . . 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
. . . . . . . . . . \ifthenelse Results (Source in Figure 92) . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Some Mathematical Operations . . . . . . . . . . . Order Relations . . . . . . . . The Tabbing Environment . . . . . . Figure and Table Location Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Spacing Commands in Math Mode . . . . . . . . . . . . . . . . . . . . . Conversions of Common Units of Measurement . . . . . . . . . . . . . . . . . . . . Accents in Math Mode . . . . . . . . Commands/Environments for Controlling Text Position Text Accents and Special Symbols . . . . . . . . . . . . . . . . . . . Operators and Quantiﬁers . . . . . . . . . . . . . . . . Commands to Control Document Style . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Some Basic Drawing Commands in PSTricks . . Commands/Environments to Organize Document . . Commands to Control Fonts in Math Mode . . . . . . . Some Common Mathematical Functions . Commands/Environments for Text Font Appearance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Transpose of a Vector . . . . Default Settings for enumerate Counters . . . . . . . . . . . . . . . Special Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Spacing Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . Boxes in PSTricks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Parameters for \psaxes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Numerals to Print Counters . . . . Reference Tables . . . . . . . . . . . . . . . . . . . . . Binary Operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Examples of Mathematical Functions . . . . . . . . . . The \mathfont Commands . . . . . . . . . . . . . . . . . . . . . . . The \kill Tabbing Command . . . . . . . . . . . . . . Commands for Counters . . . . . . . . Variable Size Mathematical Operation Symbols . . . Intrinsic Name Parameters . . . . . . . . . . . . . . . . . . . . . . 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
. . . . . .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 . . . . . . . . . . Writing Special Characters . . . . . . . . . . . . . . . . Set Notation . . . Some Symbols in Logic . . . . . . . . . . . Greek and Special Letters . . Frequently Used Mathematical Symbols . . . . . . . . . . . . . . . . . . . . Some Accents for Letters . . . . Margin Parameters . . . Notation Using mathbb Fonts from amssymb Package . . . . . . . . . . Relation Symbols . . . . . . . . . . . .
. . . . . . . . . Triangles and Lines . . Commands and Parameters in Picture Environment .vii 42 43 44 45 46 Arrows . . . . . . . . . . . . . 116 116 117 117 117
. . . . . . . . . . . . . Circles. Special Symbols in Both Text and Math Modes . Dots. . . . . . . . . . . . . . . . . . . . . . . . . . . . . Variable Size Symbols . . . . . . . . . . . . .
omitting many of the things L TEX 2ε can do. Frank Mittelbach. L TEX (pronounced Lahtek or Laytek) is a collection of macros built on top of TEX that “represents a balance between A functionality and ease of use” [9. L TEX 2ε is the current version. which should be just a few classes. and he permanently changed the way mathematical A documents are prepared. It was 10 years later that he published his seminal book [8]. p. it will take longer. based on TEX.viii
Preface
A The majority of this book is about using L TEX 2ε [2. low cost publications in mathematics and related disciplines that caused Knuth (pronounced Kahnooth) to invent TEX (pronounced Tek) in the late 1970’s. he actually ended up deﬁning an entire branch of research in computer science. especially getting used to functioning at the command line. 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. Knuth [8]. such as italic font or a complex mathematical expression. a descendant of A L TEX. you type text freely until you need something special. For one who is well acquainted with computers. not a word processor. This is a typesetting program. like ai +ε
1 + (x − µ)2 dx Φ(ε) . or article. xiii]. this book is designed to given by the The L TE A be a succinct introduction. particularly unix. A A comprehensive coverage of L TEX and the many enhancements to it is A X Companion [5]. the basics that are covered should take less than 10 hours. and Rainer Schöpf [2]. using L TEX 2ε to produce high quality results. David P. Alan Jeﬀrey. but he published articles along the way. — Harvey J.
ε→0+
lim
ai
It was the desire to have high quality. Carlisle. Originally believing that he could write a program in less than a year that could typeset documents. Colorado
. originated by Donald E. In any case. By contrast. You enter some editor that saves plain text ﬁles. the ﬁner points require more study. Then. 10]. designed by Leslie Lamport [9]. developed by a team of volunteers: Johannes L. For one who is just learning how to use a computer. Greenberg Denver. thesis. Chris Rowley. and one could do all of the exercises.
A Happy L TEXing. Braams. Exercises are provided for guided instruction.
Mårten Svantesson.ix
Acknowledgments
The author thanks the many contributors in the comp. Bernd Schandl. Graversen whose indepth review has made this version much better than my original. Jonathan Fine. I especially thank Kasper B.tex newsgroup. USA. Last. MA. Thomas Ruedas. and MiKTeX [13] is available free of charge for DOS systems.ac. These all describe how to search and browse the FTP sites for software and documents. and 3. I also received useful comments from people who read an earlier draft that I made available on the web. Andrea Dean. particularly Donald Arseneau. http://www.
. Sebastian Rahtz. is at the Comprehensive TeX Archive Network (CTAN) [4].text. Herman Bruyninckx. David Carlisle. One student. David Haller. at three host sites (and many mirrors):
1. The best A source of these. http://www. who were very generous with taking time to answer so many questions on a regular basis. Robin Fairbairns. Anton Schwaighofer. provided feedback that led to several points of clariﬁcation. notably William Briggs. and Matt Swift. Timothy Murphy. but not least.
A Sources of L TEX Software
A The basic L TEX software system is available free of charge for unix systems. 2. Axel Reichert. Holder. who A taught me L TEX in the ﬁrst place. http://ctan. I thank Allen G.uk/ in Cambridge.org/ in Boston.dante. Germany (in German). UK. Dan Luecking.de/ in Mainz. and additional packages that extend the L TEX capabilities (to which I refer in this book).tug. Denis Girou.tex.
book.tex % notes to yourself can go here \documentclass[options]{style} optional speciﬁcations — e. which creates several output ﬁles. \end{document}
Anything following % is ignored (used for comments).tex. as well as more advanced topics for customizing your document. the intrinsic document styles are: article. For example. at the command line. and slides. . the command line is the DOS command line. A Are you ready to write your ﬁrst L TEX document? Copy the source ﬁle shown in Figure 2 and name it myfile. depending upon the style. enter: latex myfile (In an MS Windows system.
A Figure 1: The Structure of a L TEX Document.
In the preamble. Most publishers have their own style. The only one you need to be concerned with now is the dvi ﬁle. its suﬃx is . Figure 1 shows the structure of this ﬁle. which is plain text.tex. like margin settings and other preamble speciﬁcations are covered. Preamble (blank lines do not matter)
Body This is the document environment. letter. which they provide for you to download.tex as a plain text source ﬁle that you create..
% This is myfile. To keep things simple. which I shall describe in greater detail throughout this book. concentrating on getting started A with using L TEX as quickly as possible. which the
. notably the American Mathematical Society (amsmath style) and the Society for Industrial and Applied Mathematics (siam style).g. All that follows is ignored (could be used for comments).) This is called compiling your source. there are many options. declaring use of packages \begin{document} . I refer to myfile. some of the options. I shall be using defaults for almost everything. . or by Start−→Run and entering cmd or command into the window. which you enter by running Start−→Programs−→MSDOS Prompt. Then. called the L TEX source. Among these are professional societies. The focus throughout this book is on the article style. report. Later.1
1
Overview
A You will create a ﬁle. Further.
tex destination. \end{document}
A Figure 2: Your First L TEX Source File
latex program (called a “compiler”) names myfile.. which you will come to know and love). If you are running under DOS. such as missing a brace (characters { and }.. (If you are familiar with
. so you have a collection of backups. This is called debugging your source.) Change 1 to another qualiﬁer each time (e. the ﬁrst thing to do is save a backup by copying your source ﬁle to another subdirectory. 2 . That way you will know that what you did in the last few minutes contains the error. . so you are advised to compile often..... You got no messages. Among the nonfatal messages you will generally see are warnings like: Overfull \hbox . Do not worry about these.. or to a diﬀerent name. but they were not fatal errors. You must ﬁnd and correct it.tex myfile1. You got messages. Something went wrong and you need to ask for help. Underfull \vbox . use copy myfile. where the destination is either a: or some backup ﬁle name.. ). Case 3.. One of three things will have occurred: Case 1. Sometimes the error message tells you what went wrong. Overfull \vbox .dvi. You got a fatal error message.2
1 OVERVIEW
\documentclass{article} \begin{document} Hello world. Many times the message is not very informative. If all went well. or some command was not recognized due to being misspelled. Underfull \hbox . .g.tex (The p is to keep the date and time of the source ﬁle. In unix do this by entering: cp p myfile. Case 2.
edit.ps. it will use .tex (otherwise. In unix this could be pico.tex
. you begin by entering a plain text editor. the viewer that comes with MiKTeX [13]. Summarizing. Under unix. including posting it on the web. To view or print a postscript ﬁle under DOS.ps. At the command line enter: xdvi myfile and it will come on your screen. xdvi does not have a print option. so you ﬁrst need to convert the dvi ﬁle to postscript. If you use a word processor.. and you must continue to specify the suﬃx as . You will need to ﬁnd out more about viewing and printing postscript ﬁles that suit your particular needs. in which case the o is not needed. press Enter at the end of a line instead of letting the word processor do it for you). xdvi. At the DOS command line you enter: YAP myfile You will see various options for viewing and printing. you need to learn how to create. when saving the ﬁle. You should even put in hard return characters (i. your installation might already have ﬁle output as the default. dvips. rather than just print. is called YAP. in which case you might ask someone for help.e.3 DOS. you can run a program called ghostview. and save plain text ﬁles. including the unix command. emacs. enter EDIT myfile.doc as its suﬃx). or you can use Notepad.) Next you want to view the result. be sure to specify plain text. If you want to check that the ﬁle is really plain text. At the command line enter: dvips myfile o (The o tells the system you want the output to go to a ﬁle. In MS Word. You can print it by any number of ways. lpr myfile. In DOS you can use EDIT at the command line. and never use tabs. vi or vim. The same conversion program can be run under DOS (and comes with MiKTeX).) If you are using DOS. and you might want to obtain myfile. you must take absolutely no advantage of its formatting. (There is more to do if you are working remotely. This is done with the program. you can view the result with the dvi viewer. If you are in a unix environment. a free software system by Christian Schenk.ps for a variety of reasons.) This will result in the creation of the postscript ﬁle myfile. nothing more need be said. however. if not. Wordpad or MS Word.
which can be printed. use your favorite search engine to ﬁnd them. 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. A This book is designed for quick entry into using L TEX.dvi dvipdf
print/post
myfile. and if all is well. which is a printer measure equal to the width of M in the current font). but do not be reluctant to read the last chapter.) The orientation here is by function. cm (centimeters). like notation.ps myfile. (There are some free text editors on the web. about 1 A E 72 in.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. beginning with how to write text.” indicated in Figure 1. enter latex myfile. Congratulations!
create/edit compile latex myfile. and em (like the letter m. It tells you how to deﬁne your own commands and how to separate them into an input ﬁle that simpliﬁes changing things. (One of A the strengths of L TEX is the community of people who provide packages for everyone to use at no cost. enter your dvi viewer.tex
view/print convert dvips myfile. except to say that they are used to fulﬁll some function.4
1 OVERVIEW
at the DOS command line and see how the ﬁle appears. I also cannot elaborate just yet on “using packages.) Once you have your source ready to compile.
. Under unix or DOS use dvips to convert the dvi ﬁle to a postscript (ps) ﬁle. The result should be one line of output: Hello World. and I shall introduce speciﬁc packages throughout this book. other units used in many parts of L T X are in (inches). Execute these commands for the source ﬁle shown in Figure 2.
A will demonstrate the superior quality of the L TEX results. but even some text. especially tables. (The power of L TEX will be evident when we get to mathematical expressions. much like you A would want in a word processor. in the preamble. \begin{document} % Begin document "environment".
\documentclass[12pt]{article} % We have defined the document to be an article using 12 point font.
lots
Skipping four lines is the same as skipping one line . Figure 4 is the source that produces the result in Figure 5.) First.
\section{This is a Section} \subsection{This is a subsection} This is the body of the subsection.5
2
Text
We begin by illustrating the most common text formatting. consider how to make sections and subsections in article style. and I can put in of blanks with no effect. I can move to a new line anytime. and how extra spaces and blank lines have no eﬀect. \subsection{Here is another subsection} \section{Here is another section} \end{document}
Figure 4: An Introductory Document Source (Result in Figure 5)
. Note the automatic numbering. % Blank lines mean nothing here.it starts a new paragraph. showing how sections and subsections are deﬁned.
and have begun the flushleft environment. Note that these are A environments. ﬂush right. and justiﬁed (the default). Text remains centered as long as we remain in this environment. \end{flushright} I am back to normal justification. \end{flushleft} \begin{flushright} This is another paragraph. . I can move to a new line anytime.1
This is a Section
This is a subsection
This is the body of the subsection.1
Fonts and Paragraphs
Figure 6 shows the source to produce diﬀerent paragraph positions: centered. \end{environment}
\begin{center} The text is centered because I have entered the center environment.2
Here is another subsection
2
Here is another section
Figure 5: An Introductory Document Result (Source in Figure 4)
2. The added space you see between the above paragraphs is due to entering those environments. but in the flushright environment. ﬂush left. You will have occasion to use all four paragraph positions.
Figure 6: Positioning Paragraphs Source (Result in Figure 7)
. . \end{center} \begin{flushleft} Now we are out of the centering environment.6
2 TEXT
1
1.
1. Skipping four lines is the same as skipping one line — it starts a new paragraph. and I can put in lots of blanks with no eﬀect. a concept you need to understand about L TEX. The general form of an environment uses the following syntax: \begin{environment} .
This precedes center environment. Now we are out of the centering environment. shown in Figures 8 and 9. and have begun the ﬂushleft environment. I am back to normal justiﬁcation.2. \begin{center} This line is centered. This continues after centering. \centerline{This line is centered. they diﬀer in that the environment skips a line before and after the paragraph. The added space you see between the above paragraphs is due to entering those environments. You will have occasion to use all four paragraph positions. Figure 7: Positioning Paragraphs Result (Source in Figure 6) Instead of the center environment. This is another paragraph. Text remains centered as long as we remain in this environment. \end{center}
Figure 8: Centering Source (Result in Figure 9)
.} This continues after centering. you can use the \centerline command. but in the ﬂushright environment.1 Fonts and Paragraphs
7
The text is centered because I have entered the center environment. This precedes centerline.
\textbf{\textit{bolditalic}} ⇒ bolditalic.) In technical writing. (The ⇒ symbol can be read as “produces. Table 1: Intrinsic Font Styles What you write This is \textbf{boldface}. typewriter.8 This precedes center environment. This is \texttt{typewriter}. as it is used when introducing a new term. For example. where font is one of {bf.
Some combinations of font styles can be produced. you will have particular use for the italic font. The general form is \textfont{text}. sans serif. small caps. The argument of \textbf is \textit{bolditalic}. For example. This is \textit{italic}. sc. italic. This is \textsf{sans serif}.
. 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. rm. sl. as seen in Table 1. sf. This continues after centering. Table 1 lists the fonts that are intrinsic in a basic latex installation. tt}. This continues after centering. slanted. usually free of charge. roman. Here is an example: \noindent This paragraph is not indented. (More fonts are available in packages. This line is centered. it. A \textit{group} is defined on a set of elements \dots ⇒A group is deﬁned on a set of elements . ⇒ ⇒ ⇒ ⇒ ⇒ ⇒ ⇒ How This This This This This This This it is is is is is is is appears boldface. This is \textsc{small caps}. This line is centered. This is \textrm{roman}. .”) Note the use of the \dots command. . produces: This paragraph is not indented. This is \textsl{slanted}. This precedes centerline. which produces the ellipsis.
. You might want to make something small or smaller or even smaller still.1 Fonts and Paragraphs
9
A Not all combinations are in the basic L TEX 2ε installation.
.) It is straightforward to underline text — just write \underline{text}. Here is an example using the quote environment. The environment syntax is useful when you want to keep the size for a large block of text. We can also frame text just by writing \frame{text}. Font size can also be varied at will.
You can make the text {\large large} or {\Large larger} or even {\LARGE larger still}. such as using {\Large\textbf{heading}} for some heading. but learning it must wait until we enter math mode. you must put \usepackage[T1]{fontenc} in your preamble to obtain:
\textbf{\textsc{bold small caps}} ⇒ bold small caps. You can also make it {\huge huge}.}. we can explicitly use the environment syntax: \begin{size} . which was generated by putting \begin{quote} before the text and \end{quote} after it:
huge
larger or even larger
still. (More on framing in §6. Notice how the paragraph spacing changes to accommodate the variation in font size. Optionally. You can make it really tiny. Now consider ways to indent a block of text. \begin{large} This is large. \end{size}. . For example. You might want to make something {\small small} or {\footnotesize smaller} or even {\scriptsize smaller still}. 71. In particular. p.
Figure 10: Some Font Sizes Source (Result in Figure 11)
You can make the text large or
You can also make it . You can make it really {\tiny tiny}. (There is no intrinsic environment for font styles.\end{large} produces the same result as {\large This is large. and the braces format is useful for short phrases. Figures 10 and 11 give the source and result for common variations. We can give frame some room around the edges by using \fbox instead. Figure 11: Some Font Sizes Result (Source in Figure 10) The use of braces to enclose a font size speciﬁcation is like an environment. These size variations can be combined with font styles.) To overline is as straightforward.2.
(Just as in the regular text. this ability is occasionally useful in computer algorithms. or a sequence of one line quotes. such as page numbers. Turing “There are reports that many executives make their decisions by ﬂipping a coin or by throwing darts.” — Alan M. two are for ranges. etc. The convention is that one dash is for hyphenation. for example in situations where a ﬁxed decision made each time would cause the algorithm to run more slowly.10
2 TEXT The construction of the real number system. Sometimes it is important to make a completely ‘unbiased’ decision.) Here is an example that was created by putting \begin{quotation} before the text and \end{quotation} after it. after their quote. I used ‘‘ ’’. We are far from certain what dreams are good for. “Computers do not dream. this can be overridden by the \noindent command. which ran into trouble due to insuﬃcient rigor in dealing with limits. by skipping a line and entering \hfill (which means horizontal ﬁll). . separated by blank lines. the longer the dash. but we know what they indicate: a great deal of information processing goes on far beneath the surface of man’s purposive behavior. Turing and Donald E. to create this more stylistic quotation punctuation. not " ". use preceding “i. Alan M. Here are some other things to notice about this example: • There are left and right quotation marks. having more than one paragraph (separated by blank lines).”
. The quotation environment is used for long quotations. Knuth. Knuth The quotes are by two pioneers of algorithms. The more minus signs you use. generally one short paragraph (as above)..
The quote environment is intended for short quotes. and three are for punctuation — i. was motivated by the need to ﬁx calculus.e.” — Donald E. any more than they play. It is also rumored that some college professors prepare their grades on such a basis. in ways and for reasons that are only very indirectly reﬂected in his overt activity. to make the line ﬂush right. • The dash that appears before each name is created by three minus signs. except the ﬁrst line of each new paragraph is indented. Their names appear on the right. notably by Dedekind cuts. The indentation is the same as the quote.e.
since he who is ignorant of it cannot know the other sciences or the things of this world.
. The space you now see just above is a small skip. And what is worse. medium and big. each line is a stanza in the verse. This was done with the \bigskip command.2. This is just an ordinary line space.1 Fonts and Paragraphs
11
• There is extra space between the two quotations. men who are thus ignorant are unable to perceive their own ignorance and so do not seek a remedy. The following was generated by putting \begin{verse} before the text and \end{verse} after it: Neglect of mathematics works injury to all knowledge. \bigskip The space you see above is a big skip. by enclosing Bacon’s verse with \textit{. — Roger Bacon The italics were speciﬁed in the usual way.
The space you now see just above is a small skip. The space you see above is a big skip. . . The space you see just above is a medium skip. Figures 12 and 13 illustrate three levels of skipping: small. \medskip \smallskip
The space you see just above is a medium skip. Figure 13: Skipping Line Spaces Result (Source in Figure 12) The verse environment indents oppositely: lines after the ﬁrst. This is just an ordinary line space.
Figure 12: Skipping Line Spaces Source (Result in Figure 13)
This is a ﬁrst line. (Designed for poetry. we shall look at a way to have a much ﬁner range of vertical spacing.
This is a first line. }. Later.
but this time it is on the last line of the verse. like a comma.
A Making Tables. make paragraphs. (The reason is that a space (or some delimiter) is needed after \LaTeX (or any keyword) in order to distinguish it completely. Knowing how to create a bibliography.Roger Bacon \end{verse}
2. Other. This is a power of L TEX and one reason why it has become standard in writing mathematical papers. L TEX provides a means to make tables with the tabular environment.12
2 TEXT
and if a stanza runs long.
A LT
Two new things appear in the example: the use of \LaTeX to produce Without the EX. report or article. This is achieved by not skipping a line after the verse: \begin{verse} \textit{Neglect of mathematics . I will show you how to do virtually any mathematical expression in line with the text. distinguished by what appears at the beginning of each item: number.
Bibliography. or in math display mode. } \hfill . There are a great many things to learn beyond the simple introducA tion when using L TEX to prepare a thesis. or your description (perhaps nothing). which requires no space.
A Mathematics. bullet. To illustrate. following \LaTeX. and its versatility puts it far ahead of word processors. vary fonts.)
.
Graphics. the result would be L TEXprovides. here is the use of a description list environment A to itemize steps involved in learning L TEX. rather than a new line. and the use of ~ (called “tilde”) to enter a space. A tilde.2
Lists
There are three intrinsic list environments. Basic Document Preparation. in particular with BibTE X. Knowing how to setup the latex source ﬁle.. again from the \hfill command. this form of indentation makes it clear.) Bacon’s name appears ﬂush right. This has progressed a great deal in the past few years thanks to many people who have provided packages free of charge. and one might want a punctuation mark. whose source is indicated by Figure 14. and list items are enough to prepare a basic document without mathematics or tables (like a resume).. even with a space after \LaTeX in the source.
– A second (nested) itemized list changes the bullet and indents another level. as in this example.. You see the nesting of two itemize lists....] This has progressed a great deal in the .2 Lists
\begin{description} \item [Basic Document Preparation..2. \end{description}
13
Figure 14: Description List Environment The text within the square brackets is an option. and the lines extend all the way to the right margin. \item [Other..] There are a great many things to learn .. Figures 15 and 16 illustrate the itemize list environment.] \LaTeX~ provides a means .. If present. using the indentation of the itemize environment. Unlike the verse environment..
• This is item 1 and our task has just begun. \item [Making Tables.. \item [Graphics. which prints bullets. • This is item 2 and we shall limit to just this few. it is printed in boldface. \item [Mathematics.. the ﬁrst line goes almost to the left margin. Figure 16: Itemize List Environment Result (Source in Figure 15)
. 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.] Knowing how to create a bibliography .. but any type of list can be nested within any other type. With no option. Next. \item [Bibliography.] Knowing how to setup . For example.} \end{description}
produces the following result: This is how one item in a description list environment looks with no optional text at the beginning. Blank lines before an item have no eﬀect. Note the indentation of each item and the spacing between items.
\begin{description} \item \textsf{This is how one item in a description list environment looks with no optional text at the beginning.] This is the power of \LaTeX~ and one . A blank line within an item does create a new paragraph.
and it’s time to number anew. \begin{itemize} \item A second (nested) itemized list changes the bullet and indents another level. With nested enumeration. \item This is item 2.
2 TEXT
Blank lines
\item This is item 2 and we shall limit to just this few. A blank line within an item does create a new paragraph. \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)
. using the indentation of the itemize environment. the numbering changes at each level. before an item have no effect. but we are not yet done. \begin{enumerate} \item Back to item 1. and we are having fun.14
\begin{itemize} \item This is item 1 and our task has just begun. \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.
\begin{enumerate} \item This is item 1. \item Two is new. Figures 17 and 18 illustrate with three levels of nesting.
One again! ii. respectively. Figure 20 illustrates a combined use of these options. To draw a line after the last row. This is item 1. Figure 19 shows an example of a 2 × 3 table. i. .2. (a) Back to item 1. enter \\ \hline (the \\ is simply part of the syntax and does not add an extra row to the table). last row spec [\\ options] \end{tabular}
As indicated. syntax: \begin{tabular}{column specs} options ﬁrst row spec \\ . c or r. abbreviated by just one character: l. Two (b) or knot 2b? Figure 18: Enumerate List Environment Result (Source in Figure 17)
2. The column speciﬁcations can have  on either side to indicate a vertical line.3 Making Tables
15
1. In the body of the table. Each column speciﬁcation can be left. and it’s time to number anew. \\. each column is separated by &. We could draw lines that span some rows and/or columns. 2. (b) Two is new. but we are not yet done. 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.3
Making Tables
which has the following
A table is made with the tabular environment. The way to vary vertical line drawing is with the column speciﬁcations: put  only where you want a vertical line. each row ends with two backslashes. This is item 2. The way to vary horizontal line drawing is by using
. center or right. and we are having fun. .
16 How it appears 110 210 120.12 220. 130 230 What you write
2 TEXT
\begin{tabular}{lcr} \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}{lcc} Name & Test 1 & Test 2 \\ \cline{11} Bob & 67 & 72 \\ Sue & 72 & 67 \\ \cline{23} \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{33} 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{33} \\ \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{22} & 3 & 4 \\ \cline{11}\cline{33} \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}{ll} \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}{ll} \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}{ll} \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}{ll} \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.
This is not the same as How sweet it is. we want the columns to be left justiﬁed. which I have put in sans serif font. on either side. such as the width of a paragraph box. a vertical line speciﬁcation. For example. The \multicolumn command can also serve to override some column speciﬁcation. such as l.}} \end{flushleft} } & \parbox[t]{1in}{\begin{flushleft}\textsf{This is column 2.2. Suppose. the third argument is the text. \multicolumn{number }{col spec}{entry}.3\linewidth}. and I might want to display something: How sweet it is. for example. If we want some heading to span several columns. then the
. but we want the headers to be centered. This must be in the range 1 to however many columns remain from the current position.}}\medskip This is not the same as \medskip\fbox{\centerline{How sweet it is. see exercise 9 at the end of this chapter and consider \parbox{. with. can be determined by some length parameter. this is done by the command.} \end{flushleft} } \end{tabular} \end{center}
19
Figure 24: \parbox Source (Result in Figure 25)
This is column 1. which I have put in sans serif font. r. The line in the source begins with & to skip column 1. rather than a ﬁxed constant. The ﬁrst is used to center ‘Test number’ over columns 2 and 3. or without. . c. Figures 26 and 27 illustrate these uses of \multicolumn..
Figure 25: \parbox Result (Source in Figure 24) Any measurement.3 Making Tables
\begin{center} \begin{tabular}{ll} \parbox[t]{3in}{\begin{flushleft} This is column 1. The second argument is any valid column speciﬁcation.. and I might want to display something: \medskip\centerline{\fbox{How sweet it is.
This is column 2. Finally. The ﬁrst argument is the number of columns to span (starting where \multicolumn is speciﬁed).
Of particular importance are the braces. . written as \{ \} to obtain { }. Table 2 (next page) illustrates this with commonly
. The second use simply centers the ‘Student’ header. enabling comments. Then. which has most of the same options. in that they delimit A something. The last use of \multicolumn centers ‘Taken in class’ over columns 2 and 3. \ delimits every L TEX command. centered with vertical lines before and after. specify the longtable environment. like {\large.4
Special Characters
We have already seen that some characters are special. and % ends a line. the vertical line at the end is missing because c was speciﬁed instead of c.5 70. The Appendix contains complete tables of these special characters (including those I do not cover explicitly in the chapters). How do we print such characters? One way is with the symbol. it appears with whatever font is active. is needed. (You obtain the package from CTAN [4].5 \\ Charleetah & 72 & 67 & 70. like \%. Unlike the ﬁrst use.5
Figure 27: Multicolumn Result (Source in Figure 26) Tables that are too long to ﬁt on one page could be broken manually. A but the longtable package enables automatic page breaks by the L TEX compiler. like \textbackslash.) In the preamble.20
2 TEXT
\multicolumn speciﬁes 2 columns. instead of the tabular environment. specify \usepackage{longtable}. }.
2. itself. Other times a keyword.) When using a keyword to specify a special character. In particular.
\begin{center} \begin{tabular}{lccc} & \multicolumn{2}{c}{Test number} \\ \multicolumn{1}{c}{Student} & 1 & 2 & Average \\ \hline Bill & 67 & 72 & 70.5 \\ \hline & \multicolumn{2}{c}{Taken in class} \\ \cline{23} \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. (Recall that the braces by themselves create a local environment. .
4 Special Characters
21
used special characters. . For example.2. except when they are used to delimit an option in the syntax. Unlike all other commands. The verbatim environment uses the usual syntax: \begin{verbatim} . \end{verbatim}
. \verb does not use braces to delimit its argument. we can write \verb@{}%$#\@ to generate the string {}%$#\ (delimited by @). are diﬀerent because they can be entered directly. One example is the description list environment. but I omitted a discussion of position options that are speciﬁed by brackets. \end{description} \begin{description} \item {[This is not option]} for item. illustrated in Figure 28. in which case they can be obtained by enclosing them in braces. which can contain any character except itself.
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. (Another is in the tabular environment (page 15). What you write
\begin{description} \item [This is option] for item. It uses any other character to delimit a string.) How it appears This is option for item. which is printed in typewriter font style. \end{description}
[This is not option] for item. . The brackets. [ ].
In addition. \= to deﬁne a tab setting. In some cases. without extra spaces. ?‘⇒¿ and \aa⇒å. and \> to move to a tab setting. yet it is needed to deﬁne the tab. Table 4: The Tabbing Environment What you see Begin: set tab 1. Table 3 shows some common examples.5
Tabbing
The tabbing environment provides an alternative to the tabular environment by letting you set your own column tabs. For example. particularly if the entire document is to be in a nonEnglish language. a complete table is in the Appendix. such as Babel [1] (also see [5. Another class of special characters are letters with accents. even if it does not relate to some ¨cˆ language — for example. \\ ends each row. An accent could be applied to any letter. so that the position of the tab is not equivalent to that of a table’s column. write Poincar\’{e} to have Poincaré and G\"{o}del to have Gödel. like Figure 26 (p. .g. A L TEX has a basic library of accents and special characters for writing in languages other than English. For that purpose there are some packages. \"{b}\~{c}\^{d} ⇒ b˜d. 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. some of which are shown in Appendix Table 29 — e.. the ﬁrst sentence continues normally. . but unlike the tabular environment.22
2 TEXT
This is how the source code was created for the ﬁgures. Table 4 shows a simple case with two basic tabbing commands. Table 5 illustrates how this is solved with the
. 20). Chapter 9]). Table 3: Some Accents for Letters What you write \"{a} \‘{e} \’{i} \~{o} \^{u} ⇒ ⇒ ⇒ ⇒ ⇒ What you see ä è í õ û
2. these are not suﬃcient.
Table 5: The \kill Tabbing Command What you see 13 sting like a bee 468on’t be late d What you write
\begin{tabbing} 13 \= sting like a bee \\ 468 \> don’t be late \\ \end{tabbing} \begin{tabbing} 468 \= don’t be late \kill 13 \> sting like a bee \\ 468 \> don’t be late \\ \end{tabbing}
13 sting like a bee 468 don’t be late
Figures 29 and 30 illustrate the tabbing environment with ﬁxed ﬁeld widths. like the following:
.
\begin{tabbing} \= \hspace*{. then speciﬁes \kill instead of \\ to suppress (or “kill”) the printing of the line. the lines are in the order we want. in order to set the tab correctly. and Word Breaks
23
\kill command. but the tab is set by the shorter string ‘13’. then it uses the name of the last ﬁeld to set what follows. It ﬁrst uses the \hspace* command for horizontal spacing. The second tabbing puts the longer ﬁeld ﬁrst. this could result in an undesirable appearance. making ‘468’ extend past the tab. When text is justiﬁed (the default). Page.6 Line.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. Page. and Word Breaks
You can cause a new line by entering \linebreak. In the ﬁrst tabbing.2.6
Line.
enter:
\nohyphens{Word breaks are hyphenations that \LaTeX\ does for you. use the space character ~. to suppress hyphenation. such as ‘ﬁgure 1’ or ‘p. 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. rather than completing the line as \pagebreak does. you specify the \nohyphens command. The \newline command forces a new line without justifying it. We would thus write figure~1 or p. For example.} ⇒This example extreme.24 \textsf{This example is \linebreak extreme. you obtain the following:
A Word breaks are hyphenations that L TEX does for you.
2 TEXT
is
The \nolinebreak command works analogously.) Then. you want to suppress hyphenation. \textsf{Here is the extreme \newline example. (You might have to download the package from CTAN [4]. Sometimes.~10. however. and follow the simple installation instructions. To prevent a line break where you want a blank.}
Then. to suppress it. even if it means extending into the right margin.
{\samepage line 1 \nopagebreak line 2 }
A Word breaks are hyphenations that L TEX does for you. The \newpage command follows the analogy with \newline in forcing a page break precisely at the point it is speciﬁed. however. The \nopagebreak command disallows a page break immediately following the next blank line. you want to suppress hyphenation. you want to suppress hyphenation. the ﬁrst sentence of this paragraph has a hyphen. Sometimes. preventing a line break.} ⇒Here is the extreme example. This can be done by specifying \usepackage{hyphenat} in the preamble. There are two commands to force a page break: \pagebreak and \newpage. however. Sometimes. Also. 10’.
. it is better style to keep certain ‘words’ together.
2. \hrulefill Analogously. . The height of one line of normal text is in the keyword \baselineskip. . . . . .3in} here. The \hspace command has no eﬀect at a line boundary. . I insert . You could write \renewcommand{\baselinestretch}{1. . This acts like renewing the \arraystretch setting and remains in eﬀect until changed. . but the \verb\hspace* \hspace*{1in} inserts the space no matter what. You can. . . The easiest way to control line spacing throughout your document is to specify usepackage{setspace} in your preamble. . . . . . Two variations of \hfill are: \dotfill . . . but the \hspace* inserts the space no matter what. . . . . . you use \vspace*{2in} to put a 2 inch margin at the top (\vspace would not insert the space). .2}\small\normalsize the associated postscript result (ps ﬁle). . . . The vertical space is not added if this goes to the top of a new page. . . .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 most versatile method to insert horizontal spaces is with \hspace and \hspace*. . . Be sure your name is on each. so vspace{\baselineskip} skips one line at the next new line.3 in here.
That is why you see the 1 inch space at the beginning of the second line.
A Exercises. . the previous sentence is written as:
The \verb\hspace command has no effect at a line boundary. For example.2} to increase the spacing modestly (actual increase depends upon the font size). Submit a printed copy of both the L TEX source (tex ﬁle) and
1. . . . . . . . . . at the very beginning of your document. . However. These have one argument: the amount of space to be inserted. . . . . . . \vspace* and \vfill. In particular. This gives you three commands: \singlespacing \onehalfspacing \doublespacing Right after you specify one of these. Write a paragraph in article [and letter] style with the following properties:
. . . . one diﬀerence is that you need to change font size before this change goes into eﬀect. that spacing will commence. . . For example. .3~in \hspace{. . . if you want to make your own title page. however. but \hspace* does.7 Spacing
25
2. \hspace would not insert the 1 inch. . specify \renewcommand{\baselinestretch}{1. . that is what \vspace* does. ⇒ I insert .
(c) Each font style in Table 1 is used for two consecutive complete words. (c) Both paragraphs are indented. • Your name appears second in letters that are not as large as the title. Produce the following table (including the accents and alignments). Write two paragraphs in article [and letter] style with each of the following properties: (a) Default indentation on both paragraphs. • Your web address appears fourth. Give the same list without numbers. 4. • Your email address appears third. with extra space preceding it. (b) No paragraph is indented. with extra space preceding it. 5. • The title appears ﬁrst in very large letters. 2.26
2 TEXT (a) Each font style in Table 1 is used as one letter in a word that has more than one letter. Give an enumeration of at least three things you like about mathematics. but larger than normal size. Produce the following table: Colors Primary Secondary Red Green Blue Orange Yellow Purple 6. • Course number and title appears next. • Date appears last.
. and it is preceded by extra space. (b) Each font style in Table 1 is used for one complete word. (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.
Use the tabbing environment to produce what you see on page 30. 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. Create a 3column text such that each column is a paragraph of arbitrary length using about 1 of the page width each.2. Produce the following table.
. 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.7 Spacing Mathematician Gabrielle Emilie Le Tonnelier de Breteuil Marquise du Châtelet Benjamin Banneker Sophie Germain Julius König Rózsa Péter 7. 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. 12. 3 10.
instead of with §7. The bibtex program that you apply to your source creates another ﬁle (which you need not examine). 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.2
3. So here we are. and I want them to use BibTE X.1
The bib File
Main body
For purpose of this introduction. The bib ﬁle contains the bibliographic database.tex
view/print convert dvips myfile.pdf
Figure 31: Adding bibtex to the Command Sequence
3.2.ps myfile. but that name is arbitrary as long as it ends with . suppose your bibliography is in a ﬁle called mybiblio. This added loop is illustrated in Figure 31.bib. about which you need not be concerned). which could extend beyond one document. Once this is successful.28
3 BIBLIOGRAPHY WITH BIBTE X
3
3. We
.1
Bibliography with BibTE X
Overview
It might seem strange to have this section so early. BibTE X [11] was developed by Oren Ptashnik and is available free of charge. 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. That is because I require students to produce an annotated bibliography early in the semester. until you do not have any “unresolved” bibliography citations. It reads a plain text ﬁle.4.dvi dvipdf bibtex
print/post
myfile.bib. you do not have to bibtex myfile again until you change your bib ﬁle or add a citation.
create/edit compile latex myfile.
year = "1989". you can put both of their names. some people add another keyword. so tex is not the same as TeX. address = "Reading.3. Knuth’s book [8] would be entered as follows: @article{tex. The use of one keyword is somewhat simplistic and could cause problems with a great number of entries because the labels must be unique. some people add a.2 The bib File
29
begin with the most important part of your bib ﬁle. ﬁeld = "value".year author :ﬁrst_keyword_in_title Example knuth. with more than 2. We cannot. Also. you can add et al. after the year (no blank). the label is speciﬁed as tex and it must be followed by a comma. the use of the Latin et al.
• Each bib entry must have a unique label.89 knuth:tex
With two authors. which are the entries you want to include in its database. } For example. Here are two styles I have seen. edition = "15th". MA". In the example.
. so it can be cited without ambiguity in the source ﬁle. publisher = "AddisonWesley Publishing Company". . . author = "Donald E.) In the ﬁrst form. Each entry has the following form: @type {label . . if there are two publications by the same authors in the same year. . . . In the second form. labels are casesensitive. for example. Before listing each style (article is one style) and the ﬁelds they can or must have (author is one ﬁeld). which you might consider: Form author .
A • The label is arbitrary. in formal writing follows this rule. here are a few things to note. but do not use any L TEX special characters or blanks. Knuth". b. if there are two publications by the same authors in the same year. title = "The {\TeX} Book". have two entries with tex as their label. (Linguistically. } Most authors develop a style to labeling bibliographic entries. You must discover what style works best for you.
month. Required ﬁelds: author or editor. year. Required ﬁelds: author.) Fields that are neither required nor optional are ignored. chapter and/or pages. title.
• There is a ﬁnal } to close the entry — @type{ . series. ﬁrst. year. type. address. title. incollection is a part of a book having its own title. Here is a list of standard entry types with their required ﬁelds. address. printed document. Required ﬁelds: author. even if they are valid ﬁelds in other types of entries. title. booklet refers to a bound. howpublished. Optional ﬁelds: volume or number. address. series. If you put Knuth Donald. but it is more readable that way. book refers to a book with an explicit publisher. For example. but it will not hurt anything. edition. year. but without an explicit publisher. edition. Required ﬁelds: author or key.
A • The ﬁeld value can be anything recognized by L TEX. publisher. . edition. such as a chapter or just some range of pages. Optional ﬁelds: editor. but most journals require the volume of the journal and the page numbers for the cited article.30
3 BIBLIOGRAPHY WITH BIBTE X • The order of the ﬁelds is arbitrary. The last ﬁeld does not require a comma at the end. and ﬁelds are separated by commas (hence the comma after the terminal quote). the volume and page numbers of an article are necessary to include even though they are optional to satisfy BibTE X syntax. month. pages. pages. • Fields do not have to be on separate lines. number. and it gives ﬂexibility if you want to add a ﬁeld or change the order. article refers to an article from a journal or magazine. month. publisher. journal. Optional ﬁelds: volume or number. Remember to put each author’s name as ﬁrst last or last. year. Required ﬁelds: author or editor. title. address. chapter series. volume or number. Optional ﬁelds: volume.
. (What is “necessary” depends upon the standard one applies. type. inbook is a part of a book. year. What are “optional ﬁelds” in BibTE X are not necessarily optional as far as having a complete bibliography citation. . Optional ﬁelds: author. even mathematical symbols in math mode. the compiler will think the ﬁrst name is Knuth and the last name is Donald. publisher. booktitle. month. month. }. title.
Required ﬁelds: author. pages. publisher. which has no person as an author. year. howpublished. proceedings Required ﬁelds: title.2 The bib File inproceedings is an article in a conference proceedings. year. unpublished is a document with an author and title. address. (Some note of explanation is required. organization. year. you must provide a key for sorting. year. month. address. misc is when nothing else ﬁts. The bibliography will be sorted with the key. (this is ignored with no error message given). Required ﬁelds: author. Optional ﬁelds: editor. enter an unrecognized ﬁeld. organization. series. month. month. used to order this entry relative to author
. Optional ﬁelds: month. title. school. year. month. title. title. note. year. year. volume or number. To have a comment that is not printed. Optional ﬁelds: type. title.
31
In addition to the optional ﬁelds listed. address. institution. such as comment = ".". address. Required ﬁelds: author. Required ﬁelds: author or key (see note below). month. title. For example. you will get an error message. Possibly there will be some standard ﬁxup. volume or number. Optional ﬁelds: author.. mastersthesis is a Master’s thesis. Optional ﬁelds: editor. edition. school. techreport is a report published by some institution. Required ﬁelds: author or key (see note below). consider the following entry for A L TEX 2ε [10]. but not formally published. address. the note ﬁeld is always an option. title. address. This lets you enter a note that will appear at the end of the citation. phdthesis is a PhD thesis. If a required ﬁeld is missing when you compile. booktitle. publisher. series. Optional ﬁelds: author. but it is best if you provide the missing ﬁeld value. year. LaTeX2e. month. even as a technical report. month. organization. title.. which vary by the type of entry. Optional ﬁelds: type. If a document has no author. Required ﬁelds: author. number. manual is some technical documentation.) Required ﬁelds: author. Optional ﬁelds: type.3.
MA". use this feature inappropriately by putting braces around everything. year = "1994". address = "{CTAN\url{/macros/latex/doc/usrguide. key = "LaTeX2e". or simply G{\"o}del. Otherwise. }
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). but here is one way. however. 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. There is no universally accepted standard for how to reference web documents.
3 BIBLIOGRAPHY WITH BIBTE X
@manual{usrguide. use the book type and specify:
. Some authors. author = "Michel Goosens and Frank Mittelbach and Alexander Samarin".2. we separate them with and (no commas). the bibtex program will produce ‘fourier analysis’ (the plain style produces article titles in lower case. such as writing {F}ourier analysis to force the capital F. the bibtex program might try to process it itself and produce an A unintended result. If it is a book. 3. title = "The {\LaTeX} Companion". except the ﬁrst letter of the ﬁrst word). institution = "Comprehensive {\TeX} Archive". title = "{\LaTeXe} for authors". publisher = "AddisonWesley Publishing Company". otherwise. type = "World Wide Web site". }
When there are multiple authors. For example. year = "199599". address = "Reading. In ordinary L TEX. The use of braces to force a particular result is necessary in other instances. The key will not be printed. the World Wide Web did not exist. Now it is a major source of information. [5] in this document has the following BibTE X entry:
@Book{companion. but this will not work in BibTE X.32 names.2 Web citations
When BibTE X was developed. This applies to accents too. so we can switch styles and use the same source (tex and bib ﬁles). we write G\"{o}del to produce Gödel.ps} (see~\cite{CTAN} for replacing CTAN)}". Instead. we write G{\"{o}}del.
~ is in many urls.’ is the ﬁrst name of the author. Carlisle and Alan Jeffrey and Frank Mittelbach and Chris Rowley and Rainer Sch{\" o}pf". address = "url ". it is deﬁned in a package. put the following declaration into your preamble: \usepackage{url}. address = "http://www. respectively. title = "{\LaTeXe} and the {LaTeX}3 Project". title = "Elements of Style".columbia. There are occasions when we want to reference an entire web site. These diﬃculties are overcome by specifying:
address = "\url{http://www. Jr. Jr.
33
Here is an example:
@book{Strunk. Eventually. }
The reference \cite{StrunkWhite} presumes there is the entry for the original publication.columbia. but you can change this to TimesRoman or sans serif font with the speciﬁcation: \urlstyle{rm} or \urlstyle{sf}. the url could contain special characters. if needed. For one thing.3. year = "1999". One A example is the L TEX 2ε reference [2]. Another feature of the url package is that \url prints special characters.
A The \url speciﬁcation is not actually an intrinsic command in L TEX. you will run into some diﬃculty with writing urls. in particular.2 The bib File
publisher = "World Wide Web". The use of the braces in the name is to be sure that the author appears as intended: William Strunk. followed by the url. note = "This is the web version of the classic book by Strunk and White~\cite{StrunkWhite}". Otherwise.edu/acis/bartleby/strunk/". Its main use is to determine where the url can be broken in order to put it on two lines. the comma would signal the bibtex program that ‘Jr. and with latex having no place to break. Also. An unsightly line with spaces could also appear after the url. use that style but include the url as a note or in the address ﬁeld. To have the \url command active in your document. not the tilde. you will see a line with lots of spaces (for justiﬁcation).
. a url could become very long. given by:
@misc{latex2e. author = "William Strunk{. author = "Johannes L. Braams and David P. like ~.edu/acis/bartleby/strunk/}".}". William Strunk’. publisher = "World Wide Web". If the document is a technical report. without the braces. The default font it uses is tt. and writing it will produce a space. and it would appear as ‘Jr.
” other times “Kluwer Academic Publishers. (That is why we needed the quotes before.html}".
. title = "My MotherMy Father".
Then.3 Additional features
One element of good style is to be consistent in your terms.) For example. For example.
}
We have seen several packages so far.latexproject/org/latex3. editor = "My Father". 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. including abbreviations and names of publishers. this is the ﬁrst use of \renewcommand. 3. we merely change the one string value and recompile. the three ﬁeld values are equivalent to:
author = "My Mother".” and still other times “Kluwer Pub. and you shall learn more about A packages in §6. title = mom # dad. One sometimes sees “Kluwer. \url{http://www. year = "1994".
to produce the publisher value = “Kluwer Academic Publishers. when we wrote literals. suppose we write
@string( mom = "My Mother" ) @string( dad = "My Father" ) author = mom. where I describe enhancements for having graphics in L TEX. about which I shall say more when I describe ways to customize your document in §??. we can enter:
publisher = kluwer. editor = dad. However.” To help be consistent and to save some work in the long run when we write many diﬀerent documents and produce more bib ﬁles.” Besides consistency. suppose we deﬁne:
@string{kluwer = "Kluwer Academic Publishers"}
Then. We can concatenate strings and/or literals with #.2. an advantage is that if some name changes.34
3 BIBLIOGRAPHY WITH BIBTE X
howpublished = "World Wide Web.
S. Kluwer Academic Press. Artémiadis. editor = "J.2 The bib File
35
Note the absence of a space between the string values in the title. }
If these were the only references. Byrnes and Jennifer L. editors. For example.
Another useful feature of BibTE X is the crossref ﬁeld for cross referencing. Results on the absolutely convergent series of functions and of distributions. Byrnes. the other values are literals):
@Proceedings{Byrnes:FAA89.K. author = "N. the result would appear as follows:
[1] N. = mom # " My Father". Here is an example [11] that is useful for guiding the sorting of references in a special circumstance:
. crossref = "Byrnes:FAA89".
[2]
BibTE X also recognizes a preamble in our bib ﬁles to enable us to deﬁne A some L TEX commands. Byrnes". Byrnes and Jennifer L. In Byrnes and Byrnes [2]. Recent Advances in Fourier Analysis and its Applications: Proceedings of the NATO Advanced Study Institute. title = "Recent Advances in {F}ourier Analysis and its Applications: Proceedings of the {NATO} {A}dvanced {S}tudy {I}nstitute". 1990.3.S. }
Then. as a literal:
title = mom # "~" # dad. J.
The same title as the above is obtained by any of the following:
title title = "My Mother " # dad. publisher = kluwer. title = "Results on the Absolutely Convergent Series of Functions and of Distributions". pages 311–316.K. The general form is
@Preamble{ string }
where string is any concatenation of literals and strings. use the space character. year = 1990. To ensure a space. we can have the following entry:
@InProceedings{Artemiadis:FAA89311. Art{\’e}miadis". pages = "311316". ~. suppose we have the following entry (kluwer is a string.
mybiblio. then by year.36
3 BIBLIOGRAPHY WITH BIBTE X
@Preamble{ "\newcommand{\noopsort}[1]{}" }
The \newcommand is something I shall describe more fully in §??. abbrv diﬀers from plain by abbreviating names of journals. we fool the bibtex program with the following speciﬁcations: Volume 1
year = "{\noopsort{a}}1973". including some provided by publishers.
. rather than by the author names. To force the ﬁrst volume to sort before the second. originally published 1971. which comes with every installation of latex. among other things (to give a more compact bibliography). For now. breaking ties with the year of publication. requiring one argument. Suppose there is a 2volume work by the same authors. so just the years appear. alpha diﬀers from plain by citing by labels. unsrt diﬀers from plain by sorting entries by the order in which they are cited.bib. producing nothing (indicated by {}). which are the same. The second command deﬁnes the format style of the bibliography to be plain. 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.
Volume 2
year = "{\noopsort{b}}1971". Here is how this can be used. There are other bibliography format styles.
This fools the bibtex program into thinking the years are a1973 and b1971. but a second edition of volume 1 is printed in 1973. does not actually print the letters. Entries are sorted by the alphabetical order of author names. thus putting volume 1 ﬁrst.3
Declaration and Citation
At the end of your source ﬁle (where you want the bibliography to appear).
3. rather than numbers. however. The deﬁnition of \noopsort. and they are labeled with numbers. before \end{document}. \noopsort. it is used to deﬁne a command. put the following commands (in either order): \bibliography{mybiblio} \bibliographystyle{plain} The ﬁrst declares the bibliography to be in the bib ﬁle. Command \noopsort ignores the argument that it receives. The bib entries would have the years in the opposite order than we want because sorting is ﬁrst by the authors.
This is done with the \nocite command. otherwise. even if the main document is spaced diﬀerently.latex} produces [8. 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. separated by commas. 46] in this document. You can insert some further citation information as an optional input argument to the \cite command. even if it is not cited explicitly. Just before the \bibliography statement.) The rule is that only those bib entries that are cited appear in the ﬁnal document. we use \nocite with their labels. 9] for this document.4 Some Controls
37
We shall use only the plain style here. A To cite particular references. delimited by [ ]. where label is what we put in our bib ﬁle entry. the ~ is used to ensure that there is a space but no line break when giving the page number as “p. For example. the L TEX command is \cite{label [. put \renewcommand{\baselinestretch}{1} \selectfont
.~46]{latex} produces [9. ]}. For example. they will appear twice if both labels are used (or if we used \nocite{*}). we will get a fatal error message.another} The bibtex program will search them sequentially for any citation.bib appear. 46” in the citation. For example. \cite[p. Repeated entry– telling us which label is repeated. If we have the same entry with diﬀerent labels. (In the option.3. There are times when we want to be sure a particular bib entry appears. . but know that many other styles have been written and are available free of charge. [8] is produced by specifying \cite{tex}.4
Some Controls
You might want to have the bibliography singlespaced. If we have the same label in both bib ﬁles. such as:
\bibliography{mybiblio. if we want to have every entry in our bib ﬁle appear. You can put more than one citation. \cite{tex. . In particular. such as \nocite{tex} to be sure Knuth’s TEX book appears. If we want only some particular list of entries to appear.. the entries must be identical. p. we specify \nocite{*}. Figure 32 shows a complete source ﬁle for having all entries in mybiblio.
3. The reason is that we can maintain one large bib ﬁle and write many documents that use it. but we do not want to cite it in the text.
such as the heading. Rich.R. Rich. How to Square a Circle.
1. Money.38
3 BIBLIOGRAPHY WITH BIBTE X
The \selectfont command is needed to activate the change in \baselinestretch.M. MacTaco. chapter 1. volume I. Submit a printed copy of the L TEX source (tex ﬁle). I. 1990. and V.F. Further. \selectfont is not needed.M. You can change the heading put by bibtex. 1999. 4. (d) A technical report.M. editor. (b) A book on the web. Smart. Tu. U. Produce a document that has only a bibliography composed of the following three entries (in the order shown). (c) An entire web site.
A Exercises. the BibTE X data (bib ﬁle). Impossible Dreams. Thus. (b) An entire book with at least three authors. [2] [3] I. but you might want to have it be References. Volume II of Rich [2]. you could specify \renewcommand{\baselinestretch}{1} \small to have the bibliography set in small font size (and single spaced). which is available at CTAN [4]. editor. that will do. 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. Produce a document with one paragraph that cites three bibliographic items. Impossible Dreams. MacTaco. and the associated postscript result (ps ﬁle). and you can change the properties. one for each of the following types: (a) An article in a journal. Produce a document that lists your entire database. Be sure your name is on each. 1990. one for each of the following types: (a) A technical report on the web. [1] I. 2. volume II. The default is Bibliography. at least one entry must have more than two authors.
. which consists of at least one entry for each of six diﬀerent document types. 3. This will let you put diﬀerent bibliographies throughout your document. second edition. Produce a document with one paragraph that cites three bibliographic items. (c) A chapter in a book. If you also want to change the font size (or style).
1960.
4. if you want to use the counter’s numerical value as an argument in a command. respectively.e. Then. The ﬁrst 100 pages seem simple.. just as the labels in the bib ﬁle entries. such as .1
Counters. \thepage produces the page number. placed where the counter’s value is set. Counters that depend upon the style.R. and equations. There are times when you just want to produce the counter value. which I assign to my Ph. The L TEX syntax for labeling a counter is \label{label }. Princeton University Press. The choice of label is arA bitrary. can be labelled and referenced in the same manner. 1945. such as pages. A where label is unique in the document. The next 100 reveal lack of understanding the ﬁrst 100. like a chapter in a book. students. Naive Set Theory. This is a good book. which I was able to print by writing \thepage. Princeton. without a label. the following labels were deﬁned (when the subsection and subsubsection were ﬁrst written):
. ﬁgures. and there are modern descendants. In the next section I describe intrinsic counters and illustrate how to label and reference them. This is done by \thecounter. specify \value{counter }. you are reading subsection 4. and a reference is a citation to a label. in this book I deﬁned: \section{Bibliography with \Bibtex} \label{sec:Bibliography} Now I can refer to §3 by \S\ref{sec:Bibliography}. To illustrate how I can reference other parts of this document.D. whose numbers I could write by \thesubsection\ and \thesection. except do not use L TEX special characters or blanks. Halmos. NJ. Van Nostrand. A label is the identiﬁcation of A a particular value. [2] G.
4
4. I shall introduce the ﬁgure and table environments.2 of section 4. Produce an annotated bibliography of the following form (note the indentations on left and right margins): [1] P. You are looking at page 39.39
5. This is a seminal book that articulates the problemsolving — i. The L TEX syntax for referencing a label is \ref{label }. NJ. Princeton. Polya. . Similarly. For example. . There are many editions. On the other hand. theoremproving — process. How To Solve It. Labels. sections. which have intrinsic counters associated with them. For example. and References
Basic Concepts
A counter is a numerical value that refers to something that is being numbered. Here I illustrate some of those that are in all document styles.2
Intrinsic Counters
A Anything to which L TEX assigns a number has a counter associated with it.
. Some people use this same form but with diﬀerent preﬁxes. . .) Equation (6).~\pageref{subsubsec:webcite} ⇒ p. if present. For that reason they are called “ﬂoating objects. can go at the top or bottom. LABELS.)
4.” or ﬂoats. such as ss for subsection and e for equation. .~\pageref{subsec:bibfile} ⇒ p. where you put it is where it will appear. even though you will get no error message. That is a matter of style. I entered the label \label{exer:likeaboutmath} (page 26). This helps me to distinguish labels for diﬀerent things. I can refer to these as follows: \S\ref{subsec:bibfile} ⇒ §3.) Because ﬁgures and tables are not split. The choice of label. was labelled \label{eqn:hessian}.2 \S\ref{subsubsec:webcite} ⇒ §3. see the showkeys package at CTAN [4]. and I used the preﬁx subsec here. I use the preﬁx eqn. I used the structure preﬁx :name. 24 that ~ is used to have a space without a linebreak. 28 For any counter. which is an element of good style. AND REFERENCES
\subsection{The bib File} \label{subsec:bibfile} \subsubsection{Web citations} \label{subsubsec:webcite} Then. (If you put it outside the caption. their exact location depends upon how much room there is. (Recall from p. You can choose any labeling convention that is meaningful to you. The label to reference a ﬁgure or table is put inside the caption.3
Figures and Tables
which have the same
In this section I describe ﬁgure and table environments. is any string you want to use A that does not contain embedded blanks or special characters used by L TEX.2 I can also refer to their page numbers: p. which I can now reference as exercise #4 by writing \#\ref{exer:likeaboutmath}. just as \ref{counter } gives its value. (If you have a lot of labels and need to keep track of them by printing each label and citation in your drafts.40
4 COUNTERS. The environment options deﬁne where the ﬂoat is to be located. gives the page number where its label is deﬁned. it will not be understood. For equations. syntax: \begin{figure}[options] [\caption{caption[\label{label }]}] .2. In the exercise to list what you like about mathematics. The four choices
. 32 p. In my choice of label. as given in [9]. [\caption{caption[\label{label }]}] \end{table}
The caption. page 60. \pageref{counter }. [\caption{caption[\label{label }]}] \end{figure}
\begin{table}[options] [\caption{caption[\label{label }]}] . such as subsubsec:webcite. so in this sentence I wrote its number by \ref{eqn:hessian} (with parenthesis added) and its page number by \pageref{eqn:hessian}.
Table 6: Figure and Table Location Options Option h t b p Meaning Locate here (where the environment is declared). which is why you sometimes see pages with some blank space in the lower portion. if possible. using the \fbox command. and they have separate counters. The table environment is not to be confused with the tabular environment. One of the enhancements is the placement option: [H]. called a ﬂoat page.4. which has no text.” the place where it is speciﬁed in the source. Locate at the bottom of the page (or the next page. The ﬁgures and tables in this document appear as the form: Figure number : caption vs. it diﬀers from a ﬁgure only in its name. and we generally use the table environment to present information in tabular form. notably pictures. which means they are to be placed “here. This is done with the \clearpage command. if this page does not have enough room). only ﬁgures and tables. For example. neither of these A conditions is necessary for their L TEX environments. followed by a ﬁgure or table. This does the same as \newpage. If there is not enough room. The latter makes tables. I did this to avoid confusion by having some ﬂoat appear pages after it is cited and discussed. This option is used in many places in this book. In this document most tables and ﬁgures are speciﬁed with [ht]. Table number : caption That’s it.
. which insists that the ﬂoat be placed here (note the capital H and no other option speciﬁed). perhaps on a page by itself. We might specify [ht] and ﬁnd the ﬂoat in an unexpected place. Figures 33 and 34 illustrate how to frame a ﬁgure with a thick border.
The placement of a ﬂoat is sometimes a source of frustration. except that it also prints all remaining ﬂoating objects. Floats can be framed. it is to be located at the top of the following page. However. but the table environment does not have to contain a table. One cause could be an accumulation of ﬂoats that should be cleared at some point before continuing. As a matter of style.3 Figures and Tables
41
are shown in Table 6. It is also advisable to specify \usepackage{float} in the preamble. we generally use the ﬁgure environment to present what we usually think of as ﬁgures. Locate at the top of the next page. Locate on a separate page.
) Thus.2in} % increase distance to border \fbox{ This is a framed figure. . they will be 1.4
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). (This is called a local setting. Further. . within section 1.
. 1. the frame in Figure 35 has thin lines and no extra padding around the border. more generally. } \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. . The counter values are printed in Arabic numerals.2. .1. The initial value of the counter is 0. You can also deﬁne the counter to be within another counter. the values of mycounter will be of the form s. shown in Table 7. For example. . \newcounter{mycounter} deﬁnes a counter whose name is mycounter. For example. . 2. AND REFERENCES
\begin{figure}[ht] \begin{center} \setlength{\fboxrule}{3pt} % make border lines thick \setlength{\fboxsep}{. upon leaving the ﬁgure environment.
4. This will cause the value of mycounter to be reset to 0 when entering a new section.2. LABELS.
Figure 34: Framed Figure with Caption at Bottom
The parameter settings have returned to their default values. instead of the printed values being 1. but you can specify the type of numeral.
Figure 35: Framed Figure with Caption at Top This is framed with default parameter values. . Also note how the caption is put at the top (see exercise 3). .1.42
4 COUNTERS. . when printed within section s. \newcounter{mycounter}[section] deﬁnes mycounter to be within the section counter. . s.
. For example. to increment mycounter by 1 and establish a label to its value at the place this is done. 4. such as illustrated in Figures 36 and 37. This can also be used to transfer the value of one counter to another. This is done with the \refstepcounter command. . b. These changes remain in eﬀect (called a global setting).4 Deﬁning Your Own
43
Table 7: Numerals to Print Counters What you see a. . For example. . For example. so we must change them back if we want to restore the defaults. but you can change the appearance to be any of those listed in Table 7 by applying the \renewcommand to \thecounter. . For example. . I. \addtocounter{mycounter}{1} adds 1 to the value of mycounter. c. . 3. consider the enumerate list environment. d. . If we just want to increment the counter by 1. .} (the “appearance” parameter is \labelenumii). where the types of numerals for the four levels are: arabic. write \refstepcounter{mycounter} \label{mylabel} Then. . i. IV. Counter values can be set to some absolute value with the \setcounter command. The second level. D. What you write \alph{mycounter} \Alph{mycounter} \arabic{mycounter} \roman{mycounter} \Roman{mycounter}
Counter values can be incremented with the \addtocounter command. whose counter is enumii. iv. roman and Alph. C. II. . . III. The default numeral type is arabic. 1. 2. alph. shown in Table 8. . we can specify \stepcounter{mycounter}. iii. \setcounter{mycounter}{5} sets the value of mycounter to 5. We can change these to be any type we want.4. which also increments its value.
This can be used for intrinsic counters too. we want to be able to label it for future reference. . A. . page). \dots ⇒ (i). \setcounter{mycounter}{\value{page}} sets the value of mycounter to the current page number (value of the intrinsic counter. ii. we can use \ref{mylabel} and \pageref{mylabel} wherever we like. For example. . \setcounter{mycounter}{0} \renewcommand{\themycounter}{\roman{mycounter}} \stepcounter{mycounter} (\themycounter). (ii). B. When using a counter for some nonintrinsic sequence. . \stepcounter{mycounter} (\themycounter). .
. had its label changed to what is speciﬁed in the source: \renewcommand{\labelenumii}{\theenumii. For example.
Write a document with at least two pages and two sections. Groups and ﬁelds B. AND REFERENCES
\renewcommand{\theenumi}{\Roman{enumi}} \renewcommand{\theenumii}{\Alph{enumii}} \renewcommand{\labelenumii}{\theenumii.
A Exercises. (a) Reference §2 by a label that you assign to section 2 (make whatever
.44
4 COUNTERS. and use the \ref or \pageref command to reference each of the following. 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). Be sure your name is on each.} \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. Introduction II. Put an enumerated list of items near the beginning of your document. Submit a printed copy of the L TEX source (tex ﬁle) and printed copy
1. Terms and Concepts A. LABELS.
List 2. . Also reference the page that they appear. . item 1 2. item 2 Now we are out of list 1 . List 1. . 4. . Produce two numbered lists such that the second starts its numbering where the ﬁrst leaves oﬀ. The other form is math display mode. ⇒ A consequence of Einstein’s postulates is that E = mc2 . produce the following: 1. 2. 2. like this: A consequence of Einstein’s postulates is that \[E = mc^2.1 . . . begin list 2. 3. The $ delimiter keeps the mathematical expression in the text. item 2
5
Math Mode
One can write mathematical expressions by entering math mode. Produce Figure 35. . Include two tables and ﬁgures in your document. List 2. List 1. . 1. . . item 1 4. .\] ⇒ A consequence of Einstein’s postulates is that E = mc2 . \]. 2. . and reference them by label. $ or \[. . . (c) Reference item #2 of your enumerated list.2 . . like this: A consequence of Einstein’s postulates is that $E = mc^2$. . . 5. For example. (b) Somewhere near the end of your document reference the page number of the ﬁrst section.2 . . 2.1 .
. . 1. .45
label name you like). 3. signiﬁed by delimiters $. Produce lists using the enumerate environment with the following appearance: 1.
but this is not universal notation. respectively. and some use A. ^.1
Mathematical Symbols
The example also illustrates the use of the superscript operator. Table 9 shows other common operations in math mode. 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
. produced by $A^c$ and $A^\prime$. some authors use Ac or A . The complement of A often appears as ∼ A. Preceding any symbol by \not puts the line through the symbol.) The braces enclose an expression that can be used to deﬁne a more com2 plex operand.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. 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. produced by $\overline{A}$. For example. xa+b is written as $x_{a+b}$ and xa is written as $x^{a^2}$. (Each of the tables in this section applies only to math mode. produced by $\sim A$.
\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. digits and accents. this applies only to letters. it. math fonts would remain bold. 8). tt} (analogous to the \textfont command. and returns to normal style in the second case. however. even when leaving and reentering. without the braces. p. ˜ {\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 A×˜⊗2 ˜ A×∞⊗∈ ˜ ×1 ⊗2 A ˜ ˜ A×⊗ ˜ 1 A×˜⊗2 ˜ ×˜⊗2 A 1 ~×~⊗2 A 1
The calligraphic style applies only to capital letters. to produce α−β =∆−δ
. Note that \boldmath is surrounded by the braces. Greek letters are deﬁned only in math mode. separate from text mode. the calligraphic fonts remain in eﬀect: $\cal P = A + B$ ⇒ P = A + B. Font style. rm. otherwise. normal. causing unintended results when applied to other symbols. For example. but not to special mathematical symbols. The calligraphic alphabet looks like this (and it is available only in math mode): ABCDEFGHIJ KLMN OPQRST UVWX YZ. \mathfont{expression}.5. does not apply to math mode because math mode has its own. and they are speciﬁed by spelling them as keywords. Write ${\cal P} = A + B$ to produce P = A + B. we can control the font style of letters with the command. For example. Unlike \boldmath. where font is one of: {bf. For example. sf. as shown in Table 11. You can make math fonts boldface by specifying \boldmath before entering math mode. {\Large $(x\div y) + z$} ⇒ (x÷y)+z . cal. where B ∪ C is boldface in the ﬁrst case. {\boldmath $x^n+y^n=z^n$} ⇒ xn + y n = z n . The following illustrates this.1 Mathematical Symbols
47
You can control the size of the font by using the usual speciﬁcation before entering math mode. For example.
“Making Greek letters is as easy as π (or Π)” (written $\pi$ or $\Pi$). as 4 .
5. Table 12 shows some of the most common of these. p. the indices on the A sums and product appear as they would in line.2
Fractions and Variable Size Functionality
To make fractions. 43] says. 1 + x2η +1 written as \[ A = \frac{x^2+y_\alpha}{1+\frac{\eta}{x^2+1}}. Figures 38 and 39 illustrate this with another example. by preceding the math mode with \large. The general form is \frac{numerator }{denominator }. We can make this appear x+y larger. Some mathematical symbols adjust their size to ﬁt the expression. but also produces proper spacing. \].\delta \]. then \bm{\beta} ⇒β. and I present more examples below. As Lamport [9.). 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. This inserts a thin space (compare the results by writing the expression with and without A the In L TEX. L TEX compilers make judgments about the layout. In the preamble specify \usepackage{bm}.) The \mathbf does not make Greek letters boldface. symbols whose size you would want to adapt to expressions are gen\.
. Here is a more complex equation in math display mode: x2 + yα A= . note the use of \.48
5 MATH MODE
write \[ \alpha . Note how the sizes of the fractions adjust automatically. (Not every Greek letter is included — see Appendix Table 36. We could use \boldmath to achieve this. but you can force either of the two styles with the \displaystyle and \textstyle commands.\beta = \Delta . but if we want x+y 4 . In the case of the integrals. we use the \frac command: $\frac{x+y}{4}$. we could write $(x+y)/4$ to make (x + y)/4. erally designed to do so. between the integrand and dx. but there is a package that not only provides the boldface font. Figures 40 and 41 illustrate this. where the numerator and denominator can be any expression.
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.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)\. in particular.
. The “default” is not always predictable. as well as sizing the expression. as though it were in display mode. using \textstyle and \displaystyle to override the default form for the 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)\. math display mode does not always use displaystyle. Figure 42 gives more examples to compare in line text and display mode.5.
\] The quantiﬁers in this last example seem a bit crowded. write \[ \forall x\exists y\ni [P(x)\wedge Q(y)]. For example. In math mode a full space is obtained by specifying \. \] To have ∀x∃y [P (x) ∧ Q(y)].\{x: x\not\prec y\} \not\subset {\cal A} ⇒ ⇒ ⇒ (−∞.) Table 14 shows some relations for ordered sets (besides those on the keyboard: < = >). shown in Appendix Table 35. Here are some examples: (\infty. 0] = {x aj bi ≡ bi x ≤ 0} aj
∀y {x : x
y} ⊂ A
. to have (x ∈ A ⇒ x ∈ B) ⇔ (A ⊆ B). so we might want to add some spacing between terms.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. \exists y ⇒ ∀x ∃y \forall x\. Here is how each looks: \forall x \exists y ⇒ ∀x∃y \forall x\. write \[ (x\in A\Rightarrow x\in B) \Leftrightarrow (A\subseteq B). and a half space by \.. including negative spacing. \exists y ⇒ ∀x ∃y (There are other spacing commands.0] = \{x\ni x \le 0\} a_j\prec b_i \equiv b_i \succ a_j \forall y\.
Compare each of the following: xi < 0f oralli = 1. and more.3
Arrays and Equations
The array environment is to math mode what tabular environment is to text mode. . xi < 0 for all i = 1.
5. . but we can also do the reverse with the \mbox command. The use of \mbox is particularly convenient in math display mode.3 Arrays and Equations
51
Table 13: Some Symbols in Logic Logical Term existential quantiﬁer universal quantiﬁer negation disjunction conjunction implication equivalence such that How it appears What you write ∃ \exists ∀ \forall ¬ \neg ∨ \vee ∧ \wedge → \rightarrow ⇒ \Rightarrow ⇔ \Leftrightarrow ≡ \equiv \ni
Table 14: Order Relations Relation less than or equal greater than or equal not equal precedes precedes or equals succeeds succeeds or equals How it appears ≤ ≥ = What you write
\le \ge \ne \prec \preceq \succ \succeq
We have seen how to embed math mode into text. . It has the form:
. written as $x_i < 0 for all i=1. . . .\dots$ written as $x_i < 0$ for all $i=1. . . which I shall illustrate in the next section.\dots$
The ﬁrst line points out that blanks mean nothing in math mode.\dots$ written as $x_i < 0 \mbox{ for all } i=1. xi < 0 for all i = 1. . and all letters are in the math form of italic (not quite the same as the italic in text mode).5.
234567 1 y −9. so it can be generated in either of two ways: with the tabular environment. . otherwise. last row spec [\\ options] \end{array}
5 MATH MODE
The column speciﬁcations and options are the same as in the tabular environment. The following table has text headers and math body. 7. (= x+y) \end{array} \] The \.7 (= x + y) The above was produced by the following use of math display mode (which is always centered): \[ \begin{array}{lcl} x &=& 5. $.2 \\ y &=& 2.5 \\ z &=& 7. .234567 & 1 \\ y & 9.234567$ & $ 1 $ \\ $y$ & $9. using \mbox for each header entry. . but the body is in math mode. or with the array environment.7(= x + y). speciﬁes a space.52
\begin{array}{column specs}options ﬁrst row spec \\ . . using the math mode designation for each body entry: $.87 $ & $12.2 \\ \hline \end{array} \]
or
\begin{center} \begin{tabular}{ccc} Variable & Current Value & Limit \\ \hline $x$ & $ 1.
.2 y = 2.87 −12.7 (= x+y) ⇒ 7.2 $ \\ \hline \end{tabular} \end{center}
You can align a series of equations to appear this way: x = 5.87 & 12.5 z = 7. Variable Current Value Limit x 1.7 \.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.
parentheses are added. anything could be used for the middle column.) We use the eqnarray environment directly (without entering math display mode). specialized to this column speciﬁcation. There is no apparent advantage to this since the same result can be produced by the ordinary array environment. but without the equation numbers. Further.) The relation need not literally be an equation. 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. (Note that \ref gives just the number. which is the same as eqnarray. 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). as above.
. there are times when we need to use more than one line for an ‘equation. Figures 43 and 44 give an example. however. This is like a 3column array with speciﬁcations {lcl}. It does. 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)
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.’ in which case we need to suppress the numbering of all but one of the rows.5.3 Arrays and Equations
53
Another environment is eqnarray. but each row is numbered: x = y = y z (1) (2)
(Another diﬀerence is that the eqnarray environment uses displaystyle.
Analogous to eqnarray*.3 \\ 21. note how x is speciﬁed.2 1.54
5 MATH MODE
For a single. This poses no particular advantage over specifying eqnarray and merely entering one row (except that column separators (&) are not used in the equation environment).2 & 1.0 22. Also.0 & 22.
. numbered equation. and the horizontal line separating the blocks is obtained by specifying \hline before the second row of the outer array. which suppresses the equation numbering.1 \\ \end{array} \right] \left( \begin{array}{cc} x_1 \\ x_2 \\ x_3 \end{array} \right).1 & 1. Notice how the vertical line was drawn by the column speciﬁcation. \end{equation*}
Figure 45: Matrix Equation Source (Result in Figure 46) Ax = 1. x3
Figure 46: Matrix Equation Result (Source in Figure 45)
Array environments can be nested. Table 15 shows other ways to denote the transpose of a vector.1 1. there is the equation* environment.1 x1 x2 .0 & 2.0 −2.3 21. {cc}.
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. To illustrate. as illustrated in Figures 47 and 48. Figures 45 and 46 show how to present a matrix equation. there is the equation environment.
rather than absolute measurements for spacing.)
. Note how the line height does not adjust to the frame. putting \vspace{.3 Arrays and Equations
\[ \left[ \begin{array}{cc} \begin{array}{ccc} A_{11} & A_{12} & A_{13} \\ A_{21} & A_{22} & A_{23} \end{array} & 0 \\ \hline 0 & \begin{array}{cc} B_{11} & B_{12} \\ B_{21} & B_{22} \end{array} \end{array} \right] \]
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. sometimes used for emphasis.2\baselineskip} after x = y + z causes extra vertical space equal to 20% of the value of \baselineskip. which you might change. such as writing $x = \fbox{y} + z$ to produce x = y + z.dx \end{array} $}
∞ 0
xe−τ x dx =
1 τ
b+c
⇒
=
a
Ψ(x) dx
We can use \fbox within math mode. (In the longrun.dx &=& \displaystyle\frac{1}{\tau} \\ \\ &=& \displaystyle\oint_a^{b+c} \Psi(x)\. causing an undesirable clash. This could be overcome by putting a vertical space command just after the expression. which is the height of one line of normal text. it is better to use parameters. For example. because the former takes into account the font size. \fbox{$ \begin{array}{lcl} \displaystyle\int_0^\infty xe^{\tau x}\. like \baselineskip.5. In particular.
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 inline 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
5. ⇒ ⇒ ⇒ ⇒ How it appears R C Z Q What it means Real values Complex values Integer values Rational values
Another alphabet is \mathscr. for which you specify \usepackage{mathrsfs} in the preamble. rather than . $\Re$.x_n \max_{x\in X}f(x) \frac{\tan(\theta + \pi)}{\ln\.x}
lim xn
lim inf log xn max f (x)
x∈X
tan(θ + π) ln x
For example. 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.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\. the real line is sometimes denoted by R.
.
Table 18: Notation Using mathbb Fonts from amssymb Package What you write† \mathbb{R} \mathbb{C} \mathbb{Z} \mathbb{Q} † In math mode. Table 18 shows how \mathbb can be used for specifying other numerical spaces. L is often used to denote the Laplace transform or the Lagrangian. which is A the L TEX special symbol.
dv . In L TEX it is produced by \nabla.dx} {\oint_{X(v)} e^{\lambda f(x)}\. ∂xi ∂xj
(6)
There are two integral signs: \int⇒ and \oint⇒ . The partial derivative symbol. H.5
Derivatives and Integrals
We can express a total derivative. . note how the outer integral is large in the following expression: b xeλf (x) dx X(v) lim Φ(v) dv. denoted by the symbol . which are both variablesize symbols. (also called “del”). \]
. .
A I leave it as an exercise to show the L TEX code that produced equation (5). which is an upside down delta (introduced by A Hamilton in 1853). .\Phi(v)\. . and {\large$\frac{\partial f(x)}{\partial x}$} to produce ∂x . ∂f (x)/∂xn ). ∂. Compare with the following and see if you can produce it:
2
f (x) =
∂ 2 f (x) . is written \partial. \end{array} \] There seems to be some crowding in this direct speciﬁcation. The Hessian is the matrix of second partial derivatives: 2
∂f (x)
(5)
f (x) =
∂ 2 f (x) . (Compare with \mathcal.)
5. 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)}\. by writing $df(x)/dx$. ∂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].dx}\right \. and its mathematical deﬁnition is the vector of ﬁrst partial derivatives: f (x) = (∂f (x)/∂x1 . we can use (x) the \frac command to produce dfdx . or. which is also used by some authors. df (x)/dx. so you can write \partial f(x)/\partial x to produce ∂f (x)/∂x. The usual notation for the gradient of a function is the nabla. For example.60
5 MATH MODE
and H is sometimes used to denote the Hamiltonian.
\. 223]): \usepackage{amsmath} in the preamble (see The L TE (u v − v u) · dS =
S τ
(u
·
v−v
·
u) dτ. there is no solution to $x^n + y^n = z^n$ for \newline $x. but is also the name of the associated counter.) Deﬁnite multiple integrals are no problem. which is not only the name of the environment. Notice how “Theorem 5. there is no solution to xn + y n = z n for x.
5. like “Theorem.1 For n > 2.y.z\in \LZ_{++}$.6 Theorems and Deﬁnitions
(Note the use of the thin space. . To have
∞ 0 0 xn 0 xn−1 x2
61
···
0
H(x1 . all text is in italic. This was deﬁned in the preamble by: \newtheorem{theorem}{Theorem}[section]
A Then.x_n)\. \end{theorem}
.1. and the spacing of the integral signs.” and a name. consider the following: (u v − v u) · dS =
S τ
(u
·
v−v
·
u) dτ.5.
Note how the domains are centered on the multiple integrals and the spacing of the integral signs. xn ) dx1 · · · dxn
write \[ \int_0^\infty \int_0^{x_n} \int_0^{x_{n1}} \cdots \int_0^{x_2} H(x_1.1” appears.
The domains of integration. but by specifying A X Companion [5. which is not produced by standard L TEX 2ε .6
Theorems and Deﬁnitions
The foundations of mathematics are axioms and rules of inference. The rules create theorems. the theorem was produced by the following L TEX code:
\begin{theorem} For $n > 2$.dx_1\cdots dx_n \] However. are better with A the following. p. which are statements whose truths are established relative to the A underlying logic. This is so fundamental that L TEX has the facility to deﬁne a special environment that includes a keyword.. .\dots. y. Consider the following example: Theorem 5. . and we have the counter value: \thetheorem=5. . z ∈ Z++ .
\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}. but this is generally not desirable for a deﬁnition. Then. here is a corollary environment: Corollary 5.62
5 MATH MODE
Other theoremlike environments can be deﬁned to have the same properties. \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. Consider the following example: Deﬁnition 5. \end{corollary} The following creates an axiom environment that is not within any other counter. Here is the syntax: \newtheorem{name}{keyword }[within] The name deﬁnes the environment name. 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. To further illustrate. 42) or by some other \newtheorem.1. so you see Theorem 5. The environment created by \newtheorem puts the text in italics. which can be intrinsic or some other counter deﬁned by the \newcounter command (p. which is valid by having been deﬁned by its own \newtheorem. and it deﬁnes a counter. This requires both a keyword. The within option deﬁnes the counter to be within some other. like theorem. and a unique name for the environment.1 The sum of cubes cannot be a cube.1. This was created by ﬁrst entering (in the preamble):
. In this document. also used as a counter. It was deﬁned in the preamble as follows: \newtheorem{corollary}{Corollary}[theorem] Note that this is within the theorem counter. the above corollary was written as: \begin{corollary} The sum of cubes cannot be a cube. I deﬁned the theorem environment to be numbered within the section. so it must be diﬀerent from all other environment and counter names. like Theorem. rather than Theorem 1.1 The circumference of a sphere is the circumference of any great circle on the sphere.
especially the thicknesses. delimiter size control commands. .
$\big(E=mc^2\big)$ ⇒ E = mc2 . .
There are. and \Bigg. One way is with a size command — for example. but they are diﬀerent. {\Large[}$E=mc^2${\Large]} ⇒[E = mc2 ]. in the text: \begin{mydefn} The \textit{circumference of a sphere} is the circumference of any great circle on the sphere. we can also enlarge these delimiters ourselves.5.7 Reﬁnements
\newtheorem{defn}{Definition}[section] Then. Huge↔Bigg). This was created by ﬁrst entering (in the preamble): {\theorembodyfont{\rmfamily} \newtheorem{mydefn}{Definition}[section]} Then. (large↔big.
. . This is more evident with the square and angular brackets:
$\Big[E=mc^2\Big]$ ⇒ E = mc2 .
5. the theorem package enables a wide range of variations over the font style (among other things). however. . Whereas \left and \right commands adjust the size of a mathematical delimiter to ﬁt the enclosed expression. For example.7
Reﬁnements
Mathematical delimiters.
The use of text font environments comes close to the corresponding math size. $\bigg\langleE=mc^2\bigg\rangle$ ⇒ E = mc2 . \Big. \bigg. must be varied to enclose some expressions.
{\large(}$E=mc^2${\large)} ⇒ (E = mc2 ). \end{defn} Compare this with the following:
63
Deﬁnition 5. \end{mydefn}
For more customization. which apply to a single character: \big.1 The circumference of a sphere is the circumference of any great circle on the sphere. like parentheses and braces. in the text: \begin{defn} The circumference of a sphere is the circumference of any great circle on the sphere.
The amsmath package has a command to put dots across any number of columns in an array. Figures 55 and 56 illustrate this. 61 for obtaining better multiple integrals). except the equations are not aligned.
The remaining reﬁnements use the amsmath package (introduced on p. Its syntax is \hdotsfor[spacing]{n}. in math mode. we could produce a b c d by spec
ifying $\left(\begin{array}{cc} a&b \\ c&d \end{array}\right)$. where spacing determines the spacing between the dots. The same result with equation numbers is obtained by the gather environment. in particular. and n is the number of columns it spans. the spacing and parentheses are not the same. For example. They behave like the eqnarray and eqnarray* environments. respectively. b \scriptsize produces a d . An alb ternative is with the amsmath smallmatrix environment: a d is obtained c by $\left(\begin{smallmatrix} a&b \\ c&d \end{smallmatrix}\right)$. \\.
. (Note that there are no column speciﬁcations.) This is not equivalent to preceding the array speciﬁcation with a text size environment. The gather and gather* environments allow the new line speciﬁcation.64
5 MATH MODE
{\LARGE$\langle$}$E=mc^2${\LARGE$\rangle$} ⇒ E = mc2 .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. While the letters inside the matrix are apc proximately the smallmatrix size.
\begin{gather*} (a+b)^2 = a^2 + 2ab + b^2 \\ {\cal L} \oplus M^\varepsilon .
...... The
.) The horizontal arrows are speciﬁed by @>>> (left to right) with any expression placed above or below.. With more generality.
The \stackrel command lets us put characters over a relation: For def example. ... For example.... Figures 57 and 58 illustrate this...... ....5..... 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... n^+\stackrel{\mathrm{def}{=}n+1 ⇒ n+ = n + 1... (Specify \usepackage{amscd} in the preamble.5]{5} \\ \end{array}\right
65
1 2 3 4 5 . the \overset and \underset amsmath commands enable us to put any characters over or under any character...... .. which makes it easy to draw commutative diagrams.....7 Reﬁnements
\left\begin{array}{ccccc} 1 & 2 & 3 & 4 & 5 \\ \hdotsfor{3} \\ & \hdotsfor{3} \\ \hdotsfor[2]{5} \\ \hdotsfor[..
$\overset{a}{X}$ $\underset{b}{Y}$ $\overset{a}{\underset{b}{Z}}$
⇒ ⇒ ⇒
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...
” But a group had not yet been deﬁned.8
Grammar
When writing mathematical expressions. . 1. For example.
5. Chapter 8]. The general guide is to treat a mathematical expression linguistically. we might see “The distinguishing property of an abelian group is the commutivity . clauses are separated by commas. For example. and phrases are appropriately punctuated. with an expression placed to its left or right. and more packages to make things A nicer. 2. note the colon before the display and the comma at its end. Here are some of the most common elements of grammar to consider. where P is a permutation matrix. Punctuate math display mode. The expression usually needs a comma or period.
\[ \begin{CD} A @>a>> \alpha @VV\beta V D @>>d> \end{CD} \] 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. In English this means that every sentence has a subject and predicate. 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. people make some common errors.66
5 MATH MODE
vertical arrows are speciﬁed by @VVV (down) or @AAA (up). and you can see an online catalog of packages at CTAN [4]. A symmetric rearrangement of a matrix has the following form: R = P t M P. .
. All possible horizontal and vertical placements are illustrated. which is incorrect to omit. Many of these are described in The L TEX Companion [5. Deﬁne before use.
Suppose A and C are expressions. Then. (a) ln ex = x (b) sin{θ + 2π} = sin θ
. C. it follows that ∆2 Fn+1 = ∆Fn . if we write Φ = au + bv. . If either A or C are compound.’ seems like it ought to be all right. Submit a printed copy of the L TEX source (tex ﬁle) and of the
associated postscript result (ps ﬁle). 0 otherwise. 6. In English. when there is no risk of confusion. “We shall use Φk . ‘If A. v). Sometimes we deﬁne the complete object.’ or ‘Suppose A. The form. (Lookup special symbols in the Appendix. ∆2 Fn = Fn+2 − 2Fn+1 + Fn = 2Fn − Fn+1 . (c) x+ = x if x ≥ 0. . 4. and only if. the second form is clearer. v) = au + bv. Reference object is located after the reference. Φ(u. Equivalence needs commas. Produce each of the following in math display mode. instead of Φ(uk . We can write either ‘If A. then C. 5. Be sure your name is on each. and the comma is used to clarify where the antecedent (A) ends and the consequent (C) begins.” The overriding principle is clarity.8 Grammar
67
3.’
A Exercises. An object has only one deﬁnition. and it is important that the reader be told of this. . For example. B. is not correct. v k ).
2. Produce the following in math display mode with the array environment and/or with the eqnarray* environment. then . For example. then tell the reader something like. we cannot later refer to Φ(u. Produce each of the following formulas in line with text (construct your own sentences that contain them. but you might want to take control over locating ﬁgures.’ The ﬁrst form is preferred if A and C are simple expressions. (b) If ∆Fn+1 = Fn . L TEX does this automatically. and include proper punctuation). √ (a) x2 = B 2 − 4AC implies x = ± B 2 − 4AC. ‘A if and only if B. however. . If .5. a A ﬁgure appears after its ﬁrst reference. this is not correct.’ should be written as ‘A if.) 1. C. The expression. 3.
.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 4.1 −1. xT ) √ 3 3 V = 2
a 0
A B 0 C
=
0 xT C
= (0. (a) A = {S ∈ S : S ∈ S} (b) F × P ≤ π a11 a12 a21 a22 CB BC
? def
(c)
9. √ A−1 (b) (0. Produce equations (5) and (6). Produce the expression in the Preface. C T x)T . .
5.2 2. . 1. Produce each of the following formulas in math display mode (with punctuation): 2m(GA ) if G = ∅.3 −2. xT )T = (0.
(c)
(−x 3 + a)2 dx. 6. Produce each of the following expressions: (a) x = y mod n = x − y = kn for some k = 0. Produce the following equation in math display mode. 7.1 2.2 −1. (b) α1 + β 2 − x2 + y 3 8.
1
. 1.3 + a1 a2 a3 b1 b2 b3 = α −β γ −δ λ θ . Produce each of the following in line with text (that you compose) and in math display mode. (a) q ∗ (G) ≥ max ∆(G).
x
. (b) A result of these assumptions is the following equation E = mc2 Einstein ﬁrst noticed this equivalence between energy (E) and mass (m). Deﬁnition 2 A matrix is nonsingular if it is not singular. + 12. What is grammatically wrong with each of the following segments. ∂f (x)/∂xj } if xj = aj + ∂f (x)/∂xj if aj < xj < bj f (x)j = min{0. (c) Complex nvectors: Cn . 14. ∂f (x)/∂xj } if xj = bj 11. and mathematical symbols to produce the following (called the truncated gradient): max{0. Produce the following: 13. Therefore. (d) Nonnegative rational nvectors: Qn .
∂f (x) ∂xj x=¯ . Produce the following symbols: (a) Extended reals: R∞ .8 Grammar
69
10. conditional assignment (with array environment). (a) A key is how to add velocities the formula is (u + v) 1 + uv c2 where c is the velocity of light. Proof: The determinate of a nonsquare matrix cannot be zero because it is not deﬁned. Theorem 1 Every nonsquare matrix has an inverse.5. Produce the following: Deﬁnition 1 A matrix is singular if its determinate is zero. the matrix is nonsingular. (b) Strictly positive integers: Z++ . This implies it has an inverse. Combine your knowledge of derivatives.
(1/2)1/4
16. (e) Theorem If x. You are to produce the mathematical expressions shown in math display mode. d f (x + t dt
+
1 2
to be
f (x))
t= 1 2 b 0 1 1 0
=−
2v + 1. Figure 1 (above) shows how to add velocities simply as vectors. Note the row and column labels outside the matrix. Adding velocity vectors: u + v. Column pointers: A=
.70
5 MATH MODE (c) Let x be an nvector and ω a scalar. The following is tricky to get the evaluation expression. y. where A is an m × n matrix and b is an mvector. then n < 3. and deﬁne y = Ax − ωb. (d) Now we consider adding velocities. Now suppose y(ω) is speciﬁed and we want to ﬁnd x. z ∈ Z+ and xn + y n = z n . 15. 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. 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. t = the right size and location.
u
u+v v
Figure 1. The remaining exercises are more diﬃcult.
6. 2. and r = right. we can do this simply with \fbox and a long arrow in math mode.71 19. By itself. \framebox also has two optional arguments to control the length of the box and the position of the text within it.
. 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.
3.
\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. Use a graphics package to draw within the document.1
Picture Environment
If all we want is a series of boxes and arrows. I illustrate each. The second optional argument is the position of the inscribed text: l = left. However. notably the picture environment. c = center. The ﬁrst optional argument of this \framebox command is the width of the box. but I do not provide a complete list of the relevant packages A (see CTAN [4] and The L TEX Companion [5]). 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. given as 2 cm for each box. it lets us stack short phrases. Use standard L TEX 2ε commands. We can make the contents of a box obey all paragraph controls in text mode by the \parbox command. For example. Use a package to import some standard graphics ﬁle.
These commands can be combined.72
6 GRAPHICS
top like middle (note how the paragraph spacing adjusts). respectively. and more control over positioning. whose source is shown in Figure 62. Going through its parts will serve to explain the various commands.b
c
Z Z Z ~ . To begin.
\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
. Figure 61 shows a more elaborate chart. as illustrated in Figures 59 and 60. which was created by the picture environment. Combined with bottom \framebox. but there is a need for more versatility. we can create vertical diagrams easily. like ovals and diagonal arrows. A basis for this is the picture environment. but it has an optional argument to align its top or bottom with the text. along with other box commands. This is done by specifying \parbox[t]{width}{text} or \parbox[b]{width}{text}.
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. This means that when
.
we enter the picture environment stating that the point of entry is the origin.1}} ⇒ v The next three commands put three diﬀerent kinds of boxes.5){\dashbox{.5 inches below the origin (i.6..43}} \put(1.. which is exclusively for the picture environment. The parameter that determines this is \unitlength.3975}} \end{picture} \end{center} \vspace{1in}
73
Figure 62: Source for Figure 61 I specify some length = 5. indicated by the coordinates (0.5.0){.1){.4}} \put( . 1){\circle{. and the default for the picture environment is 1 pt.65. The complete syntax is: \put(x..y){stuﬀ }.55}} \put(.2}} \put(..0) \put( 0.0){. but also the height. The ﬁrst is similar to \framebox in text mode. where stuﬀ can be text or some picture object.5){\dashbox{.. it could p be a column in a table deﬁned within the tabular environment. or ph it could be in the middle of a sentence.01}(.05){1} \put( 1.3){.1){\vector(1. The
.55}} \put( 0. as in beginning a paragraph with \noindent. I am specifying 5 inches.0){\circle*{.) The ﬁlled circle shows where (0.1){\vector(1.25)} \put(. Then.35.85.1}(1..1){\fbox{$\begin{array}{c}a\\b\\c\end{array}$}} \put(. The ﬁrst \put in Figure 62 speciﬁes the position at the origin...0) is in this picture. Every picture command begins with \put..0). which is not described here.1..e.2.5){\vector(4.1 Picture Environment
\begin{center} \setlength{\unitlength}{1in} \begin{picture}(0.5). y = −.7.7.5}} \put(.2. y) coordinates are relative to where the position is when the picture environment is entered. 1){\oval(. b = bottom.55. but its syntax is diﬀerent.3)[br]{bottom right} } \put(.7.3.1 inches (centered at the origin): \put(0.1){. and this extends the position options to a second character: t = top.895){\line(1.0){. 0){\circle*{.5){\framebox(.05){oval} \put(0.3){center} } \put(1. (There is an alternative way to begin the picture environment.65.1}} \put( 0.35){\vector(1.0){.3}} \put(.3)[tl]{top left} } \put( 1. and the stuﬀ is a ﬁlled circle with diameter .5){\vector(0.32. each beginning at . This could be at the left margin. just as the smiley face appears here (see Exercise 1).1){\vector(1.. The (x.1. In picture mode it enables control over not only the width..
3 inches. which are lines with arrow heads. Otherwise. I use the \fbox command. and len is the amount of change above or below the original point (it does not matter what the magnitude of ∆y is. resulting in fewer dashes to compose the box.height)[posn]{text}
In the example shown in Figure 62. located at coordinates (−. the line is vertical. This is the same as I used in text mode. The next \put puts a dashed box. the speciﬁcations are width = . 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.65. Now we come to the \circle speciﬁcation. We know the center of the circle is at (−.01 inches.65. where . [br]. only its sign matters). and the text is at the bottom right because of the optional speciﬁcation. with diameter = . except here I use it to frame an array. If ∆y = 0. The same applies to the \oval speciﬁcation.25 (inches). The next dashed box has the dash length set equal to . Both \vector and \line have the same syntax:
\line(∆x.1 inches. having the same dimensions as the framed box. deﬁned as usual in math mode: the array has three rows and one column. if ∆x = 0. The new point is determined by moving
. The oval.∆y){len}
If ∆x = 0.7 inches and height = . the line is horizontal. Now the code begins to draw the vectors.height ? %
width
After the \oval speciﬁcation. the actual ∆y change in x is still len. −1) (from \put). Unlike the box family. the position is centered because that is the default.∆y){len} \vector(∆x.2 inches. we cannot include the centering of text within the circle command. only its sign matters).2 inches.74
6 GRAPHICS
general form of the \framebox command in the picture environment is as follows:
\framebox(width. and the slope of the line is ∆x . The box length is set to 1. This is undoubtedly confusing.5 × . but that is not where we want to put the inscribed text to be centered. followed by putting text that required some trial and error to locate. itself.5 measures the entire width:
' & $ 6 . with the length of the dash set to . −1). which is centered. has dimensions . so consider Figure 63. The “1” inside the circle required another \put. and some trial and error was needed to establish its position.
Then. The ﬁrst \vector command in Figure 62 starts at (−. for example. If we we set len = 1. 5 Fixing len = xt − x0 . We could setup a leastsquares estimation problem. If xt = x0 .
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 trial and error in selecting the parameters tends to be just as eﬃcient. The ∆y closest we could come is 4 . we have some work to do. we would set ∆x = 15 . we want (xt . which I calculated to be from the “top left” box to the “center” box.3 to obtain the correct xcoordinate.3. y0 + 1. Suppose our original point is (x0 . and 1 if yt > y0 . how should we set the slope parameters? 13 Ideally. Suppose. is not necessarily the best overall approximation. yt ) = (x0 + 1. If xt = x0 . y0 ) along the line with slope ∆x until the new xcoordinate is ∆y x0 + len. and we want our destination point to be (xt . −1).6. the new ycoordinate is y0 + len ∆x .65. the calculation is simple: set ∆x = 0.1 Picture Environment
75
∆y from (x0 . we could have problems with approximating the results. len = yt − y0 . there is an important restriction: ∆x. y0 ). The actual length of ∆y 2 the line segment is len 1 + ∆x .5). then searching for a nearest slope approximation. Either way. yt ).
. but the restrictions do not permit this. ∆y = −1 otherwise. ∆y must be integervalued and within −6 to 6.
That 2 accounts for the initial position given by \put(. then an arrow at the same end points.. speciﬁed as {. yc ) = (−.3. Finally. but we
?
.3}. and its xcoordinate is to the right by the length of the radius: (x1 .35). so the coordinate where the arrow begins (called its tail ) is (−.3}}. this involves more calculations because the arrow is not simply horizontal or vertical. −.2) since . The length is determined by where we want the arrowhead: at the top of the circle. −.3. so the right edge of that box is at x = −. as speciﬁed with \vector(1.3) = . . by computing the coordinates of the tail and head. where the 2 box starts at (x0 . −.0){. drawn left to right. −.55.3. Since the arrow is downward.35){\vector(1.3. The uncertainty is the width of the box.5) = (−.5) + . which is what is speciﬁed: top left ⇐ \vector(0. −1) = (−. and its width is . as speciﬁed. −1).3.2}. y0 ). so ∆x = 0 − (−. but drawn right to left. −5). and we set len = ∆y = .5) and h = .2 is 2 the diameter speciﬁed by \circle{. We must add the radius.1) = (−. (x0 . Thus. yh ) = (xt . The circles y coordinate is −.1). y0 ) = (−1.65. −.76
66
GRAPHICS
width

height
?
box begins here → c
In Figure 61 the “top left” box starts at (−1. yt ) = (−1 + 1 2 .0). which is its center.65.3) = (−.4} 1m The next arrow is doubleheaded.4.5). given by \vector(0. we need to determine the length. That box begins at (0. yc − yt + r) = (−.. We want the coordinate of the end of the arrow (called the head ) to be ﬂush to the left side of the “center” box.7. the vertical position changes by moving up half of the height. Starting at y = −.3. −. we know only that the center of the box was put at (0. Now consider the next \vector. so we obtain the coordinates of the arrow’s tail: (xt .3.1. so we use two \vector commands to draw one arrow left to right. It required these computations to determine the complete picture command: \put(. ∆x = 0 and ∆y < 0. In this case. −1).5 + 1 . We begin the same way. The left end point is at the ycoordinate of the center of the circle. which is a vertical arrow from the same box to the circle below it. y1 ) = (xc + r. −1 − (−.65. y) = (x0 + 1 h.65. Further..4).65 + . and this needs some trial and error.1){.5). the position of the arrowhead is (xh .35). so that is where we \put the ﬁrst arrow. which is 1 (.35.3. y0 ) and h = height. The initial position is calculated simply as the midpoint of the bottom edge of the box: (x. The head is to be ﬂush with the left edge of the \fbox. so ∆x = 1 and ∆y = 0. The arrow is to be horizontal.
5) and has a width of 1.395) = . the end point was determined to be x = 0. and there is very limited control over line thicknesses. • Some calculations and some trial and error are needed to align objects and lines.5). Now the true slope of the line we want is .3975} to obtain the line shown in Figure 61. called Bezier approximations. These can make using the picture environment time consuming and rather unpleasant.55}. p. • There is no direct way to control the size or style of the arrow heads. The last vector also required trial and error. Table 46 (in the Appendix. In this case. The closest slope we can have is with (∆x. which is what is speciﬁed.6. 2 Thus. However. With just a few iterations. so the midpoint of the bottom edge is at (1. we specify \line(1. 1). and we can plot curves. the end points were determined to be from (1. The reverse arrow begins at (0.3975. in which case you specify the parts you use instead of pstall — see [15] for loading individual portions.55. but here are some things to note: • Only boxes can have inscribed text. 117) gives the commands in the picture environment. Given this slope.4 + .5). There are packages to extend the picture environment.6.2 PSTricks
77
do not know the width of the box. which is why we have \vector(1.0){. but you can obtain it at CTAN [4].895) to (1.6. the circles and ovals require separate \put commands. which can take some trial and error to position.1){. −. (It is not standard with MiKTeX.4 do not allow this. so len = ∆x = . −1).2. I shall cover these in the next section with a powerful package called PSTricks.2.)
. the best choice of len can be found as the average of the deviations: len = 1 (. (You can use parts. −. −. There is a better way!
6. ∆y) = (1. due to not having the corner of the oval coordinates. requiring recalculations and more eﬀort for the new positions. but the restrictions . • Moving a portion of the picture can be tedious.2
PSTricks
PSTricks [15] was written by Timothy Van Zandt. The former was found by trial and error. −. to a set of points.) In the preamble specify \usepackage{pstall} for the entire system. 0). and is provided free of charge. but the latter was computed by knowing that the “bottom right” box starts at (1. and its slope is (−1.395 .
Hobby. • Drawing curves is simple. which makes it potentially more versatile.78
6 GRAPHICS
One thing you need to know is that not all of the pst results can be seen with a dvi viewer. typically available free of charge. for short) is designed to overcome diﬃculties with using the picture environment.fillcolor=gray}
A fundamental command in pst is \rput. There are many packages [4]. can have inscribed text. themselves. • Shapes are highly variable. and slopes need not be calculated. In particular. 7]. can be set with the \psset command: \psset{parameter = value[. PSTricks (pst. this is not the only way to put objects. . . the default unit of measurement is 1 cm. The defaults. but unlike the \put command in the picture environment. It is more diﬃcult to learn than PSTricks. ]}. Table 19 gives
. written by John D. and Bezier approximations of four points are available. All of the pst commands have options to override default settings for relevant parameters. This is especially true of commands that involve rotations. The commands. Some require converting to postscript and viewing the ps ﬁle. Another widely distributed picturedrawing system is MetaPost [6. but it does with MetaPost. can specify where to put them. that do many of the things done by PSTricks (and some additional things). • Objects can be named (as nodes) and lines and arrows can be drawn between them by naming the tail and head. and the default ﬁll color is white. but MetaPost is more openended in its design. especially on varying the types of ﬁle outputs (PSTricks is tied to postscript). including plots of points that can come from a data ﬁle. For example. • Arrow heads are adjustable. in addition to boxes. • Lines and arrows have the same command. Here are some of the features of PSTricks that I shall illustrate. Many A of these are described in The L TEX Companion [5]. but we can change them by specifying:
\psset{unit=1in. pdflatex (not covered here. see [4]) does not work with PSTricks. themselves.. also provided free of charge. • Only one command is needed to put lines through a sequence of points. identifying any of a great variety of arrowheads simply. thereby eliminating the need for calculation or trial and error. some of which were listed above. • Circles and ovals.
same as \psline{}. then I shall show some additional shapes and commands. PSTricks extends the rectangle in \framebox by having a variety of shapes.2)
pspolygon(x0 . yn ) to (x0 . . y){r}
\pscircle(5. determined by a: .2)
psline{a}(x0 . y) with radius = r.0) \psline{*} (0. we write \pscircle[fillstyle=solid](0. y0 ).6. A parameter used by these commands is the distance
. yn )
\pspolygon(0. This is meant to be an introduction. y0 ) and opposite corner at (x1 . <> double arrow’ <. . use the fillstyle parameter.3)(6.3)
In using these commands. and objects that could be made solid. Draws closed polygon with given coordinates. I can put that circle right here: commands use the linewidth parameter to control the thickness of the lines used in the drawing. to produce a solid circle with radius . . Draws circle centered at (x.0)(10.2)
Draws rectangle with a corner at (x0 . the unit of measurement was set to 1 mm. y)(rx . (there are more!).1 cm. y1 ). > forward arrow. we do want the \rput command in order to put text into various objects. The User’s Guide [15] is freely available and clearly written. I shall illustrate the commands in Table 19 ﬁrst. .backward arrow. The origin is determined by where you are when issuing a pst command. For each command.0)(10. shown in Table 20. showing the ease and versatility of PSTricks.2)(1. For those examples.2 PSTricks
79
some of the common commands to draw objects and lines. so many features are not presented here. y) with horizontal radius = rx and vertical radius = ry Draws line or arrow. along path given by coordinates.0)(5. All no environment is entered.0)(5. (xn .0) \psline{<>} (0.
pscircle(x. Table 19: Some Basic Drawing Commands in PSTricks psframe(x0 . (xn .0){2}
psellipse(x. .no arrow. except ﬁgure is closed by drawing line from (xn . .1} (having already set fillcolor=gray).0){. y1 )
\psframe(0. centered at the origin. Thus. .1)(10. y0 )(x1 . Draws ellipse centered at (x. ry )
\psellipse(3. like boxes and circles. yn )
\psline{}(0. y0 ) .0) (0. For example. we can specify relevant options as [parameter = value]. y0 ) . The idea of a box is to have some shape enclose text.
where the default value of len is 3 pt.) The pst ﬁgures are drawn after specifying \psset{unit=1mm.fillcolor=white}. . we obtain this oval by writing: . (As usual. For example.height}{stuﬀ} Here are some examples: scales stuﬀ keeping the same aspect ratio scales the width and height individually
.80
6 GRAPHICS
between the border and the text inside. Boxes need not be enclosed (like \makebox). . we obtain \psovalbox{this oval}. 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. and they can be scaled by specifying one of the following: \scalebox{size}{stuﬀ} \scalebox{width. other parameters include linewidth. linecolor. . linestyle. . 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=. and fillcolor.
the \rput command puts a node. 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. In PSTricks. These can be connected by \psline. The \ncline command has the same arrow options as \psline.6. The name is set to A.2 PSTricks Halving the circle
\scalebox{. After entering the centering environment and setting the default units of measurement. with the \rnode command. the objects being joined can be referenced by name. by themselves and as enclosures for boxes.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. Knuth So far I have described a variety of shapes. including variations of arrowhead shape. 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. Knuth}
The source code is shown in Figure 64. each enclosed with a frame. The syntax for \rnode is:
\rnode{name}{stuﬀ }
The next two commands put nodes named B and C. Consider the following example: Node A Node B Node C
Who is the founder of \TeX? \rotatedown{Answer: Donald E. the named objects are called nodes. but with the following syntax:
\ncline{a}{name of node A}{name of node B }
. with a great variety of styles.
2.1){\rnode{C}{\psovalbox{Node C}}} \ncline[nodesepA=3pt]{A}{B} \ncline[nodesepA=5pt]{<}{A}{C} \ncline{<>}{B}{C} \end{center} \vspace{.82
6 GRAPHICS
The ﬁrst \ncline in Figure 64 draws a plain line from node A to node B.linearc=. or for both end points. like B and C. the line would touch Node A text.loopsize=. (nodesepA and nodesepB are keywords and have nothing to do with the names we assign to our nodes.2]{>}{3}{3} \Bput[5pt]{loop} % \Bput keeps label horizontal and 5pt is % the space added between label and arc \ncloop[angleA=180. is shown in ﬁgure 65. node separation can be speciﬁed for either end point.5.2.arm=. or nodesep=n. The [nodesepA=3pt] option gives 3 pt separation between the end of the line and node A.)
\begin{center} \psset{unit=1cm} % Nodes \cnodeput(2. Its source. which is not what we want. The default value is nodesep=0pt. which is what we want when the nodes are enclosed boxes.5in}
Figure 64: PSTricks Source for Connecting Nodes Figure 66 shows a graph that could represent any number of things.)
\begin{center} \psset{unit=1cm} \rput( 0. 0){\rnode{A}{Node A}} \rput(2. 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. (Try adding one line at a time and observe each eﬀect. 0){1}{1} \cnodeput(0. 1){3}{3} \cnodeput[doubleline=true]( 2. The separation is exaggerated to 5 pt in the arrow from node C to node A. 0){2}{2} \cnodeput( 2. respectively.arm=. In general.5.1){\rnode{B}{\psframebox{Node B}}} \rput( 2.2]{<}{4}{4} \Bput[5pt]{loop} \end{center} \vspace{1cm}
Figure 65: Graph Source (Result in Figure 66)
.linearc=.1){4}{4} \pnode(3. Otherwise. by specifying nodesepA=n.loopsize=. using PSTricks. nodesepB=n.
3)
⇒
Question: What is the pst command to draw the parabola given by y = ax2 + bx + c.4) \parabola{<>}(4.0)(4.1)(1. and (x1 .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. where a = 0?
b b Answer: \parabola(0.
2
\pscurve{(>}(0.
\psgrid[subgriddiv=1. − 4a + c) The following shows two commands: pscurve and psccurve. The examples that follow use the following pst settings:
\psset{unit=. y0 )(x1 .1)
\psccurve(0.3)(2.5cm.1)(1. y0 ) is one point on the parabola.1)(1.6. whose command syntax is:
\parabola{a}(x0 .griddots=10.0)(1.1)
. y1 ). For example.showpoints=true}
(The showpoints=true setting is what causes the points to be included in the picture you see. the latter being a closed curve that joins the last point with the ﬁrst.0)(1.showpoints=false](1.gridlabels=7pt](1.1)(2.1)(1.0) \parabola*[fillcolor=black. given points.
where (x0 . y1 ) is the (unique) point having dy/dx = 0. perhaps approximately.1)(1. c)(− 2a .) We begin with the parabola. \parabola* speciﬁes ﬁlling the parabola.1)(1.
6)(10. perhaps produced by mathematical software like gnuplot c . (Setting showpoints=false
.4)
We can read data from a ﬁle.8)(39.Oy=0.)
15 C 10 F 5 0 50 60 D 70 80 90 100 B A
score
\psset{unit=2mm.0) \rput(14. MATLAB c .0)(11. y2 )(x3 . which can be separated by a comma or just blank and can have parenthesis.0)(30.0)(1. The following histogram was plotted by the source code in Figure 67. The command syntax is:
\psbezier[parameters]{a}(x0 . 7){\textsf{A}}
Figure 67: Source Code for Drawing Histogram of Test Scores After setting the units of measurement to 2 mm.5)(50.84
6 GRAPHICS
The Bezier curve joins two end points and comes as close as possible to two intermediate points.5)(50. y0 )(x1 . Maple c . braces.3)(2.11)(29.0)(1. the data ﬁle is read and its points plotted with the \fileplot command.17) \rput[r](60. The data ﬁle just needs pairs of coordinates.2)(19.8)(39.2){\large score} \psline(1. The data ﬁle had y = number of students with test score = x + 50.11)(29.0) \rput(45. 4){\textsf{D}} \psline(20.0) \rput(35.dy=5.Dy=5.0)(40.dx=10.Dx=10. y3 )
\psbezier(0. 8){\textsf{F}} \psline(11.10){\textsf{B}} \psline(40. (The oﬀset of 50 was used in establishing the origin in the plot.13){\textsf{C}} \psline(30.2)(19.0)(20.0) \rput(25.dat} \psaxes[Ox=50. Mathematica c .0) \rput( 5. Octave c . showpoints=false} \fileplot[plotstyle=dots]{mydata. y1 )(x2 .1)(3.6)(10.ticks=y]{<>}(60. and SPLUS c . or nothing around each pair. which I shall explain.
90) [90.3) ⇒
\psaxes[unit=.5cm]{>}(0.0)(2. and there are 11 dot styles.) The data ﬁle is plain text and has the following entries: % This is mydata.65) [65.5cm]{>}(4. y0 )(x1 .1) ⇒
. y2 ) is the Northwest corner.70) [70.1. This is suppressed for the xaxis in Figure 67 by specifying the option. (x1 .dat} ⇒
Next.0)(1.2. ticks=y.80) [80. itself. axes are superimposed with the \psaxes command:
\psaxes[params]{a}(x0 . and (x2 . it is assumed to be equal to the origin. speciﬁed by plotstyle=dots. As in \psline. y1 ) is absent.6. y0 ) is the origin.1. 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. The other parameter settings are described in Table 21. If (x1 .)
\psaxes[unit=. dx=dy=0. is just the points. the origin is assumed to be at (0. y2 )
where (x0 .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. (The default values.100]
The plot. respectively. cause the spacing to be equal (approximately) by using Dx÷\psxunit and Dy÷\psyunit. 0).plotstyle=dots]{mydata.2 PSTricks
85
suppresses plotting the points in the \psline commands. Here is one of the alternatives:
\fileplot[dotstyle=+. y0 ) is absent. y1 ) is the Southeast corner. y1 )(x2 .0)(8. such as plotstyle=line. if (x0 . There are other plot styles.
ﬂush right (indicated by [r]) at the coordinates (60.) On unix. \psaxes and \rput.edu/octave/. The remaining commands draw the histogram boxes and put the letter grade above each box in sans serif font.che. or there is a unix conversion utility. including historical context. is gnuplot. which can produce eps ﬁles of plots. It comes with MiKTeX and basic unix installations. we can import it using the Graphics Bundle [3]. −2).
6. \ps2epsi. Once the ﬁle is in eps format. which converts jpeg ﬁles to eps. and those that plot mathematical functions or data. This is available free of charge at FTP://ftp. notably jpeg2ps. The unix systems xv and Image Magick can do this for a large variety of graphic ﬁle formats. and SPLUS c . but this does not exhaust the PSTricks commands.) Many systems that let us draw ﬁgures. we obtain the data plot. like Maple c . A basic plotting system for functions and data. Thus. Mathematica c . MATLAB c . you could use \psfig. which converts Windows Metaﬁles (wmf) to eps. There are two packages that provide
. Another way to obtain an eps ﬁle is with conversion.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. There are free conversion systems on MS Windows. (If you can get a ps ﬁle. rput[r](60. this is scaled (simply. provided free of charge. by specifying \psset{unit=1mm}): We shall stop here. See [15] for lots more.dartmouth. Octave extends the capabilities of gnuplot and is also available free of charge.3
Importing pictures
A The way to import a picture into L TEX is to convert it to encapsulated postscript (eps). have an option to export an eps ﬁle. Leaving oﬀ the “score.” Figure 68 shows the sequence of how each \psline and \rput adds to the picture. including bitmap (xbm). and the export options include the eps ﬁle format. gif and jpeg ﬁles. written by David P. is given by Keith Reckdahl [12]. Carlisle. There are also commercial systems.wisc. for both unix and DOS that produces eps ﬁles.edu/pub/gnuplot/. at http://www. (He also goes deeper into customizing placements of pictures in ﬁgures. when I superimpose the commands \fileplot. xfig is an excellent system to draw ﬁgures. including many examples.2){\large score} puts “score” in large font. An exceptionally clear description of this. To ﬁt the picture and the code next to it.
Oy=0.ticks=y]{<>}(60.8)(39.0) \rput(45.0) \rput(25. the other is graphicx.6.0) \rput(14.8){\textsf{F}} 50 60 70 80 90 100 \psaxes[Ox=50. dx=10.0)(30.
\fileplot[plotstyle=dots]{mydata.5)(50.8)(39.5)(50.2)(19.0)(1.0)(11.0)(40.13){\textsf{C}} \psline(11.Dy=5. as speciﬁed in the preamble by \usepackage{graphicx}. Figure 69 shows a ﬁgure that was imported with the following statement:
.10){\textsf{B}} 50 F D 60 70 C 80 90 100 \psline(20.7){\textsf{A}} 80 90 100 50 F D 60 70 C B 80 90 100 \psline(30.17)
60
70
80
90
100
Figure 68: Sequence of PSTricks Commands to Draw Histogram To include an eps ﬁle. For example.4){\textsf{D}} 50 F \psline(1.Dx=10. Here I use graphicx.3 Importing pictures
87
essentially the same capabilities but with diﬀerent syntax.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(40.11)(29.11)(29.0) \rput(35.0)(20. simply specify \includegraphcs[options]{ﬁlename}.dy=5.6)(10. One is called graphics.2)(19.6)(10.0) \rput(5.
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.eps} \end{center}
1 0. scale=. The height speciﬁcation (!) says to maintain the aspect ratio.height=!.6 0.4 0. read about the PSfrag package.6 0. but with the width and height set as
\begin{center} \includegraphics[width=2in. we might want to specify width=\textwidth.8 0.eps}\end{center}
1 0. by specifying the sin function over the indicated grid).4 −0.4 0. ﬁle.8 0.88
6 GRAPHICS
\begin{center}\includegraphics[scale=. Grant and
.2 0 −0.6 −0. and let it ﬁll the entire width of the page. by Michael C.2 0 −0.6 −0. which prints the ﬁgure (in this case by MATLAB.5.2 −0.2 −0.8 −1 −8
−6
−4
−2
0
2
4
6
8
Figure 70: Specifying Dimensions in \includegraphics For a very large picture.4 −0.5]{sin.height=1in]{sin. If you ﬁnd yourself importing eps ﬁles but would like to make some A changes in L TEX.
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. with eps ﬁles produced by MATLAB and xﬁg.6. rotation. The documentation gives examples. The box could contain text.e. It has two basic operations: (1) edit A some string or position in the ﬁgure (i. it is more eﬃcient to specify that option in the \includegraphics.. Although it is feasible to perform the operation after importing a graphic. or almost any stuﬀ. It can also perform scaling. whose documentation is at CTAN [4]. and (2) translate L TEX commands that you put in the ﬁgure in the ﬁrst place. Here are some examples:
. the eps ﬁle).3 Importing pictures
89
David Carlisle. Importing graphics is only one of the functions of graphicx. pictures. which comes with a basic installation (including MiKTeX). and sizing of an arbitrary box.
25\textwidth.eps}
\includegraphics[height=.5in.90
6 GRAPHICS
\includegraphics {protractor.
.origin=c]{protractor. Submit a printed copy of the L TEX source (tex ﬁle) and printed
© vm 4'
vm 3
3. Use the picture environment to draw the smiley face on page 73. height=!]{protractor.eps}
\includegraphics[width=. 4). 2. angle=90. 1. where \thicklines is speciﬁed and \unitlength = 1mm vm 1
E vm 2 c
A Exercises. Use PSTricks to draw Figure 3 (p.width=!. Draw the following graph with the picture environment. Be sure your name is on each.eps}
copy of the associated postscript result (ps ﬁle).
91
rhombus
5. but this section did not describe all that is needed. (They were drawn here with PSTricks. Use whatever means you prefer (or that your instructor requires) to include each of the following ﬁgures in your document.6. include it in your document.
β α
6. Use PSTricks or the picture environment to draw the following. 7.) (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.3 Importing pictures 4. Then. Make a ﬁgure in some system that lets you save it as an eps ﬁle (or use some conversion program). so you must obtain the PSTricks User’s Guide [15].
(35. 20) (25. (35. 10)130) 50) (15.92 (b) Network with arc data: 2 4
6 GRAPHICS
(15. 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)
. 40) (45. 30) 50) (45.
You might also need to initialize the page counter with \setcounter{page}{0}. If you see a page number on your cover page. right after \maketitle. The necessary parameters are \author and \title. This is usually done just following \begin{document}. such as in the example shown in Figures 71 and 72. the following does this while specifying 12pt font as another option:
. it depends upon your management style. followed immediately by \maketitle. suppress this by adding \thispagestyle{empty}. Multiple authors are separated by \and.) Specifying \date is optional (\maketitle puts in the current date if the date is not deﬁned). (The jagged edges in Figure 72 mean that there is more space between the title and the top of the paper. which can be deﬁned anyplace before the \maketitle. For example. The cover page is by itself and is not numbered.93
7
7. Typically these are put into the preamble. titlepage must be speciﬁed as an option in the \documentclass command. or right after \begin{document}.1
Making Special Parts
Cover Page
The easiest way to make a cover page is with the \maketitle command.
\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).
For example. and other information about each author can be added.titlepage]{article}
Addresses. This is because the added width of author information makes it too long to ﬁt on one line. Figure 73 shows how the authors appear when the \author deﬁnition in Figure 71 is changed to the following:
\author{Michel Goosens \\ 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. All three authors would be put on separate lines if the address information were extended further. using \\ to create new lines. using diﬀerent footnote marks for each one. \maketitle puts the third author on a separate line. aﬃliations.94
7 MAKING SPECIAL PARTS
\documentclass[12pt. Switzerland
Frank Mittelbach Mainz. or if the names were very long. Germany \and Alexander Samarin \\ Geneva Switzerland}
As illustrated in Figure 73. The \thanks command does this by creating a footnote. Switzerland \and Frank Mittelbach \\ Mainz. Germany
Alexander Samarin Geneva. Figures 74 and 75 illustrate this along with some variation in the date.
A The L TEX Companion
Michel Goosens Geneva.
} \LaTeX} Rich \thanks{Smart University} Grand \thanks{StreetSmart} All \addtocounter{footnote}{2} \footnotemark
Figure 76: Authors with same footnote (Result in Figure 77)
. Prior to that.7.M. Ionia } \date{210 {\sc bc} (revision of earlier version. the value of the footnote counter is set back to the ﬁrst footnote mark.
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.}\\ Syracuse. which puts the footnote mark without any new text. \and U. Sicily \and Pythagoras \\ Samos. Ionia
210 bc (revision of earlier version. This uses the \footnotemark command.
\title{Doing \author{I. 510 bc)
1 Renamed.R.}} \index{\texttt{$\backslash$thanks}} \index{footnote} \author{Archimedes\thanks{Supported by the army. 2 Supported
by the army.C. Sicily Pythagoras Samos.1 Cover Page
\title{Pieces of $\pi$\thanks{Renamed. \and I. 510 {\sc bc})}
95
Figure 74: Footnotes in the Cover Page Source (Result in Figure 75)
Pieces of π 1
Archimedes2 Syracuse.
\end{abstract}
Figure 78: Making an Abstract Source (Result in Figure 79)
.C. Grand2
I. itself.)
\begin{abstract} This shows that the ratio of the circumference to the diameter of any circle is the same constant value.2
Abstract
The abstract environment is in all document styles. All1
1 Smart
University
2 StreetSmart
Figure 77: Authors with same footnote (Source in Figure 76)
7. Figures 78 and 79 illustrate this. Rich1
U. with the header: Abstract. The abstract. specify titlepage as an option in \documentclass (even if you do not intend to use \maketitle).R. the abstract is placed far from the top of the paper. This environment is deﬁned to produce an abstract on a separate page (placed wherever you put the environment speciﬁcation). except article. in boldface and centered. We further prove that this constant is bounded by $\frac{223}{71} < \pi < \frac{22}{7}$. To have it.96
7 MAKING SPECIAL PARTS
A Doing L TEX
I. (Like the cover page.M. which is not shown in Figure 79. denoted $\pi$. is one paragraph and is printed without indentation.
L TEX provides the \addcontentsline command. which is on page v. followed by the list of tables. denoted π.
\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. For example. Just above each declaration. That is why you see the Table of Contents on page i (ﬁrst numbered page. I declare \listoffigures. it is placed wherever you put the command. 71 7 Figure 79: Making an Abstract Result (Source in Figure 78)
7. you can include lists of ﬁgures and tables with the \listoffigures and \listoftables commands.7. respectively. which should be right after the cover page.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. We further prove that this constant is bounded by 223 < π < 22 . Then.3
Other Front Matter
The \tableofcontents command makes a table of contents. Each of these are put on a new page. like sections and A subsections. To include other front matter. just after the cover). indicated by the toc speciﬁcation.
. Then. I use the \addcontentsline to add it to the table of contents. 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. The table of contents generally includes numbered parts. the table of contents in this document was obtained with the speciﬁcations given in Figure 80. The section parameter tells the latex program to format it like a section — ﬂush left.
The only change we require is another header name. . After a successful compilation. but we do not want it to have a section number. Put \printindex just before \end{document}. This is achieved by the \section* command. . The last portion in the back of any book is its index. but with no number (and the section counter remains unchanged). 3. Put \makeindex at the end of the preamble. you use \index{entry}. Put \usepackage{makeidx} in the preamble.
. with all references resolved. \section*{Preface} puts “Preface” in the same style as any section. For now. To have entries in the index. For example. 2. §?? has more to say about using this command to customize many things. where the * suppresses the numbering. This is done by redeﬁning the \abstractname parameter used by the abstract environment.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. \end{appendix}. For example. the index in this book contains several occurrences of ‘ﬂushleft’. 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. we put the bibliography (see §3 and §8. we have three things to put into our source ﬁle: 1. The \renewcommand enables us to do this. This could be done with the appendix environment: \begin{appendix} . I also thank the contributors to the Comprehensive \TeX\ Archive Network (CTAN). We might ﬁrst want to have appendices that follow the main text. compile again. This is analogous to the use of bibtex (p.6). \end{abstract}
Alternatively. and is illustrated in Figure 81. we might want something to look like a section (and automatically added to the table of contents).4
Back Matter
After the main part of the document is ﬁnished.
7. This could also be desirable in a long report. In the text. enter at the command line: makeindex myfile Then. 28). To make an index.
as shown in ﬁgures 84 and 85.dvi dvipdf makeindex
print/post
myfile.
1 2
ﬁrst second
Figure 83: Setting a Footnote Result (Source in Figure 82) \thefootnote gives the counter for footnotes.
create/edit compile latex myfile.
. More generally. here is my second\footnote{second}. You can also change to common footnote symbols by specifying \renewcommand{\thefootnote}{\fnsymbol{footnote}}. I specify \index{flushleft}. To put the entry as a subordinate. nomencl makes a list of nomenclature.ps myfile. Figures 82 and 83 illustrate. and gloss makes a glossary.pdf
Figure 81: Adding makeindex to the Command Sequence There are packages to make other back matter: acronym makes a list of acronyms.
Figure 82: Setting a Footnote Source (Result in Figure 83) Here is my ﬁrst footnote1 . use !. For example. and you can change from numbers to letters by \renewcommand{\thefootnote}{\alph{footnote}}.5
Footnotes
The \thanks command is one way to have a footnote on the cover page.7.tex
view/print convert dvips myfile. which can be controlled by \setlength.
7. and was shown on page 95.5 Footnotes
99
at each occurrence. \index{package!makeidx} puts the entry ‘makeidx’ under the entry ’package’. here is my second2 . The distance between the line that underlies part of the last line of text to the footnote is \footnotesep. the \footnote command can be used anywhere.
Here is my first footnote\footnote{first}.
Be sure your name is on each. but also the abstract.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 }. acknowledgments and references. Here are two examples simply to abbreviate commands with long names:
\newcommand{\ul}{\underline} \newcommand{\mc}{\multicolumn}
. here is my second\footnote{second}.
8. which enable you to deﬁne your own commands and change parameter values of existing commands.
8
Taking Control
This section introduces you to fundamentals of customizing your document. A key to these changes are the \newcommand and \renewcommand commands. and whatever is whatever you want the command to do.
A Exercises. Submit a printed copy of both the L TEX source (tex ﬁle) and
1. 2. It is still in the context of an introduction.100
8 TAKING CONTROL
\renewcommand{\thefootnote}{\fnsymbol{footnote}} Here is my first footnote\footnote{first}.
∗ †
ﬁrst second
Figure 85: Setting a Footnote Result (Source in Figure 84)
the associated postscript result (ps ﬁle).
Figure 84: Setting a Footnote Source (Result in Figure 85) Here is my ﬁrst footnote∗ . and Conclusions. here is my second† . Extend exercise 1 to have acknowledgments and references (using BibTE X). Combine exercises 1 and 2 and add a table of contents showing not only all sections and subsections. Write an article with a title page and abstract. Make the main body have at least three sections: Introduction. Main Results. where n is the number of arguments. choosing only a few of the things you can change. 3.
particularly of notation.0) \put( 2. no matter which mode we are in.7)} \end{picture} }}
(\mbox is used to ensure text mode).. Then. we can use it in other new commands. it is replaced by $G_{i+j}$ to put it into math mode ﬁrst. so it is useful to put them in a separate ﬁle. in the preamble I speciﬁed: \newcommand{\chkbox}{$\Box^\surd\. like the null space of a matrix.. Now \Box ⇒ and.. and obtain the correct result. then add to the preamble:
\newcommand{\usenul}{\mynul}
and specify \usenul in the document. simply change the one line to:
\newcommand{\usenul}{\nul}
(where the publisher’s command name is \nul). For example. A way to do this is to choose a diﬀerent name. we can specify \Gs{subscript}. having deﬁned the \Box command. If we are already in math mode. We can choose one and deﬁne \newcommand{\nul}{\ensuremath{\mathcal{N}}}. The latex compiler will not let you use a name that is already being used. Suppose we have a key term. but we want them to work in any mode. Some publishers have their own notation. If you need to use the publisher’s. in either a tabular or an array environment.8. √ \chkbox ⇒ Some commands are speciﬁcally for math mode.$} Then. For example. if \Gs{i+j} is speciﬁed in text mode. A related use is when the command requires some lines of code. if you specify \newcommand{\fbox}. This is achieved by the command: \ensuremath{math stuﬀ }. Then. Another reason to have our own commands is for consistency. Thus. say mydefs. and there are still more symbols people use. consider \newcommand{\Gs}[1]{G_{#1}}. so we must be careful not to override them with ours. The second lets us write \mc{3}{c}{stuff} to enter a multicolumn. we use the \input command to have the latex compiler read
. like mynul. Consider the following example:
\newcommand{\Box}{\mbox{\begin{picture}(0. Some authors write N (A).tex (note the . The preamble can become very long as we add our commands.0){\framebox(7.1 Your Own Abbreviations and Commands
101
The ﬁrst lets us write \ul{something} to underline something. spanning 3 columns and centered. you will get a fatal error message since there is already a \fbox command. otherwise. we can write \nul(A) to obtain N (A) (and we can be in text or math mode when we write this). For example. \Gs{i+j} is replaced by G_{i+j} to produce Gi+j .tex suﬃx). some write nul A.
43). the preamble of this document contains the command:
\input{mydefs}
(The suﬃx . in this document the “Table of Contents” was obtained by specifying the following in the preamble:
\renewcommand{\contentsname}{Table of Contents}. so duplication of work is avoided. it will number the ﬁrst section as 0. Making the document class an article will also solve the problem. The general form is
\renewcommand{\thecounter }{something}
Another example is to change section numbering in a report document style. Table 22 shows the common names we might want to change. Making the ﬁrst level division chapters will overcome the numbering problem. so the numbering will be chapter. The ﬁrst level of division is assumed to be a chapter.2
Your Own Names.subsection]. This might not be appropriate due to other considerations. If you have no chapters.
8.102
8 TAKING CONTROL
it wherever it is placed. We saw an example of this in changing the counters for enumeration lists (p. . since the section is the ﬁrst level division.) Diﬀerent source ﬁles could simply input this same ﬁle. Titles and Numbers
There are times when we prefer some name other than the default. similar to a book. .1. but the format of chapters is diﬀerent. such as entering the document into a database using
. For example.
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.section[. In particular.tex is assumed.
4
Your Own Margins and Spacing
The default margins and spacing are set with purposeful values.
8.1mm}{2. the theorem follows. is to make the @ character a letter.8. and end is what is executed upon leaving the environment. restores @ to its special meaning (\@ is for certain spacing. suppose . \end{proof} produces: Proof: First. where you want it to be counted as a report.3 Your Own Environments
103
BibTE X. not as an article. . \begin{proof} First. The succeeding command. and the \renewenvironment command enables us to revise an existing environment.
8.1mm} \end{description}\end{flushleft} } % end proof Then. and you will usually not need to change them. The margins
. \makeatother. equal to about 2 spaces). Thus. When you do. For example. begin is what is executed upon entering the environment.3
Your Own Environments
The \newenvironment command enables us to deﬁne our own environments. however. the following creates a proof environment: \newenvironment{proof} {\begin{flushleft} \begin{description} \item \textit{\textbf{Proof:}}~ } % begin proof {\hfill\rule{2. the theorem follows. . n = number of arguments (omit [0] for n = 0). \makeatletter. suppose \dots \linebreak Thus. They have the same syntax:
newenvironment{name}[n]{begin}{end } renewenvironment{name}[n]{begin}{end }
where name is the name of the environment. The way to do this is as follows:
\makeatletter \renewcommand\thesection{\@arabic\c@section} \makeatother
The preceding command. they can be changed by setting certain parameters in the preamble.
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 twosided. The geometry package provides easy speciﬁcations for page layout. Margin settings can be negative. For example. in particular.0pt 29. (See Table 25 for conversion factors. the current settings (shown 2 in Table 23) break down the horizontal parts as follows:
72.295pt \textheight 585. The body expands to the right unless we also change \oddsidemargin.96999pt \paperwidth 614. The \hspace* and \vspace* commands provide a great deal of control over horizontal and vertical spacing.295 pt 390 pt 5.0pt \topmargin 21. showing their default values (used in this document).0pt 12.14. 1 pt = 72.38744pt \textwidth 360. Table 23: Margin Parameters Parameter \footskip \headsep \headheight \hoffset \oddsidemargin Current Setting† 30.5 in
end body stuﬀ
113. This might be accompanied by increasing the text length. respectively.0pt \voffset 0. \textwidth=6in increases the text width to 6 inches.) For example.025 pt 1.0pt 0.0pt 25. 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. Table 24 lists some you can set with the \setlength command.27 pt 39 pt 1 in 54 in
begin body stuﬀ
6. We might want some global settings to make repeated use of these unnecessary.564 in
We can increase the text width by setting \textwidth=length in the preamble. if we are using 8 1 × 11 paper.104
8 TAKING CONTROL
of the document are controlled by the parameters shown in Figure 86 and described in Table 23.0pt † Printed using \theparameter. we raise the body 1 inch by specifying \topmargin=1in in the preamble.396 in 8.27 in.
. for example.
8.4 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
.
This item is fairly close to the ﬁrst one. This list speciﬁed \setlength{\itemsep}{12pt} right after \begin{itemize} 2.0pt. (Defaults are restored after leaving. and I have saved it by: \setlength{\mylength}{\itemsep}. This item is farther from the ﬁrst one. • What you see next is with \setlength{\itemsep}{10pt}.
1. What you see next is with \setlength{\itemsep}{0pt}. the lists in §2.2 (p. varying by the value of \itemsep. This must be set before the \begin command. For \itemsep this can occur just after the \begin statement. This list speciﬁed \setlength{\itemsep}{0pt} right after \begin{itemize} 2. • Next is back to normal by \setlength{\itemsep}{\mylength}. indentation at beginning of paragraph.
. • We are back to normal with \itemsep = 4.5pt plus 2.) For example. they must be set within the list environment. 12) are spaced by default values.106
8 TAKING CONTROL
Table 24: Spacing Parameters Parameter
\itemsep \parindent \parsep \parskip
Meaning space added to \parsep between items in a list. • See the above spacing between items.
Figure 87: Varying \itemsep to control item spacing in a list The left margin parameter is \leftmargini for a level1 list (note the i at the end).0pt minus 1.5pt plus 2. space between paragraphs. Here is what happens when we change \itemsep: • The default value of \itemsep is 4. Figure 87 shows two enumerate lists. Figure 88 gives an example. 1.0pt. space between paragraphs in the same item of a list.
In the case of list parameters.0pt minus 1.
3in}l} \.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 pcolumn to put horizontal space between the other two columns.6 line spacing: \renewcommand{\arraystretch}{1.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. Note how congested it is.3 line spacing: \renewcommand{\arraystretch}{1.B\.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 1.6} B xB = bN + 1 θδbN 2 B ∗ xB > bB + 1 θδbB 2
1 πN B = cB + 2 θδcB 1 πN N < cN + 2 θδcN
. so we want to increase its vertical spacing.
\[ \begin{array}{lp{.8.
followed by your second argument. .
How much is 2+2? \printsol{\sol}{4}
Figure 92: \ifthenelse Source (Results in Figure 93) \def\sol{0} . If so. Suppose that you want no page numbering if there is only one page. just before \end{document}. . If there is more than one page. it will do nothing. it will print Solution.
. put \label{lastpage}. }#2}}{} }
This will test the condition if your ﬁrst argument equals 1.108
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 . 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. How much is 2+2? Solution. This is illustrated in ﬁgures 92 and 93. 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. and the ﬁrst statement applies: \pagestyle{empty}. Just after \begin{document}. . there is only one page. How much is 2+2? \def\sol{1} . The syntax is \ifthenelse{condition}{action if true}{action if false} For example. if not. You can do this as follows. If not. You will have control of the whole document. there is more than one page. 4
Figure 93: \ifthenelse Results (Source in Figure 92) Another example is to decide on page numbering. suppose you want to use headings in the header with page numbers. and the second statement applies: \pagestyle{headings}.
8. suppose your preamble contains
\newcommand{ \printsol}[2]{ \ifthenelse{\equal{#1}{1}}% {\textsf{\textbf{Solution. If so. First. . or just some lines whose space you want to control.
1989. . . MA. Frank Mittelbach and Alexander Samarin. The bib style ﬁle. When citing Knuth’s book. specify the following:
\begin{thebibliography}{n} \bibitem[what appears]{label (that you cite)} entry . so each entry appears however you put it. (It works if you specify 99.e. we can exercise the option:
\bibitem[Knuth. This is one reason it is usually better to use BibTE X. MA. \textit{The \TeX\ Book}.) Each \bibitem is an entry.6 Your Own Bibliography
109
8. and use thebibliography environment instead. \end{thebibliography}
These will appear in the document’s list of references even if they are not cited. and you can enter whatever you like. They can be cited in the same way described in §3: by \cite{companion} and \cite{tex}. \bibitem{tex} Donald E. 1989]{tex} Donald E. With you in control. even though you lose control over what appears (i. 28). respectively. Knuth. we obtain [2] in the text. You will have complete control over the formatting. AddisonWesley Publishing Company. with label the unique identiﬁer used by the \cite command. Reading..8. . Knuth. even if there are inconsistencies in style. it generally helps the reader to know something about the citation. and there will be no sorting — the list of references will appear in the order you put them. Here is a complete example with two references.
in which case \cite{tex} ⇒ [Knuth.6
Your Own Bibliography
You can choose not to use BibTE X. 1994. which I formatted to agree with the plain style of BibTE X:
\begin{thebibliography}{99} \bibitem{companion} Michel Goosens. Reading. such as
. \end{thebibliography}
where n is the width of the widest label you want to allow. 1989] is preferred to [2] because it immediately gives the reader information about the document without having to ﬂip to the bibliography section. for example. [Knuth. 1989]. but if you are writing a report and have control over the formatting. The option is an alternative to having the references numbered. \textit{The \LaTeX\ Companion}. they will be numbers). Thus. Alternatively. 15th edition. there is no format monitoring. as described for BibTE X in §3 (p. . AddisonWesley Publishing Company. Instead of the BibTE X commands. Some publishers give you no choice. . \bibliography{mybiblio} and \bibliographystyle{plain}.
As you begin to use packages. SIunits.text. Most installations come with more than the basic plain. including those commonly found in automata theory. If you want to have several bibliographic units in one document. cpascal and listings). and xypic. Here are some packages that give you special symbols: chemsym. or you can join their newsgroup at comp. qsymbols. Figure 94 shows the preamble used for this document. 36). which you obtain from CTAN [4]. and its three variations (given on p. it is necessary to become aware of updates. such as [Knuth.. applies in either case. These packages provide even more versatility in how the citations appear (see [4] or [5. the algorithm package enables an environment to write source code with standard language elements. which give the author and year. Also.
. You can visit CTAN from time to time. and there are others with similar properties or for particular programming languages (viz. depending upon your technical area. 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. Chapter 13]). 1989] instead of [8].tex (can be done in Google by clicking on Groups).bst. available from CTAN [4].110
8 TAKING CONTROL
plain. Alternative bst ﬁles are achicago (from the frankenstein package). Besides what you can do yourself to elevate the quality of the results. wasysym. apalike and plainnat (from natbib). such as at the end of each chapter of a book. there are many packages. The graphtex package specializes in all sorts of graphs. use the bibunits package. You will ﬁnd other packages useful.
Afterwards.}} ifthenelse {condition}{true}{false} index more math symbols (viz. beginning with Table 25. which gives conversion among three common units of measurement.54 1
Table 26 is a guide to how most of the remaining tables are organized.to write \textbf{\textsc{...45 in .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.03515 2.pstall} \usepackage{hyphenat} \usepackage{fancyvrb. Table 45 gives special symbols that can be used in either text or math mode.APPENDIX
111
\usepackage{amsmath} \usepackage{amssymb} \usepackage{amscd} \usepackage{bm} \usepackage{graphicx. It is designed like a reference manual for easy lookup.27 28..
.. \mathscr) like multicolumn enables more control over newtheorem \url{.
Table 25: Conversions of Common Units of Measurement pt in cm pt 1 72. and Table 46 gives the commands for the picture environment.3937 cm ..01384 1 .}
\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.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 ..
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. 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
.
å Å ¿ \c{c} \d{d} \b{b} \dots \aa \AA ?‘ x ˙ ¯ z v ˇ ø Ø ¡ \..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 ¯ .
rather than j. and \jmath. rather than i.y x\. Z {\mathcal \theta \vartheta \iota \kappa \lambda \mu \nu$ \xi \Lambda \Xi \Pi \ell A. to avoid the clash between the accent and dot.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
..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.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 .)
Table 35: Spacing Commands in Math Mode
What you write x 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.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. \bigcirc \ddots \bullet \smile \diamond \triangleleft \bigtriangledown \perp \bot \vdash \ \parallel
⊥ ⊥
∠ 
..
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 coauthoring. 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 override (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 styleoption 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.latexproject.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/texarchive/; Germany: ftp.dante.de/texarchive/; 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. AddisonWesley 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.belllabs.com/who/hobby/ MetaPost.html/.
MA. Reading.0. 1993–98.
[11] Oren Patashnik.ps (see [4] for replacing CTAN). New Jersey. 1998–99 (version 1. 1995–99. 1994). 15th edition. Technical report. 1989. CTAN. World Wide Web.
.org/applications/PSTricks/. AddisonWesley Publishing Company. http://www. AddisonWesley Publishing Company. [8] Donald E.miktex. CTAN /info/epslatex.html. World Wide Web. [15] Timothy Van Zandt. http://www.ps (see [4] for replacing CTAN). Hobby. http://www. MA.
A [9] Leslie Lamport.edu/ depts/adn/infwww/ps/btxdoc.ps.uic. AT&T Bell Laboratories. Using imported graphics in L TEX 2ε .2). CTAN /macros/latex/doc/usrguide. A [10] L TEX 2ε for authors. L TEX: A Document Preparation System. TE X Catalogue Online. MiKTE X Local Guide.gz. //www. PSTricks: PostScript macros for Generic TeX.
http:
[14] Graham Williams. 1988. Available at http://cm. Reading. Comprehensive TEX Archive. World Wide Web site Version 2. Computing Science Technical Report no. 1993.ctan.de/.
[13] Christian Schenk. 1995–97.ps. Drawing Graphs with MetaPost. Murray Hill. BibTE Xing.belllabs.tug.com/cs/cstr/164.org/texarchive/help/Catalogue/catalogue. The TEX Book. 164.120
REFERENCES
[7] John D. Knuth. World Wide Web. 1986 (also see 2nd edition.
A [12] Keith Reckdahl.
93 \mathbb. 63 \addtocounter. 25 \hspace. 51. 59 \mathscr. 104 \overbrace. 42 \fboxsep. 24 \pagenumbering. 97 \listoftables. 40. 47 \mbox. 41 \newtheorem. 47 \cdots. 97 \pageref. 118 \dvipdf. 61. 36 \bigg. 74 \listoffigures. 25 \dots. 71. 56. 41 \cline. 36 \bibliography. 99 \dvips. 43 \arraystretch. 93 \dotfill. 74 \frame. 24. 107 \author. 16 \dashbox. 55 \documentclass. 11 \multicolumn.Index
\Bigg. 24 \noindent. 48. 62
. 9 \hfill. 37 \nohyphens. 58 \medskip. 72 \makeindex. 53 \left. 93 \def. 55. 74 \footnotesep. 24 \nopagebreak. 118 \newcounter. 63 \bigskip. 62 \nocite. 60 \framebox. 100. 28. 55 \baselinestretch. 117 \itemsep. 101. 3 \ensuremath. 25. 65 \pagebreak. 63 \boldmath. 25 \imath. 36. 101 \newcommand. 42 \fbox. 114 \input. 24 \newpage. 71. 41. 99 \frac. 11 \big. 103 \newline. 37 \clearpage. 117 \linebreak. 118 \displaystyle. 98 \maketitle. 97 \makebox. 56 \overline. 23 \line. 1. 4. 23 \label. 25. 93 \baselineskip. 114 \kill. 8 \dvi2tty. 106 \jmath. 25 121 \hspace*. 10 \nolinebreak. 23. 15. 63 \Big. 56 \overset. 7. 12. 24 \oddsidemargin. 108 \fboxrule. 56. 10. 25 \hline. 59 \mathfont. 39. 9. 19. 72 \date. 61 \centerline. 42 \newenvironment. 101 \equal. 57. 19 \cite. 54 \hrulefill. 108 \bibliographystyle. 48.
55 \widehat. 97 \textstyle. 42. 3. 5 \substack. 8 \thanks. 1 conditional assignment. 1 DOS. 53. 25. 3 dvips. 43 \title. 98 \section. 11 \sqrt. 25. 72. 106 \parskip. 25 \vspace. 16 command line. 62 \renewcommand. 48 \stackrel. 99 \smallskip. 84 bibtex program. 4 emftoeps. 5 \selectfont. 65
INDEX \unitlength. 56 counting words. 9. 39. 95 \thecounter. 33 \usepackage. 56. 56 \underline. 1 comment. 94 \theenumi. 119 \psset. 77 \value. 18. 57 \raisebox. 34. 98 \prod. 65 \tableofcontents. 48 \providecommand. 24 \section*. 60 \prime. 43 \subsection. 56 \\. 43. 54 \printindex. 25. 43 \ref. 41. 65 \stepcounter. 16. 118 cover page. 1 amsmath. 10 debugging. 15 pdftex. 118 column speciﬁcation. 108 \parbox. 43 \setlength. 118 compile. 74 \verb. 38 \setcounter. 60 document styles. 43 \thefootnote. 72 \parindent. 98. 106 \partial. 47 box. 20 comments. 21. 58. 24. 19. 59.122 \pagestyle. 56 \underset. 1. 86
. 78 \raggedright. 48 \textwidth. 93 accents. 97. 72 coauthoring. 104 \textfont. 72. 117 \samepage. 103 \right. 100. 2 derivative. 57 \refstepcounter. 106 \parsep. 39. 119 \renewenvironment. 61 Bezier curve. 28 dvi viewer. 43 \vector. 35 dash. 78 titlepage. 93 cross referencing. 56 \widetilde. 40. 118 \vfill. 1 boldmath. 22 AMS. 28 body. 73 \url. 25 \vspace*. 61. 93 \underbrace.
6 corollary. 62 description. 8 boldmath. 9 smallmatrix. 34–36 bst (bib style). 15. 86 tex. 95. 47 italic. 21 document. 9 longtable. 38 list environment. 109 theorem. 16 math mode. 3. 6 gather. 81 Laplace transform. 43 graph. 42
123
. 4. 32. 65 tabbing. 8. 40 ﬂoat page. 14 itemize. 8 small caps. 8. 40 font size. 47. 41. 54. 54 ﬁgure. 22 Roman. 8 slanted. 12. 93 enumerate. 18. 16 sans serif. 98 jpeg2ps. 8 footnote. 22 table. 8 underlined. 16 typewriter. 57 ﬂushright. 1. 110 dvi. 72 quotation. 58 nonEnglish. 78 eps.INDEX environment. 41 ﬂushleft. 25 bibliography. 64 itemize. 9 boldface. 20 picture. 96. 82 Hamiltonian. 54. 99 fractions. 1. 62 center. 1 line spacing. 40. 3 global setting. 98 appendix. 64. 86 Lagrangian. 24 index (making of). 56 thebibliography. 57 axiom. 86 jpeg. 12 description. 10 hyphenation. 13 large. 13 local setting. 40. 9. 59 latex command. 61 verbatim. 78. 12 enumerate. 6. 11. 59 landscape. 86 ﬂoat. 55 font style bold small caps. 1. 6 abstract. 36. 51. 53 equation. 1. 60 horizontal ﬁll. 64 gather*. 98 array. 32 wmf. 53 eqnarray*. 19. 21. 47 calligraphic. 28. 10 quote. 11 ﬁle bib. 14 eqnarray. 48 ghostview. 21 verse. 29. 41 tabular. 41 ﬂoating object. 47 Greek. 86 ps (postscript).
5 superscript. 61. . 104 ~. 54 trigonometric functions. 45. 81. . 25 list. 24. 41 fontenc. . 78 MiKTeX. 78. 22 bibunits. 33 wasysym. 1. 15 ticks. . 29. ix. 58 stacking. 20 makeidx. 77. 110 hyphenat. 11. 79. 6 preamble. 65 subscript. 1 spacing. 22 table. 99 psfrag. 63 url. 46 tabbing commands. 110 bm.124 make index. 5 SIAM. 1. 81
INDEX
section. 111
. 110 fancyvrb. 85 transpose. 12. 103 \textspec . 48. 110 page numbering. 52 mathbf. 45. 106 math mode. . 3. 37 Underfull . 20. 4. 108 listings. 87 graphtex. 77 qsymbols. 16. 110 xypic. 86 nodes. 38. 10 rotate. 2 warning. 84. 108 showkeys. 21 in url. 110 chemsym. 11. 107 vertical. 59 moreverb. 58 units of measurement. 110 quotation marks. 98 math display mode. 110 geometry. 110 longtable. 48 cpascal. 40 SIunits. . 103. 99 algorithm. 110 theorem. 4. 110 setspace. 65 amsmath. 24 horizontal. 81 package. 65 subsection. 64 amssymb. 25. 47 matrix equation. 88 pstricks. 24 ifthen. 2 Repeated entry. 58 babel. 2 MetaPost. 108 paragraph positions. 12. 98 mathrsfs. 9 frankenstein. 1. 99 graphicx. 25 line. 46 stack. 33 special function. 104 gloss. 15. 111 ﬂoat. 25 special character. 50. 110 amscd. 48. 21. 8 acronym. 54 message. 72. 111 nomencl. 2 Overfull . 35.
118 xdvi. 28. 3. 86 wc. 118 word count. 3 YAP. 3
125
.INDEX unix.