Writing Latex

Writing LATEX
Some do’s and don’ts
Jendrik Stelzner
September 15, 2019

Version: 0.5 Commit: b86684a
Available at https://github.com/cionx/writing-latex.
Please send comments and corrections to stelzner@uni-bonn.de.
Preface
Das Auge liest mit.
German saying
This text presents some thoughts about what (not) to do in mathematical typesetting
using LATEX. Examples are provided using the environment tcblisting from the
package tcolorbox. This text is not intended as a first introduction to LATEX. Some of
the following remarks are subject to individual taste.
ii
Contents
Preface ii
List of tables vii
List of figures viii
1 Before you start writing 1

1.1 Use version control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 The compilation process . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2.1 Use LuaLATEXor XƎLATEX . . . . . . . . . . . . . . . . . . . . . 1
1.2.2 Don’t ignore warnings . . . . . . . . . . . . . . . . . . . . . . . 1
1.3 Code quality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3.1 Make use of whitespace to organize the code . . . . . . . . . . 2
1.3.2 Write one sentence per line . . . . . . . . . . . . . . . . . . . . 3
1.3.3 Properly indent your source code . . . . . . . . . . . . . . . . . 3
1.3.4 Use % for problematic line breaks . . . . . . . . . . . . . . . . . 3
1.4 Splitting up a project into multiple files . . . . . . . . . . . . . . . . . 4
1.4.1 The commands \usepackage, \include, \input . . . . . . . . . . 4
1.4.2 Use \includeonly . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.5 Choosing the document class . . . . . . . . . . . . . . . . . . . . . . . 8
1.5.1 Use the KOMA-Script classes . . . . . . . . . . . . . . . . . . . 8
1.5.2 Deciding on a document class . . . . . . . . . . . . . . . . . . . 8
1.6 Use \frontmatter and its friends . . . . . . . . . . . . . . . . . . . . . 10
1.7 Be consistent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2 Useful and important packages 12

2.1 microtype for better typesetting . . . . . . . . . . . . . . . . . . . . . . 12
2.2 mathtools for mathematics . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.3 amsthm for theorem-like environments . . . . . . . . . . . . . . . . . . . 12
2.4 tikz-cd for commutative diagrams . . . . . . . . . . . . . . . . . . . . 15
2.5 cleveref and hyperref for referencing . . . . . . . . . . . . . . . . . . 16
2.5.1 Recalling basic referencing . . . . . . . . . . . . . . . . . . . . . 16
2.5.2 cleveref . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.5.3 hyperref . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.5.4 Order of inclusion . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.6 csquotes for quotation marks . . . . . . . . . . . . . . . . . . . . . . . 20
2.7 enumitem for configuration of lists . . . . . . . . . . . . . . . . . . . . . 22
2.7.1 Style of numbering . . . . . . . . . . . . . . . . . . . . . . . . . 22
iii
Contents
2.7.2 Resuming lists . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

2.7.3 Spacing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.7.4 Global settings . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
2.7.5 Nested lists . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.7.6 Cloning lists . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
2.8 biblatex for bibliography . . . . . . . . . . . . . . . . . . . . . . . . . 33
2.8.1 The basic setup . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
2.8.2 Creating the bib-file . . . . . . . . . . . . . . . . . . . . . . . . 34
2.8.3 Citing the references . . . . . . . . . . . . . . . . . . . . . . . . 40
2.9 booktabs for tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
2.9.1 Recalling basic tables . . . . . . . . . . . . . . . . . . . . . . . 42
2.9.2 Vertical lines . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
2.9.3 Problems and solutions . . . . . . . . . . . . . . . . . . . . . . 45
3 Writing text 47
3.1 Obey orthography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3.2 Proper punctuation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3.2.1 A sentence ends with punctuation . . . . . . . . . . . . . . . . 47
3.2.2 Punctuation in commutative diagrams? . . . . . . . . . . . . . 48
3.2.3 Lists contain punctuation . . . . . . . . . . . . . . . . . . . . . 50
3.2.4 Hyphen and dashes . . . . . . . . . . . . . . . . . . . . . . . . . 51
3.2.5 Correct spacing after a period . . . . . . . . . . . . . . . . . . . 53
3.3 Text layout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
3.3.1 Use non-breakable space . . . . . . . . . . . . . . . . . . . . . . 56
3.3.2 Don’t use \\ or \newline . . . . . . . . . . . . . . . . . . . . . . 57
3.3.3 Text consisting of lists . . . . . . . . . . . . . . . . . . . . . . . 58
3.4 Mathematics in text . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
3.4.1 Don’t break inline mathematics . . . . . . . . . . . . . . . . . . 61
3.4.2 Don’t begin a sentence with a mathematical symbol . . . . . . 62
3.4.3 Don’t put two formulas next to each other . . . . . . . . . . . . 63
3.4.4 Don’t begin a line with a mathematical symbol (optional) . . . 64
3.5 Text in mathematics . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
3.5.1 Use \text . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
3.5.2 Put space around text in display mathematics . . . . . . . . . . 66
3.5.3 Use \intertext and \shortintertext . . . . . . . . . . . . . . . 67
3.6 Mathematics as text . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
3.6.1 Don’t abuse math as text . . . . . . . . . . . . . . . . . . . . . 70
3.6.2 Logical symbols don’t replace text . . . . . . . . . . . . . . . . 71
3.6.3 Don’t use “iff” . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
4 Good behaviour inside math mode 73

4.1 Using the right commands and symbols . . . . . . . . . . . . . . . . . 73
4.1.1 General symbols . . . . . . . . . . . . . . . . . . . . . . . . . . 73
4.1.2 Operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
4.1.3 Negations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
iv
Contents
4.1.4 Don’t use \limits . . . . . . . . . . . . . . . . . . . . . . . . . 76

4.1.5 Use extensible arrows instead of \overset and \underset . . . . 78
4.1.6 Know your ellipses . . . . . . . . . . . . . . . . . . . . . . . . . 79
4.1.7 Know your matrices . . . . . . . . . . . . . . . . . . . . . . . . 82
4.2 Avoid bad notation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
4.2.1 Don’t use \subset . . . . . . . . . . . . . . . . . . . . . . . . . 84
4.2.2 Don’t underline . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
4.2.3 Use dx instead of dx . . . . . . . . . . . . . . . . . . . . . . . . 85
4.2.4 Don’t force fancy fractions . . . . . . . . . . . . . . . . . . . . . 85
4.3 Defining new commands . . . . . . . . . . . . . . . . . . . . . . . . . . 86
4.3.1 \DeclareMathOperator . . . . . . . . . . . . . . . . . . . . . . . 86
4.3.2 \operatorname . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
4.3.3 Don’t abuse \mathrm . . . . . . . . . . . . . . . . . . . . . . . . 88
4.3.4 \newcommand . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
4.3.5 \DeclarePairedDelimiter . . . . . . . . . . . . . . . . . . . . . 89
4.3.6 The package xparse . . . . . . . . . . . . . . . . . . . . . . . . 91
4.4 Stretch your arrows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
4.5 Beware of spacings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
5 Mathematical layout 97
5.1 Environments for display mathematics . . . . . . . . . . . . . . . . . . 97
5.1.1 Don’t use $$ $$ or eqnarray . . . . . . . . . . . . . . . . . . . . 97
5.1.2 The environments \[ \] and equation* . . . . . . . . . . . . . 97
5.1.3 The environment gather* . . . . . . . . . . . . . . . . . . . . . 98
5.1.4 The environments align* and alignat* . . . . . . . . . . . . . 99
5.1.5 Don’t use center plus $ $ . . . . . . . . . . . . . . . . . . . . . 101
5.1.6 The environments gathered, aligned and alignedat . . . . . . . 101
5.1.7 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
5.2 Where to break and align formulas . . . . . . . . . . . . . . . . . . . . 102
5.2.1 When to break a formula . . . . . . . . . . . . . . . . . . . . . 103
5.2.2 Where to break and align a formula . . . . . . . . . . . . . . . 106
5.2.3 Don’t break formulas badly . . . . . . . . . . . . . . . . . . . . 110
5.3 Aligning nearly aligned formulas . . . . . . . . . . . . . . . . . . . . . 113
5.4 Proper spacing before multi-line display mode environments . . . . . . 115
5.5 Use the cases environment . . . . . . . . . . . . . . . . . . . . . . . . 116
5.6 Proper placement of the qed-symbol . . . . . . . . . . . . . . . . . . . 117
5.7 Tagging and numbering . . . . . . . . . . . . . . . . . . . . . . . . . . 121
5.7.1 Don’t autonumber all formulas . . . . . . . . . . . . . . . . . . 121
5.7.2 \tag . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
5.7.3 \notag . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
5.8 Multi-line set descriptions . . . . . . . . . . . . . . . . . . . . . . . . . 123
Bibliography 124
List of examples 125
v
Contents
Index 131
vi
List of tables
2.1 The commands \*name*ref. . . . . . . . . . . . . . . . . . . . . . . . . 19
2.2 Possible labels for the environment enumerate. . . . . . . . . . . . . . . 23
3.1 Kinds of hyphen and dashes. . . . . . . . . . . . . . . . . . . . . . . . 52

3.2 The Clebsch–Gordan coefficients in different languages. . . . . . . . . 52
3.3 Spacings after a period (with a zoom of factor 3.5). . . . . . . . . . . . 53
4.1 Right symbols and wrong symbols. . . . . . . . . . . . . . . . . . . . . 74

4.2 Binary operation and operator version. . . . . . . . . . . . . . . . . . . 75
4.3 Negated versions of popular symbols. . . . . . . . . . . . . . . . . . . . 75
4.4 Extensible arrows in amsmath, mathtools and extarrows. . . . . . . . . 80
4.5 Ellipses with a specific functions. . . . . . . . . . . . . . . . . . . . . . 81
4.6 Ellipses with a specific orientation. . . . . . . . . . . . . . . . . . . . . 81
4.7 Kinds of matrices. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
4.8 Standard distances for column sep. . . . . . . . . . . . . . . . . . . . . 93
4.9 Standard distances for row sep. . . . . . . . . . . . . . . . . . . . . . . 94
vii
List of figures
1.1 Splitting up a project into files. . . . . . . . . . . . . . . . . . . . . . . 5
3.1 Structure tree of the definition. . . . . . . . . . . . . . . . . . . . . . . 65

3.2 Structure tree of the mathematical expression. . . . . . . . . . . . . . 65
5.1 Deciding on a math environment. . . . . . . . . . . . . . . . . . . . . . 103
viii
Chapter 1
Before you start writing
A dozen mistakes and bad choices can be made before even the first word of actual text
is written. In this chapter we talk about how to prevent this from happening. Some
of the points – those about code quality and code management – are not unique to
writing LATEX.
1.1 Use version control

Use some kind of sensible version control for your code. The author uses Git but has
also heard good things about Mercurial. Sending code to yourself on Facebook is not a
proper system of version control.
1.2 The compilation process

1.2.1 Use LuaLATEX or XƎLATEX
Use LuaLATEX or XƎLATEX for native Unicode support (i.e. without including any kind
of additional packages). The author recommends LuaLATEX because the microtype
package his only a limited functionality under XƎLATEX.
1.2.2 Don’t ignore warnings

There are two ways in which LATEX will tell you that something went wrong: Errors
and warnings.
The occurrence of an error means that LATEX was unable the process the source code
and gave up at some point in the process. In the case of a warning LATEX again wasn’t
able to process the source code, but decided to produce some output nevertheless.
Hence the only difference between an error and a warning is how LATEX decided to
proceed with it. But there is no intrinsic difference between the two of them.
It is typically hard to ignore errors, as no output file will be generated. But warnings
also shouldn’t be ignored: While the occurrence of an error means that you get no
output at all, the occurrence of a warning means that you’re getting a faulty output.
1
1.3 Code quality
1.3 Code quality

1.3.1 Make use of whitespace to organize the code
The LATEX compiler almost never cares about unnecessary whitespace in your code.
Use generous amounts of whitespace to organize your code in a sensible way. Don’t do
the following:
Example 1.1: Code is too cramped
1 \[\begin{pmatrix*}[r]a^2+b^2&a^2-b^2\\-a^2+b^2&-a^2-b^2\end{pmatrix*}\]
Instead do the following:

Example 1.2: Code is less cramped
1 \[
2 \begin{pmatrix*}[r]
3 a^2 + b^2 & a^2 - b^2 \\
4 -a^2 + b^2 & -a^2 - b^2
5 \end{pmatrix*}
6 \]
You can (and should) even do the following:

Example 1.3: Code isn’t cramped
1 \[
2 \begin{pmatrix*}[r]
3 a^2 + b^2
4 &
5 a^2 - b^2
6 \\
7 - a^2 + b^2
8 &
9 - a^2 - b^2
10 \end{pmatrix*}
11 \]
This last one is easiest to write and easiest to navigate.

Also put at least one space between any two symbols that do not belong together.
Don’t do the following:
Example 1.4: Unrelated symbols need to be separated
1 $y=\sin(x)-e^x$
2
1.3 Code quality
Example 1.5: Unrelated symbols are separated
1 $y = \sin(x) - e^x$
One could (and probably should) go even further and do the following:
Example 1.6: Unrelated symbols are separated even better
1 $y = \sin ( x ) - e^x$
1.3.2 Write one sentence per line

Related the previous point, write at most one sentence per line. This has at least three
advantages:
• When an error or warning occurs LATEX will (try to) tell you the line of source code
in which the problem occurs. Having your source code split up over multiple lines
makes it much easier to find the problem.
• Organizing the source code in lines will in most cases make the tools provided by
your version control system more efficient to use.
• Sentences which are overly loong become easily recognizable.
1.3.3 Properly indent your source code

Writing LATEX means writing source code. Make sure that your source code is properly
indented.
1.3.4 Use % for problematic line breaks

Sometimes a sensible line break in your source code can lead to unwanted effects in the
resulting output. Consider the following example:
Example 1.7: Wrong spacing before a footnote
1
1 Here is some text.
Here is some text.
2 \footnote{Here is a footnote for
1 Here is a footnote for this text.
this text.}
Note the unwanted space between the period and the superscript of the footnote. This
unwanted space comes from the line break that occurs between them in the source
code. By ending the first line with % we can “comment out” this line break. We hence
do the following:
3
1.4 Splitting up a project into multiple files
Example 1.8: Right spacing before a footnote
1 Here is some text.%

Here is some text.1
2 \footnote{Here is a footnote for
1 Here is a footnote for this text.
this text.}

Any slightly larger project should be split up into multiple files.
1.4.1 The commands \usepackage, \include, \input

There are (at least) three useful ways to include another file into your project:
\usepackage, \includeand \input.
• Most of the preamble information – including of packages, configuration of styles,
configuration of the look and feels of the document – should be put into one or
multiple sty-files. These file(s) can then be included into the document by using
\usepackage.
• The commands \include and \input insert the text of the specified documents
at the position where they are used, but have slight differences:
◦ The command \include ensures that the specified content starts on a new
page, and also that the following content begins on a new page. The
command \input simply inserts the specified content at the given position
without any such additional formatting.
◦ The command \include cannot be nested: The specified file is not allowed
to include the command \include again.
One should use the command \include for chapters and the command \input for
any smaller level of text organization. One should in particular break the main text
into smaller units until each of them can be put into its own file of sensible length.
Suppose for example that your text consists of three chapters, each of which consists
of two sections. Then you should have at least ten files:
• A master file main.tex. This file includes all other files in some way, and this is
the file which needs to be compiled.
• A file mystyle.sty in which packages are included and options are set.
• Three files like chapter1.tex, chapter2.tex and chapter3.tex for the chapters.
• Six files for the sections, say section1.tex up to section6.tex.
These files should include each other as shown in Figure 1.1. We suppose for simplicity
that all ten files are contained in the same directory. The master file main.tex should
look roughly as follows:
4
main
mystyle chapter1 chapter2
section1 section2 section3 section4 section5 section6
Figure 1.1: Splitting up a project into files.
Example 1.9: Layout of main.tex
1 \documentclass[a4paper, 10pt]{scrreprt}
2
3 \usepackage{mystyle}
4
5 \title{A Report}
6 \author{John Doe}
7
8 \begin{document}
9
10 \maketitle
11
12 \include{chapter1}
14
15 \end{document}
The file mystyle.sty may look as follows:

Example 1.10: Layout of mystyle.sty
1 %%%%% PACKAGES
2
3 % general mathematics
4 \usepackage{mathtools}
5 \usepackage{amssymb}
6
7 % commutative diagrams
8 \usepackage{tikz-cd}
9
10 %%%%% NEW COMMANDS

11
5
12 % new operators
13 \DeclareMathOperator{\End}{End}
14 \DeclareMathOperator{\Hom}{Hom}
The file chapter1.tex should look roughly as follows:

Example 1.11: Layout of chapter1.tex
1 \chapter{Name of the first chapter}

2
3 Some introduction to this chapter before the first section appears.

4
5 \input{section1}
6 \input{section2}
The file section1.tex should look roughly as follows:

Example 1.12: Layout of section1.tex
1 \section{Name of the first section}

2
3 Text of the first section.
1.4.2 Use \includeonly

As a document grows so does the time that is needed to compile it. It is therefore often
desirable to compile only certain parts of the document. One naive solution to this
problem is to simply not include the currently unwanted parts, e.g. by commenting out
the corresponding lines of \include and \input. This approach has the problem that
references to the now-excluded parts of the document will not compile properly. This
problem can be fixed by using the command \inludeonly.
Suppose that your main document, say main.tex, included two other files, say
chapter1.tex and chapter2.tex, via the command \include. Hence the file main.tex
should look roughly as follows:
Example 1.13: Basic example of main.tex
2
3 \begin{document}
4
7
8 \end{document}
6
By putting \includeonly{chapter1} in the preamble we tell LATEX to only compile this

specific chapter.
Example 1.14: Using the command \includeonly
2
3 \includeonly{chapter1}
4 \begin{document}
5
8
9 \end{document}
If in the above situation chapter 1 contains some references to chapter 2 then these
references will still compile correctly, even though only chapter 1 is compiled.
The command \includeonly has two quirks, which follow from the way LATEX handles
referencing:
When LATEX compiles the file main.tex an auxiliary file main.aux is created. This
file contains various information about the compiled document. It does in particular
contain a list of all labels found in main.tex during the compilation process. The next
compilation process can then access these information to properly properly typeset all
references that refer to these labels.
A feature of the command \include (which the command \input does not have)
is that such an auxiliary file is also generated for the included file. So in the above
example the files chapter1.aux and chapter2.aux will be created. These files will in
particular contain lists of all the labels found in chapter1.tex and chapter2.tex. This
allows LATEX to access the labels in chapter2.tex even though only chapter1.tex is
compiled.
The auxiliary file chapter2.aux remains unchanged because LATEX does not go trough
the file chapter2.tex as long as the code \includeonly{chapter1} is used. This leads
to the two quirks mentioned above:
• Before inserting the code \includeonly{chapter1} one needs to compile the whole
document, including chapter2.tex, a least once. This needs to be done so that
LATEX can properly create the auxiliary file chapter2.aux.
• Changes made to the file chapter2.tex will not be noticed by LATEX while the
code \includeonly{chapter1} is present. This means in particular that LATEX
won’t notice any new labels that are added in this file. In this case one need to
stop excluding chapter1.tex for at least on compilation process to ensure that
the auxiliary file chapter2.aux is refreshed.
The command \includeonly actually accepts as its argument a list of files to be
included:
7
1.5 Choosing the document class
Example 1.15: Syntax of \includeonly
1 \includeonly{file_1, ..., file_n}

1.5.1 Use the KOMA-Script classes
Instead of the standard classes article, report or book the the more modern KOMA-
Script classes scrartcl, scrreprt and scrbook should be used. They provide more
functionalities then the standard classes and a more unified interface.
The KOMA-Script classes make slightly different design choices than the standard
classes:
• The KOMA-Script classes have a different chapter title layout than the standard
classes. The standard look can be regained by setting the option chapterprefix
to true:
Example 1.16: Using the option chapterprefix
1 \documentclass[10pt, a4paper, chapterprefix = true]{scrbook}
• The default font of the KOMA-Script classes for headings and description labels
is a bold sans-serif font. It can be changed to the standard serif font with help of
the command \setkomafont.
Example 1.17: Changing KOMA fonts
1 \addtokomafont{disposition}{\rmfamily}
2 \addtokomafont{descriptionlabel}{\rmfamily}
One can also use the undocumented option egregdoesnotlikesansseriftitles.
Example 1.18: Using the option egregdoesnotlikesansseriftitles
1 \documentclass[10pt, a4paper, egregdoesnotlikesansseriftitles]{scrbook}
1.5.2 Deciding on a document class

The three standard classes to consider for a mathematical project are scrartcl, scrreprt
and scrbook. These three classes differ in what functionalities they provide and what
they standard settings are.
The class scrartcl is the most basic one. It provides the standard sectioning
commands \section, \subsection and \subsubsection. An abstract can be set with
the abstract environment and by giving \documentclass the option abstract = on:
8
Example 1.19: Using the environment abstract
1 \documentclass[10pt, a4paper, abstract = on]{scrartcl}

2
3 \begin{document}
4
5 \begin{abstract}
6 Our results are cool and our proofs have no details.
7 \end{abstract}
8
9 \end{document}
The class scrreprt extends the previous class: It provides the additional sectioning
command \chapter that precedes \section. The title is now printed its own page,
although this can by setting the option titlepage to false:
Example 1.20: Using the option titlepage
1 \documentclass[10pt, a4paper, titlepage = false]{scrreprt}
It also provides the sectioning command \abstract after which all chapters will be
counted towards the appendix. The abstract environment is still available in the same
way as before. (But the abstract is printed only on the second page, after the title
page.)
The class scrbook provides the additional sectioning command \part that pre-
cedes \chapter. This class also provides the (very useful) additional sectioning com-
mands \frontmatter, \mainmatter and \backmatter. The abstract environment isn’t
available anymore. This class distinguish between left and right pages, i.e. even and
odd pages. This can be disabled by using the option oneside:
Example 1.21: Using the option oneside
1 \documentclass[10pt, a4paper, oneside]{scrbook}
By default chapters will begin on the right side. This can be disabled with the
option openany:
Example 1.22: Using the option openany
1 \documentclass[10pt, a4paper, openany]{scrbook}
When the option oneside is used then chapters can start on both the left pages and
the right pages. So option openany is then not needed.
In practice one should use the class scrartcl if only sections and subsections are used,
and otherwise the class scrbook. Instead of scrreprt one should directly use scrbook:
The additional commands \frontmatter, \mainmatter and \backmatter that scrbook pro-
vides are extremely useful (see Section 1.6), and the visual difference between scrreprt
9
1.6 Use \frontmatter and its friends
and scrbook (which is their main difference) can easily be adjusted by passing the
option openany to \documentclass:
Example 1.23: Using the option openany
1 \documentclass[10pt, a4paper, openany]{scrbook}
1.6 Use \frontmatter and its friends

The class scrbook provides the commands \frontmatter, \mainmatter and\backmatter.
Together with the sectioning command \appendix these can be used to separate the
document into four parts: The front part, the main part, the appendix and the back
part. These parts behave slightly differently:
• The front part uses lower case roman numerals for page numbers. In the generated
Pagepdf-file these pages will again be numbered with roman numerals (this will
be important for the main part). Chapters in this part of the document will not
be numbered, but will appear in the table of contents.
This part of the document should include the title page, preface, table of content
and introduction. Basically everything that occurs before the first chapter.
• The main part uses Arabic numerals for page numbers and the page number is
reset when the main part starts. The first page of the main part is therefore
numbered “1”. In the generated pdf-file this first page will also be numbered as
such. This means that going to page 15 of the pdf-file will indeed give page 15
of the document. The chapters in this part of the document are numbered and
appear in the table of contents.
This part of the document does contain the vast amount of the document. It
contains all the chapters excluding the ones belonging into the appendix.
• The appendix part continues the numbering of the main part. It both resets the
chapter number and changes its style to upper case letters. The first chapter in
this part will hence be numbered “A”. If the option chapterprefix = true is set
then the word “Chapter” at the beginning of a new \chapter will be replaced
by “Appendix”.
This part of the document should contain all appendices.
• The back part continues the numbering of the appendix (and thus the numbering
of the main part). The chapters appearing in this part of the document are again
unnumbered, and no “Chapter” is printed at the beginning of a new \chapter is
printed.
This part of the document should contain the bibliography and the index. An
entry for the bibliography can be added to the table of contents by setting the
option bibliography to totoc.
The document layout with scrbook should therefore look roughly as follows:
10
1.7 Be consistent
Example 1.24: General document layout with scrbook
1 \documentclass[a4paper, 10pt,bibliography=totocnumbered]{scrbook}
2
3 % preamble
4
5 \begin{document}
6
7 \frontmatter
8 \maketitle
9 % dedication
10 % preface
11 \tableofcontents
12 % introduction
13
14 \mainmatter
15 % most of the document
16
17 \appendix
18 % the appendices
19
20 \backmatter
21 \printbibliography
22 % index
23 \end{document}
1.7 Be consistent
One of the most important aspects of both mathematical writing and the use of
mathematical notation is consistency. If you’re making crappy choices then at least do
them consistently, in the same way. Consider the following example:
Example 1.25: Inconsistency looks bad
1 If $X$ and $Y$ are two objects in a category $\mathcal{C}$ then it may happen
that the set $\operatorname{Hom}_{\mathcal{C}}(X,Y)$ is empty even though the
set $Hom_{\mathcal{C}}(Y,X)$ is non-empty.
If X and Y are two objects in a category C then it may happen that the set
HomC (X, Y ) is empty even though the set HomC (Y, X) is non-empty.
The inconsistent use of Hom and Hom makes the already bad Hom even worse.
11
Chapter 2
Useful and important packages
2.1 microtype for better typesetting
Use the package microtype. It makes your document look nicer and helps you to
circumvent overfull hboxes. Simply including the package is enough to let it work its
magic
2.2 mathtools for mathematics

The package amsmath is standard for mathematical typesetting. The package mathtools
is an extension of amsmath that fixes some of its problems and also provides some new
(and often times very useful) functionalities. The package mathtools automatically
loads amsmath, so instead of amsmath include mathtools.
Note that the important mathematical packages amssymb and amsthm are not auto-
matically loaded by mathtools and therefore still need to be included by hand. Two
other useful packages for mathematical content are stmaryrd (which provides some
more symbols) and occasionally extarrows (which provides certain kinds of extensible
arrows, see Table 4.4).
2.3 amsthm for theorem-like environments

To define theorem-like environments – like lemma, proposition, theorem, corollary,
etc., include the package amsthm. New environments can then be defined with the
command \newtheorem:
Example 2.1: Syntax of \newtheorem)
1 \newtheorem{internal name of the environment}{name to be printed}
Example 2.2: Using the command \newtheorem
1 % in the preamble
2 \newtheorem{proposition}{Proposition}
3 % in the main body
12
4 \begin{proposition}
5 Every finite subgroup of $k^\times$ is cyclic.
6 \end{proposition}
Proposition 1. Every finite subgroup of k × is cyclic.
The variant \newtheorem* defines unnumbered theorem-like environments:

Example 2.3: Using the command \newtheorem*
1 % in the preamble
2 \newtheorem*{claim}{Claim}
3 % in the main body
4 \begin{claim}
5 The symmetric group $S_3$ is the smallest non-abelian group.
6 \end{claim}
Claim. The symmetric group S3 is the smallest non-abelian group.
If multiple theorem-like environments are defined then their have by default indepen-
dent counters:
Example 2.4: Theorem-like environment use different counters by default
1 %in the preamble

2 \newtheorem{idea}{Idea}
3 \newtheorem{problem}{Problem}
4 %in the main text
5 \begin{idea}
6 Fly to the moon in a car.
7 \end{idea}
8 \begin{problem}
9 Cars ’dont fly.
10 \end{problem}
Idea 1. Fly to the moon in a car.

Problem 1. Cars don’t fly.
For mathematical texts this behavior is pretty bad, as it makes it harder to find a
specified result.1
To solve this problem we define a new counter alltheorems and tell all theorem-
like environments to use this counter. To define the now counter we use the com-
1 If page 492 features Lemma 112 and Proposition 43 then where is Remark 20?
13
mand \newcounter:
Example 2.5: Syntax of \newcounter
1 \newcounter{name}[dependence]
The explicit code looks as follows

Example 2.6: Setting up a common counter
1 % in the preamble
2 \newcounter{alltheorems}
3
4 \newtheorem{assumption}[alltheorems]{Assumption}
5 \newtheorem{consequence}[alltheorems]{Consequence}
6
7 % in the main text

8 \begin{assumption}
9 Cats hunt mice.
10 \end{assumption}
11
12 \begin{assumption}
13 Tigers are cats.
14 \end{assumption}
15
16 \begin{consequence}
17 Tigers hunt mice.
18 \end{consequence}
Assumption 1. Cats hunt mice.

Assumption 2. Tigers are cats.
Consequence 3. Tigers hunt mice.
In practice one often wants the counter to be bound to the surround section or even
chapter. This can be achieved by binding the new counter to the section counter:
Example 2.7: Binding a new counter to the section level
1 % in the preamble
2 \newcounter{sometheorems}[section]
3 \renewcommand{\thesometheorems}{\thesection.\arabic{sometheorems}}
4 \newtheorem{corollary}[sometheorems]{Corollary}
5

7 \section{Free abelian groups}
14
2.4 tikz-cd for commutative diagrams
9 \begin{corollary}
10 Every subgroup of a free abelian group is again free abelian.
11 \end{corollary}
12
13 \begin{corollary}
14 Every subgroup of $\mathbb{Z}^n$ admits a basis.
15 \end{corollary}
16
17 \section{More free abelian groups}

18
19 \begin{corollary}
20 Every subgroup of a subgroup of $\mathbb{Z}^n$ admits a basis.
21 \end{corollary}
2.4 Free abelian groups

Corollary 2.4.1. Every subgroup of a free abelian group is again free abelian.
Corollary 2.4.2. Every subgroup of Zn admits a basis.
2.5 More free abelian groups

Corollary 2.5.1. Every subgroup of a subgroup of Zn admits a basis.
The addition [section] to the definition of the new counter ensures that the resulting
counter sometheorems resets every time the counter section is increased (which happens
every time a new section begins). We also change the way the counter sometheorems
is printed, namely printing it in the form “(section number).(counter number)” with
both numbers printed in Arabic numerals.
2.4 tikz-cd for commutative diagrams

There are many packages for drawing commutative diagrams. Many of them have a
rather restricted functionality, and quite a lot of them product bad looking output.
You should use the package tikz-cd. We refer to the very readable manual [CTN18]
for an explanation of this package.
15
2.5 cleveref and hyperref for referencing

2.5.1 Recalling basic referencing
To refer to a numbered part of the document, like a theorem, an item of a list, a
chapter or a section, one should never write down this number explicitly in the code.
The referencing system of LATEX should be used instead. Using this referencing system
always consists of two steps: Setting a label at the position that you want to refer to,
and then referencing this label at the desired position.
\label and \ref
The command for setting a label is \label:
Example 2.8: Syntax of \label
1 \label{labelname}
There are multiple commands for referencing this label, the most basic of which is \ref:
Example 2.9: Syntax of \ref
1 \ref{labelname}
Referencing a label with \ref will print the number of whatever the specified label
appears in.
Example 2.10: Using the commands \label and \ref
1 \begin{theorem}
2 \label{vector spaces are free}
3 Every vector space admits a basis.
4 \end{theorem}
5 It follows from \ref{vector spaces are free} that every $k$-vector space is
isomorphic to $k^{\oplus I}$ for some suitable index set $I$.
Theorem 1. Every vector space admits a basis.

It follows from 1 that every k-vector space is isomorphic to k ⊕I for some
suitable index set I.
Instead of giving just a simple number it is customary to also specify what hides
behinds this number, e.g. a theorem, table or figure. The specified name and number
should then be separated by a tie ~ (as explained in Section 3.3.1) to ensure that no
line break occurs at this position. The name of the referred to environment is usually
capitalised.
16
Example 2.11: Specifying the kind of reference
1 \begin{theorem}
2 \label{every vector space has a basis}
3 Every vector space admits a basis.
4 \end{theorem}
5 It follows from Theorem~\ref{vector spaces are free} that every $k$-vector
space is isomorphic to $k^{\oplus I}$ for some suitable index set $I$.
Theorem 2. Every vector space admits a basis.

It follows from Theorem 1 that every k-vector space is isomorphic to k ⊕I for
some suitable index set I.
\eqref
To refer to an equation it is customary to put the resulting number in parentheses.

This is done via the command \eqref.
Example 2.12: Syntax of \eqref
1 \eqref{labelname}
The command \eqref is used in the same way as the original \ref:
Example 2.13: Using the command \eqref
1 A classic result in mathematics shows

2 \begin{equation}
3 \label{important formula}
4 1 + 1 = 2 \,.
5 \end{equation}
6 Note that the identity \eqref{important formula} shows that ’Fermats conjecture
on the sum $a^n + b^n = c^n$ cannot be generalized to the case $n = 1$.
A classic result in mathematics shows
1 + 1 = 2. (2.1)
Note that the identity (2.1) shows that Fermat’s conjecture on the sum
an + bn = cn cannot be generalized to the case n = 1.
Always use descriptive label names. Cryptic sequences of seemingly random letters
will backfire on you down the road.
17
2.5.2 cleveref
There are two related problems with the above way of using hardcoding the type of
a reference: One has to remember or look up what type of environment the label
labelname refers to, and if this type is changed (e.g. by promoting a proposition to a
theorem) then the hardcorded types need to be manually adjusted.
These problems can be circumvented by using the cleveref package. The author
recommends to load this package with the options capitalise and noabbrev:
Example 2.14: Loading the package cleveref
1 \usepackage[capitalise, noabbrev]{cleveref}
The command \cref
The package cleveref provides the command \cref
Example 2.15: Syntax of \cref
1 \cref{labelname}
The command \cref differs from the more primitive \ref in that it automatically
inserts the right kind of type before the reference.
Example 2.16: Using the command \cref
1 \begin{lemma}
2 \label{dim is well-defined}
3 Every two bases of a vector space have the same cardinality.
4 \end{lemma}
5
6 \begin{remark}
7 One can generalize \cref{dim is well-defined} to non-commutative rings:
8 If $R$ is some ring and $M$ is a semisimple $R$-module then for every
irreducible $R$-module $L$ the multiplicity of $L$ in $M$ is well-defined.
9 \end{remark}
Lemma 3. Every two bases of a vector space have the same cardinality.
Remark 4. One can generalize Lemma 3 to non-commutative rings: If R is
some ring and M is a semisimple R-module then for every irreducible R-module
L the multiplicity of L in M is well-defined.
The used options options capitalise and noabbrev have the following effects:
• The option capitalise ensures that the printed type of the reference will begin
with an upper case letter. In Example 14 we would otherwise get “lemma 3”
18
lower case upper case default case

singular \lcnamecref \nameCref \namecref
plural \lcnamecrefs \nameCrefs \namecrefs
Table 2.1: The commands \*name*ref.
instead of “Lemma 3”.

• The option noabbrev ensures that printed types won’t be abbreviated. One wil
otherwise get “eq. (5)” instead of “equation (5)”.
The command \cref has the variant \Cref which ensures that the inserted type will
start with an upper case letter. One should always use \Cref instead of \cref at the
beginning of a sentence, even if the option capitalise is set.
The commands \*name*ref
The package cleveref provides another family of useful commands aside from \cref
and \Cref. An overview of these commands can be found in Table 2.1. They can be
used to print the type of a reference without its number. This can be used be used to
circumvent hardcoding types into the source code:
Example 2.17: Using \lcnamecref
1 \begin{theorem}
2 \label{weak cayley}
3 Every group embeds into a non-abelian group.
4 \end{theorem}
5 The above \lcnamecref{weak cayley} is actually a corollary of ’Cayleys~theorem.
Theorem 5. Every group embeds into a non-abelian group.
The above theorem is actually a corollary of Cayley’s theorem.
By using the various referencing commands introduced so far it is now possible to avoid
(nearly?) every kind of hardcorded type.
2.5.3 hyperref
If the resulting pdf-file is supposed to be navigated digitally then the package hyperref
should be used. This package puts hyperlinks in the pdf-file whenever some kind of
reference is used.
19
2.6 csquotes for quotation marks
2.5.4 Order of inclusion

The packages cleveref and hyperref are a bit peculiar when it comes to where they
have to be included in the preamble. The general rule is that the package hyperref
should be included as the very last package to ensure that it interacts properly with
all other used packages. There are some rare exceptions to this rule, one of which
happened to be cleveref.
If you’re defining some common counter for your theorem-like environments (which
you should do, as explained in Section 2.3) then this needs to be done after cleverref
was included. Otherwise cleveref will have problems knowing what names to print
when the command \cref is used.
Overall your preamble should the following order for these things:
Example 2.18: Order of preamble with cleveref and hyperref
1 % most packages
2 ...
3 \usepackage{amsthm}
4 ...
5
6 % the last packages

7 \usepackage{hyperref}
8 \usepackage{cleveref}
9
10 % defining theorem-like environments

11 \newcounter{everything}
12 \newtheorem{theorem}[everything]{Theorem}

Dealing with quotations marks in LATEX by hand can be a pain in the ass, for at least
two reasons: Different languages use different kinds of quotation marks, and finding the
right combination of LATEX code to get the correct ones can be a non-trivial problem.
One way to circumvent this problem is by using the package csquotes. This package
provides the command \enquote:
Example 2.19: Syntax of \enquote
1 \enquote{text}
This commands inserts quotation marks around the specified text:

Example 2.20: Using the command \enquote
1 \enquote{This is a quote.} “This is a quote.”
20
The command \enquote can automatically adjust the quotation marks to conventions
of the used language when this language is specified through the package babel. This
is done by loading the package csquotes which the option babel set to true.
Example 2.21: \enquote chooses the right kind of quotation marks
1 \begin{tabular}{@{}ll@{}}
2 \toprule
3 American English
4 &
5 \selectlanguage{american}
6 \enquote{quote}
7 \\
8 British English
9 &
10 \selectlanguage{british}
11 \enquote{quotation}
12 \\
13 German
14 &
15 \selectlanguage{ngerman}
16 \enquote{Zitat}
17 \\
18 French
19 &
20 \selectlanguage{french}
21 \enquote{citation}
22 \\
23 \bottomrule
24 \end{tabular}
American English “quote”

British English ‘quotation’
German „Zitat“
French « citation »
The command \enquote automatically handles nested quotation marks:

Example 2.22: Nested quotes with \enquote
1
“This is a ‘quote’ inside a quote.”
\enquote{This is a \enquote{quote}
inside a quote.}
So for dealing with quotes of any kind use the package csquotes.
21
2.7 enumitem for configuration of lists

In LATEX there are three different kinds of list environments: Numbered lists are
provided by the environment enumerate:
Example 2.23: Using the environment enumerate
1 \begin{enumerate}
2 \item
3 Assumption 1. Assumption
4 \item
5 ??? 2. ???
6 \item
7 Contradiction 3. Contradiction
8 \end{enumerate}
Unnumbered lists are provided by the environment itemize:
Example 2.24: Using the environment itemize
1 \begin{itemize}
2 \item
3 This is a list item. • This is a list item.
4 \item
5 This is also a list item. • This is also a list item.
6 \item
7 And yet another list item. • And yet another list item.
8 \end{itemize}
The environment description uses no predefined symbols for the list items and instead
expects a descriptive text from the user:
Example 2.25: Using the environment description
1 \begin{description}
2 \item[Field]
3 A special kind of ring. Field A special kind of ring.
4 \item[Ring]
5 A generalization of fields. Ring A generalization of fields.
6 \end{description}
The package enumitem is immensely useful for configuring the style and behavior of
these list environments. It provides (among others) the following features:
2.7.1 Style of numbering

For the environment enumerate the style of the numbering can be changed by using
the option label
22
option description
\alph* lower case alphabetic
\Alph* upper case alphabetic
\roman* lower case Roman numerals
\Roman* upper case Roman numerals
\arabic* Arabic numerals
Table 2.2: Possible labels for the environment enumerate.
Example 2.26: Changing the numbering style of enumerate lists
1 \begin{enumerate}[label = (\alph*)]
2 \item
3 First entry. (a) First entry.
4 \item
5 Second entry. (b) Second entry.
6 \item
7 Third entry. (c) Third entry.
8 \end{enumerate}
For a list of possible labels see Table 2.2. One can similarly change the symbol
for itemize lists via the option label:
Example 2.27: Changing the symbol for itemize lists
1 With the standard symbol:

2 \begin{itemize}[label =
{\textbullet}]
3 \item With the standard symbol:
4 First entry.
5 \item • First entry.
6 Second entry.
7 \end{itemize}
• Second entry.
8 Now with a different symbol:
Now with a different symbol:
9 \begin{itemize}[label =
{\textopenbullet}] ◦ First entry again.
10 \item
11 First entry again. ◦ Second entry again.
12 \item
13 Second entry again.
14 \end{itemize}
23
2.7.2 Resuming lists

One can resume lists with the option resume*. The option resume works similarly
to resume but doesn’t copy any style options from the previous list.
Example 2.28: Resuming lists with resume*
1 Some text before the first \texttt{enumerate} environment.

2 \begin{enumerate}[label = \roman*.]
3 \item
4 First entry.
5 \item
6 Second entry.
7 \end{enumerate}
8 Some text between the \texttt{enumerate} environment.
9 \begin{enumerate}[resume*]
10 \item
11 Third entry.
12 \item
13 Fourth entry.
14 \end{enumerate}
15 Some text after the second \texttt{enumerate} environment.
16 \begin{enumerate}[resume]
17 \item
18 Fifth entry.
19 \item
20 Sixth entry.
21 \end{enumerate}
22 Some text after the third \texttt{enumerate} environmen.
Some text before the first enumerate environment.

i. First entry.
ii. Second entry.

Some text between the enumerate environment.
iii. Third entry.
iv. Fourth entry.

Some text after the second enumerate environment.
5. Fifth entry.
6. Sixth entry.
Some text after the third enumerate environmen.
24
2.7.3 Spacing
One can change the various spacings involved in the list environments. For a full
overview we refer to [CTN19a], and in particular to [CTN19a, Figure 1]. We give here
only some of the more important configurations.
Multiple paragraphs in a list item

If a list item contains multiple paragraphs of text then the new paragraphs are by
default not indented but instead separated from the preceding paragraph by some
additional vertical space:
Example 2.29: Indentation and distance of pararaphs in lists, default behavior
1 \begin{enumerate}
2 \item
3 Lorem ipsum dolor sit amet, 1. Lorem ipsum dolor sit amet,
consectetur adipiscing elit, sed do consectetur adipiscing elit, sed
eiusmod tempor incididunt ut labore do eiusmod tempor incididunt
et dolore magna. ut labore et dolore magna.
4
5 In massa tempor nec feugiat.

In massa tempor nec feugiat.
Nunc aliquet bibendum enim Nunc aliquet bibendum enim
facilisis gravida. facilisis gravida.
6 \end{enumerate}
One can adjust the indentation of the paragraphs with the option listparindent and
the additional vertical spacing between paragraphs with the option parsep. With this
we can regain the standard indentation pattern in lists:
Example 2.30: Indentation and distance of pararaphs in lists, changed behavior
1 \begin{enumerate}[listparindent =
\parindent, parsep = 0pt]
2 \item
3 Lorem ipsum dolor sit amet,
1. Lorem ipsum dolor sit amet,
consectetur adipiscing elit, sed do consectetur adipiscing elit, sed
eiusmod tempor incididunt ut labore do eiusmod tempor incididunt
et dolore magna. ut labore et dolore magna.
5 In massa tempor nec feugiat. Nunc aliquet bibendum enim
Nunc aliquet bibendum enim facilisis gravida.
facilisis gravida.
6 \end{enumerate}
Distance between items

One can control the distance between list itemns with the option itemsep:
25
Example 2.31: Distance between list items with itemsep
1 A list with standard item distance:

2 \begin{enumerate}
3 \item
4 First entry.
5 \item
6 Second entry.
7 \end{enumerate}
8 Now a list without additional item distance:
9 \begin{enumerate}[itemsep = 0pt]
10 \item
11 First entry again.
12 \item
13 Second entry again.
14 \end{enumerate}
A list with standard item distance:

1. First entry.
2. Second entry.
Now a list without additional item distance:

1. First entry again.
2. Second entry again.
Indentation
The environment enumerate has by default two kinds of indentation: An indentation of
the labels with respect to the left margin, and an indentation of the list items with
respect to the label.
Example 2.32: enumerate environment with standard indentation
Lorem ipsum dolor sit amet,

2 \begin{enumerate}
3 \item 1. consectetur adipiscing elit, sed
4 consectetur adipiscing elit, sed
do eiusmod tempor incididunt
ut labore et dolore magna ali-
do eiusmod tempor incididunt ut
labore et dolore magna aliqua.
5 \end{enumerate}
qua.
We can remove both indentations by using the option wide with the value 0pt:
26
Example 2.33: enumerate environment without indentation

2 \begin{enumerate}[wide = 0pt] Lorem ipsum dolor sit amet,
3 \item
4 consectetur adipiscing elit, sed 1. consectetur adipiscing elit, sed do
do eiusmod tempor incididunt ut eiusmod tempor incididunt ut labore
labore et dolore magna aliqua. et dolore magna aliqua.
5 \end{enumerate}
We can add back the identation between the item labels and the item text by setting
the option leftmargin to the value *:
Example 2.34: enumerate environments with only identitation between label
and text

2
\begin{enumerate}[wide = 0pt,
leftmargin = *]
3 \item
1. consectetur adipiscing elit, sed do
eiusmod tempor incididunt ut la-
bore et dolore magna aliqua.
5 \end{enumerate}
One can also add back the indentation between the left margin and the item labels by
also using the option labelindent:
Example 2.35: enumerate environment with indentation between margin and
label, and label and text

2 \begin{enumerate}[wide = 0pt,
leftmargin = \parindent, Lorem ipsum dolor sit amet,
labelindent = \parindent]
3 \item 1. consectetur adipiscing elit, sed
4 consectetur adipiscing elit, sed do eiusmod tempor incididunt ut
do eiusmod tempor incididunt ut labore et dolore magna aliqua.
5 \end{enumerate}
One can also combine the above two indentations to regain something that looks like
the standard indentation:
27


2 \begin{enumerate}[wide = 0pt,
leftmargin = *, labelindent =
\parindent] 1. consectetur adipiscing elit, sed
3 \item
do eiusmod tempor incididunt
ut labore et dolore magna ali-
qua.
5 \end{enumerate}
But Example 36 does actually have an advantage over the standard indentation, namely
that it properly handels long labels:
1 The standard identation:

2 \begin{enumerate}[label = {(Latin \arabic*)}]
3 \item
4 Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod
tempor incididunt ut labore et dolore magna aliqua.
5 \end{enumerate}
6 An indentation with adjusted parameters:
7 \begin{enumerate}[label = {(Latin \arabic*)}, wide = 0pt, leftmargin = *,
labelindent = \parindent]
8 \item
9 Dolor sed viverra ipsum nunc aliquet bibendum enim.
11 \end{enumerate}
The standard identation:

(Latin 1) Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod
tempor incididunt ut labore et dolore magna aliqua.
An indentation with adjusted parameters:
(Latin 1) Dolor sed viverra ipsum nunc aliquet bibendum enim. In massa
tempor nec feugiat.
The identitation for the environment itemize can be adjusted in the same way as for
the environment description.
For the environment description we can proceed similarly as for enumerate. We
first observe the standard identation for this environment: There is a fixed identation
between the left margin and the items texts. The item labels are aligned at the
28
left margin.
Example 2.38: description environment with standard indentation
1 Text, text, text, text, text, text,

text.
Text, text, text, text, text, text, text.
2 \begin{description}
First item Text, text, text, text,
3
text, text, text, text, text, text,
\item[First item]
4 Text, text, text, text, text,
text, text, text, text, text, text.
text.
5 \item[Second item]
Second item Text, text, text, text,
6
text, text, text, text, text, text,
Text, text, text, text, text,
7 \end{description}
text.
We can remove the identation by again using the option wide.
Example 2.39: description environment with no indentation
1 Text, text, text, text, text, text,

text.
2 \begin{description}[wide = 0pt]
Text, text, text, text, text, text, text.
3 \item[First item]
First item Text, text, text, text,
4
Text, text, text, text, text,
text, text, text, text, text.
5 \item[Second item]
Second item Text, text, text, text,
6 Text, text, text, text, text,
text, text, text, text, text.
7 \end{description}
2.7.4 Global settings

Global settings can be set via the command \setlist:
Example 2.40: Syntax of \setlist
1 \setlist[kind of list]{options}
Consider the following example:
29
Example 2.41: Global settings for list environments
1 \setlist[enumerate]{label =
\roman*)}
2 \begin{enumerate}
3 \item i) First entry.
4 First entry.
5 \item ii) Second entry.
6 Second entry.
7 \end{enumerate}
2.7.5 Nested lists

When lists are nested one can use different settings for each list.
Example 2.42: Different settings for nested lists
1 \begin{enumerate}[label = \Roman*)]
2 \item
3 First entry.
4 \begin{enumerate}[label = \alph*)]
5 \item
6 First entry, first subentry.
7 \item
8 First entry, second subentry.
9 \end{enumerate}
10 \item
11 Second entry.
12 \begin{enumerate}[label = \arabic*)]
13 \item
14 Second entry, first subentry.
15 \item
16 Second entry, second subentry.
17 \end{enumerate}
18 \end{enumerate}
I) First entry.
a) First entry, first subentry.
b) First entry, second subentry.
II) Second entry.
1) Second entry, first subentry.
2) Second entry, second subentry.
30
One can also use different global settings for different depths:
Example 2.43: Global settings depending on level
1 \setlist[enumerate, 1]{label = (\roman*)}

2 \setlist[enumerate, 2]{label = (\alph*)}
3 \begin{enumerate}
4 \item
5 First entry.
6 \begin{enumerate}
7 \item
8 First entry, first subentry.
9 \item
10 First entry, second subentry.
11 \end{enumerate}
12 \item
13 Second entry.
14 \begin{enumerate}
15 \item
16 Second entry, first subentry.
17 \item
18 Second entry, second subentry.
19 \end{enumerate}
20 \end{enumerate}
(i) First entry.

(a) First entry, first subentry.
(b) First entry, second subentry.
(ii) Second entry.
(a) Second entry, first subentry.
(b) Second entry, second subentry.
The counter of the first depth and second depth can be accessed via the counters enumi
and enumii:
Example 2.44: Accessing level counters in settings for list environments
1 \setlist[enumerate, 1]{label = (\arabic*)}

2 \setlist[enumerate, 2]{label = (\arabic{enumi}.\alph*)}
3 \begin{enumerate}
4 \item
5 An entry.
6 \begin{enumerate}
7 \item
31
8 Again an entry.
9 \item
10 Again an entry.
11 \end{enumerate}
12 \item
13 Another entry.
14 \begin{enumerate}
15 \item
16 Yet another entry.
17 \item
18 Yet another entry.
19 \end{enumerate}
20 \end{enumerate}
(1) An entry.
(1.a) Again an entry.
(1.b) Again an entry.
(2) Another entry.
(2.a) Yet another entry.
(2.b) Yet another entry.
2.7.6 Cloning lists

New list types can be constructed by cloning an already existing one via the com-
mand \newlist:
Example 2.45: Syntax of \newlist
1 \newlist{new list}{original list}{depth of new list}
One can then set global options for this new list type without changing the global
options for the original type. This can be used to construct list environments that
serve special purposes. In the following example we create a new list environment that
is meant for listing equivalent statements:
Example 2.46: Custom clones of list environments
1 % clone enumerate as equivalenceslist, allowing up to 2 levels

2 \newlist{equivalenceslist}{enumerate}{2}
3 % set the formatting
4 \setlist[equivalenceslist,1]{label = (\roman*)}
32
2.8 biblatex for bibliography
5 \setlist[equivalenceslist,2]{label = (\alph*), leftmargin = *}

6 % an example
7 Let $M$ be an $R$-module.
8 For every collection of elements $x_1, \dotsc, x_n \in M$ the following
conditions are equivalent:
9 \begin{equivalenceslist}
10 \item
11 For every $R$-module $N$ and every choice of elements $y_1, \dotsc, y_n \in
N$ there exists a unique module homomorphism $f \colon M \to N$ with $f(x_i) =
y_i$ for every $i = 1, \dotsc, n$.
12 \item
13 The elements $x_1, \dotsc, x_n$ are a basis of $M$, i.e.
14 \begin{equivalenceslist}
15 \item
16 the elements $x_1, \dotsc, x_n$ are linearly independent, and
17 \item
18 the elements $x_1, \dotsc, x_n$ are a generating set of $M$.
19 \end{equivalenceslist}
20 \end{equivalenceslist}
Let M be an R-module. For every collection of elements x1 , . . . , xn ∈ M the

following conditions are equivalent:
(i) For every R-module N and every choice of elements y1 , . . . , yn ∈ N there

exists a unique module homomorphism f : M → N with f (xi ) = yi for
every i = 1, . . . , n.
(ii) The elements x1 , . . . , xn are a basis of M , i.e.
(a) the elements x1 , . . . , xn are linearly independent, and
(b) the elements x1 , . . . , xn are a generating set of M .

2.8.1 The basic setup
A bibliography in LATEX comes about from the interplay of three different actors:
• A bib-file that contains the various references and their information.
• A bibliography package that provides commands for citing these references.
• A back end program that accesses the bib-file to extract the needed information
and pass them to LATEX.
One should choose biblatex for the bibliography package and biber for the back end
program. For this the package biblatex needs to be loaded with the option backend
set to biber:
33
Example 2.47: Loading the package biblatex
1 \usepackage[backend = biber]{biblatex}
The author also likes to use the following options:

• By default the occurring references will simply be numbered as [1], [2], [3], etc.
Often references of the form [Eis04] are preferable, which is achieved by setting
the option style to alphabetic.
• Setting the option dateabbrev to false ensures that month names like “September”
are not abbreviated as “Sept.”
• Setting the option urldate to long ensure that dates concerning URLs are written
out as “September 4, 2109” instead of “09/04/2019”.
2.8.2 Creating the bib-file

To most important step of creating a bibliography is to collect the references and their
various metadata in a bib-file. For every reference we need to add an entry to this
bib-file. These entries have the following form:
Example 2.48: Syntax for an entry of the bib-file
1 @type{label,
2 key1 = {value1},
3 key2 = {value2},
4 key3 = {value3},
5 ...
6 }
Instead of curly braces { } one can also use quotation marks " " on the right hand
side of the equality signs.
The specifier @type in Example 48 will be be replaced by something like @book
or @article to explain what kind of work this entry is. A list of all possible types can
be found in [CTN19c, 2.1]. This specified type will determine which of the given data
will be printed in the bibliography and how these printed date are formatted.
The given text label has no influence on the bibliography itself. It will be used to
add the citations to this reference in the main text.
How to choose data for the bibliography

One should follow two guidelines when adding information to the bibliography.
• Provide as much data as possible. The specified type will determine which of
these data will be printed. To find out which type will use which information we
refer again to [CTN19c, 2.1, 2.2].
34
• How the printed data are to be formatted is for LATEX – and more specifically
biblatex – to decide. So don’t try to preformat the provided date in the bib-file.
Try in particular to give full, unabbreviated names whenever possible.
If you feels strongly about certain data being printed, or how certain data should be
formatted when printed out, then you should not try to abuse the bib-file for this.
Instead tell these complains to biblatex by changing the appropriate settings.
Some good sources for finding the data that a bibliography requires are MathSciNet
and the websites of the publishers. (Springer is quite good at providing all the needed
information on the websites of their books.) Most of the needed data can also be found
in the cited resource – e.g. book or article – itself.
In the following we will look at some specific examples of bib-file entries.
Entry for a single book

The bib-file entry for a single book should look as follows:
Example 2.49: bib-file entry for a single book
1 @book{fultonharris2004,
2 title = {Representation Theory},
3 subtitle = {A First Course},
4 author = {Fulton, William and Harris, Joe},
5 edition = {1},
6 year = {2004},
7 pagetotal = {xv+551},
8 publisher = {Springer-Verlag New York},
9 series = {Graduate Texts in Mathematics},
10 number = {129},
11 isbn = {978-0-387-97527-6},
12 doi = {10.1007/978-1-4612-0979-9}
13 }
The resulting output in the bibliography (see Example 58) will look as follows:
William Fulton and Joe Harris. Representation Theory. A First Course. 1st ed.
Graduate Texts in Mathematics 129. Springer-Verlag New York, 2004, pp. iv–
15. xv+551 pp. isbn: 978-0-387-97527-6. doi: 10.1007/978-1-4612-0979-9
The type @books tells LATEX that the entry is a (single) book. The various keys have
the following functions:
title This key specifies the title of the book. The expected value of this key is a text.
subtitle This key specifies the subtitle of the book. The expected value of this key is
a text.
author This key specifies the author(s) of the book. There are some things to be aware
of here:
35
• The name of an author needs to be given in the format Lastname, Firstname. This
is needed so that biber can properly process this data.
• If multiple authors are given then they need to be separated by the word and.
edition This key specifies the edition of the book. The value should be given as a
number for proper processing, but can in an emergency also be given as a text.
year This key specifies the year the book was published. One could also specify a moth
with the key month. The values for both keys are expected to be numbers.
One could also use the key date takes arguments of the form year, year-month
or year-month-day. Here the value of year is expected to be a four digit number, and
the values of month and day are expected to be two digit numbers (which may contain
a leading zero).
pagetotal This key specifies the total number of pages of the book. The value should
be an integer but can also be an arbitrary text.
There is however a drawback to simply providing a text: Normally the number of
pages is followed by the text “pp.” or “p.”, depending on whether the reference consists
of only a single page. To do so biblatex always tried to interpret the input as a number.
But if this interpretation fails then neither “pp.” nor “p.” will be added.
Books often start with pages that are numbered with Roman numbers, followed by
pages that are numbered by Arabic numbers. In this case the total number of pages
should be given in the form “(Roman number)+(Arabic number)”.
As explained above, this will lead to the problem of biblatex being unable to interpret
the input as a number, which leads by default to a missing “pp.” in the output. For
this problem one can adjust the settings of biblatex to always include “pp.” after the
total page number of a book. This can be done as follows:
Example 2.50: Adjusting the formatting of pagestotal
1 \DeclareFieldFormat[book]{pagetotal}{#1~\ppno}
publisher This key specifies the publisher of the book. The expected value is a text.
Note that “Springer” is not a proper reference for a publisher.
series Many mathematical books are part of some series, e.g. “Graduate Texts is
Mathematics” or “Cambridge Studies in Advanced Mathematics”. Such a series can be
specified with the key series, which expects as its value a text.
number This key specifies the number of the book in the previously specified series.
The values of this key is (counterintuitively) treated as a text.
If you copy your bibliography data from somewhere else then there is a very high
chance that instead of the key number the key volume is used. This is relicts from the
past that isn’t correct with biblatex.
isbn This key specifies the isbn number of the book. The value of this field is treated
as a text.
doi This key specifies the DOI of the book (if it has one).
36
Entry for a book with multiple volumes

Sometimes a book is just one volume in a small collection of books. In this case one
should use the type @mvbook to define the overall information of these books, and then
an entry of type @book which is subordinate to the previously created entry. Let’s
consider an example:
Example 2.51: bib-file entry for a book with multiple volumes
1 @mvbook{benson,
2 title = {Representations and Cohomology},
3 author = {Benson, David John},
4 publisher = {Cambridge University Press},
5 series = {Cambridge Studies in Advanced Mathematics},
6 volumes = {2}
7 }
8
9 @book{benson1991,
10 crossref = {benson},
11 volume = {1},
12 number = {30},
13 title = {Basic Representation Theory of Finite Groups and Associative
Algebras},
14 edition = {1},
15 year = {1991},
16 pagetotal = {xii+246},
17 isbn = {978-0-521-36134-7},
18 doi = {10.1017/CBO9780511623615}
19 }
We can then refer to the entry of type @book as usual, to get the following output in
the bibliography (see Example 58):
David John Benson. Representations and Cohomology. Vol. 1: Basic Represen-
tation Theory of Finite Groups and Associative Algebras. 2 vols. Cambridge
Studies in Advanced Mathematics 30. Cambridge University Press, 1991,
pp. xii+246. isbn: 978-0-521-36134-7. doi: 10.1017/CBO9780511623615
There are three new keys to talk about here:
volumes This key describes the total number of volumes. The value of this key is
expected to be an integer.
volume This key describes the volumes of the specified volume. The value of this key is
expected to be an integer.
crossref This key expresses that the given entry is subordinate to some other entry.
The expected value is the label of the superior entry.
37
Entry for an article

We now consider an example for citing an article:
Example 2.52: bib-file entry for a single book
1 @article {diamond_lemma,
2 title = {The Diamond Lemma for Ring Theory},
3 author = {Bergman, George Mark},
4 year = {1978},
5 month = {2},
6 journaltitle = {Advances in Mathematics},
7 issn = {0001-8708},
8 volume = {29},
9 number = {2},
10 pages = {178--218},
11 doi = {10.1016/0001-8708(78)90010-5}
12 }
The output in the bibliography (see Example 58) will look as follows:
George Mark Bergman. “The Diamond Lemma for Ring Theory.” In: Advances
in Mathematics 29.2 (February 1978), pp. 178–218. doi: 10 . 1016 / 0001 -
8708(78)90010-5
The type @article tells LATEX that the entry is for a (journal) article. Many fields
are as for the type @book, so we will focus on the changes:
journaltitle This key specifies the name of the journal that the article was published
in. The expected value for this key is a text.
issn This key specifies the ISSN (International Standard Serial Number) of the journal
in question. The value is treated as a text.
volume, number These keys specify in which volume of the journal the article appeared,
and in which number of the volume. The value for volume should be an integer, and
the value for number should be an integer too (although it is treated as text).
pages This key specifies in which page range the article appeared. It doesn’t matter
how many dashes are used to separate the two page numbers. It also doesn’t matter if
the dash(es) are surrounded by space.
It is customary to specify page ranges in the form pages = number--number because
this gives a right looking output even if this argument were simply to be interpreted as
text. (Which lesser packages than biblatex may do.)
Entry for an online resource

We consider now an example where we cite an online resource. For this we use the
generic type @online.
38
Example 2.53: bib-file entry for an online resource
1 @online{cayley_graph,
2 title = {Cayley graphs and the geometry of groups},
3 author = {Tao, Terence},
4 date = {2010-06-10},
5 url =
{https://terrytao.wordpress.com/cayley-graphs-and-the-geometry-of-groups},
6 urldate = {2019-09-06}
7 }
Terence Tao. Cayley graphs and the geometry of groups. July 10, 2010.
url: https://terrytao.wordpress.com/cayley-graphs-and-the-geometry-of-
groups (visited on September 6, 2019)
We have three (or rather two) new fields to discuss:

date This key specifies when the linked to resource was created. This key takes
arguments of the form year, year-month or year-month-day. Here the value of year is
expected to be a four digit number, and the values of month and day are expected to be
two digit numbers (which may contain a leading zero).
Instead of date one can also use the overall less specific keys year and month.
url This key specifies the URL of the online resource.
urldate This key specifies the date on which the online resource was accessed. This is
an important information since content online may change over time.
Entry for an arXiv article

To cite an article published on arXiv one shouldn’t use the key url because biblatex
has a built-in way of dealing with arXiv articles. For an article that uses the current
identifier scheme consider the following example:
Example 2.54: bib-file entry for an online resource
1 @online{leinster2014,
2 title = {The bijection between projective indecomposable and simple
modules},
3 author = {Leinster, Tom},
4 date = {2014-10-14},
5 eprint = {1410.3671v1},
6 eprinttype = {arxiv},
7 eprintclass = {math.RA},
8 urldate = {2019-09-12}
9 }
39
Tom Leinster. The bijection between projective indecomposable and simple

modules. October 14, 2014. arXiv: 1410 . 3671v1 [math.RA]. (Visited on
September 12, 2019)
We have used three new fields:
eprinttype This key specifies what kind of resource the entry is. The used value arxiv
results in a predefined formatting of the resulting output.
eprint This key specifies the identifier of the resource.
eprintclass This field specifies additional information about the resource.
2.8.3 Citing the references

Suppose now that we have added an entry to our bib-file, as outlines in Example 48.
In the actual LATEX project we can then refer to this entry with the command \cite:
Example 2.55: Syntax of \cite
1 \cite[details]{label}
2 One can also refer to the overall collection:
3 \testcite{benson}
Let’s consider an example where we cite the references given in the previous examples:
Example 2.56: Using the command \cite
1 We assume that the reader is familiar with the representation theory of the
symmetric groups as discussed in \cite[\S 4]{fultonharris2004}.
2 The reader may also want to check out \cite{benson1991} and \cite{cayley_graph}.
3 For a nice proof of the Poincaré--Birkhoff--Witt~Theorem we refer to \cite[\S
3]{diamond_lemma}.
We assume that the reader is familiar with the representation theory of the sym-
metric groups as discussed in [FH04, §4]. The reader may also want to check out
[Ben91] and [Tao10]. For a nice proof of the Poincaré–Birkhoff–Witt Theorem
we refer to [Ber78, §3].
We will also need to add the bibliography into the LATEX document. Suppose for
this that the bib-file is called references.bib. We then have to do two things:
• We need to tell LATEX how the bib-file is called. This is done via the com-
mand \bibliography:
Example 2.57: Syntax of \bibliography
1 \bibliography{name of bib-file}
In our case we have to add \bibliography{references.bib}.
40
• We have to add the command \printbibliography at the position in source code

where we want the bibliography to be printed. This typically happens near the
end of the document.
• We set the option bibliography of \documentclass to totoc to add the bibliography
to the table of contents. (see Example 24).
In our running example(s) we would get the following bibliography:
Example 2.58: Using the command \printbibliography
1 \printbibliography
Bibliography
[Ben91] David John Benson. Representations and Cohomology. Vol. 1: Basic
Representation Theory of Finite Groups and Associative Algebras.
2 vols. Cambridge Studies in Advanced Mathematics 30. Cambridge
University Press, 1991, pp. xii+246. isbn: 978-0-521-36134-7. doi:
10.1017/CBO9780511623615.
[Ber78] George Mark Bergman. “The Diamond Lemma for Ring Theory.”
In: Advances in Mathematics 29.2 (February 1978), pp. 178–218.
doi: 10.1016/0001-8708(78)90010-5.
[FH04] William Fulton and Joe Harris. Representation Theory. A First
Course. 1st ed. Graduate Texts in Mathematics 129. Springer-Verlag
New York, 2004, pp. iv–15. xv+551 pp. isbn: 978-0-387-97527-6.
doi: 10.1007/978-1-4612-0979-9.
[Lei14] Tom Leinster. The bijection between projective indecomposable and
simple modules. October 14, 2014. arXiv: 1410.3671v1 [math.RA].
(Visited on September 12, 2019).
[Tao10] Terence Tao. Cayley graphs and the geometry of groups. July 10,
2010. url: https://terrytao.wordpress.com/cayley-graphs-and-
the-geometry-of-groups (visited on September 6, 2019).
Compiling the bibliography

To get the output of Example 56 and Example 58 we actually have to compile the
document in the right way:
Suppose that the bib-file has created, we have put the citations in the text via \cite
and that we have placed \printbibliography in the source code. Suppose that our
main file is called main.tex and the bib-file is called references.bib. To get the desired
bibliography and cross-references between it and the text we need to complete three
steps:
41
2.9 booktabs for tables
1. We compile the document main.tex. The compiler will note down in an auxiliary
file main.bcf which labels are cited in this document
2. The back end program biber will go through the auxiliary files main.bcf and
write down all the requested information in a new auxiliary file main.bbl.
3. We compile the document main.tex again. The compiler will read the various
data given in the auxiliary file main.bll and, using the settings and commands
from the package biblatex, will typeset both the citations in the main text and
create a bibliography.
If you’re using a specialized LATEX editor or IDE (like TEXStudio, kile, etc.) and have
everything properly configured then your editor should take care of the above steps
automatically when(ever) the document is compiled. But if you’re compiling by hand
in the console then you will need three commands:
Example 2.59: Compiling in the console (with bibliography)
1 latex main.tex
2 biber main.bcf
3 latex main.tex

2.9.1 Recalling basic tables
The table itself
Recall that a table is constructed with the environment tabular:
Example 2.60: A basic table
1 \begin{tabular}{lcr}
2 longtext & text & text \\
3 text & longtext & text \\
4 text & text & longtext
5 \end{tabular}
longtext text text

text longtext text
text text longtext
The labels l, c, r specify the alignment of the corresponding column (left aligned,
centered, and right aligned).
42
Horizontal lines
The command \hline at the beginning of a row introduces a horizontal line that
separates this row from the previous one. To get multiple parallel lines (e.g. a double
line) one uses multiple instances of \hline directly after each other.
Example 2.61: Full horizontal lines in tables
1 \begin{center}
2 \begin{tabular}{ccc}
3 top left & top center & top right \\
4 \hline\hline
5 text & text & text \\
6 \hline
7 text & text & text
8 \end{tabular}
9 \end{center}
top left top center top right

text text text
text text text
Partial horizontal lines

To separate only the columns i, . . . , j by horizontal line, the command \cline is used:
Example 2.62: Syntax of \cline
1 \cline{start column-end column}
The command \cline has the same placement as \hline:
Example 2.63: Partial horizontal lines in tables
1 \begin{center}
2 \begin{tabular}{ccccc}
3 text & text & text & text & text \\
4 \cline{2-4}
5 text & text & text & text & text \\
6 \cline{1-2} \cline{4-5}
7 text & text & text & text & text
8 \end{tabular}
9 \end{center}
43
text text text text text

2.9.2 Vertical lines

One can put a vertical line between two columns by adding the symbol | between
the corresponding alignment symbols. Inserting this symbol multiple times will give
parallel vertical lines.
Example 2.64: Vertical lines in tables
1 \begin{tabular}{l||c|c}
2 first row & text & text \\
3 second row & text & text \\
4 third row & text & text
5 \end{tabular}
first row text text

second row text text
third row text text
Different kinds of lines

The above effects can also be combined:
Example 2.65: A table with all kinds of lines in it
1 \begin{center}
2 \begin{tabular}{|l||l|r|}
3 \hline
4 \textbf{Country} & \textbf{Town} & \textbf{Population} \\
5 \hline\hline
6 France & Paris & 2,229,621 \\
7 \cline{2-3}
8 {} & Marseille & 855,393 \\
9 \hline
10 Germany & Berlin & 3,520,031 \\
11 \cline{2-3}
12 {} & Hamburg & 1,787,408 \\
13 \hline
14 Japan & Tokyo & 8,637,098 \\
44
15 \cline{2-3}
16 {} & Yokohama & 3,697,894 \\
17 \hline
18 \end{tabular}
19 \end{center}
Country Town Population

France Paris 2,229,621
Marseille 855,393
Germany Berlin 3,520,031
Hamburg 1,787,408
Japan Tokyo 8,637,098
Yokohama 3,697,894
2.9.3 Problems and solutions

The above table has a huge problem: It’s ugly. This is due to various reasons:
• The spacing between the horizontal lines and the text below them is both bad
and inconsistent.
• The above table breaks the first rule of table club: Never, ever use vertical lines.
• The above table also breaks the second rule of table club: Never use double lines.
The last two points are easy to fix. The package booktabs gives a way for fixing the
first problem:
• The package provides the commands \toprule, \midrule and \bottomrule as
replacements for \hline. The command \toprule is to be used only for the line
on top of the table. The command \bottomrule is similarly only to be used for
the line on the bottom of the table. The horizontal line \midrule is meant to
separate the main part of the table from the top part and bottom part.
• The command \cmidrule is the replacement for \crule.
One should also try to minimize the number of horizontal lines. The above example
should hence look as follows:
Example 2.66: Using the rules of booktabs
1 \begin{center}
2 \begin{tabular}{llr}
3 \toprule
4 \textbf{Country} & \textbf{Town} & \textbf{Population} \\
5 \midrule
45
6 France & Paris & 2,229,621 \\

7 {} & Marseille & 855,393 \\
8 Germany & Berlin & 3,520,031 \\
9 {} & Hamburg & 1,787,408 \\
10 Japan & Tokyo & 8,637,098 \\
11 {} & Yokohama & 3,697,894 \\
12 \bottomrule
13 \end{tabular}
14 \end{center}
Country Town Population

France Paris 2,229,621
Marseille 855,393
Germany Berlin 3,520,031
Hamburg 1,787,408
Japan Tokyo 8,637,098
Yokohama 3,697,894
Note that we could leave out all non-essential horizontal lines because the table has
a very regular form. We refer to [CTN16] for more information about typing tables.
46
Chapter 3
Writing text
Un texte mathématique est,
d’abord, un texte.
Michèle Audin, Conseils aux

auteurs de textes
mathématiques
In this chapter we will talk about the more text focused aspects of writing mathematics
with LATEX. We will in particular talk about the interplay between the mathematical
content and the text that surrounds it.
3.1 Obey orthography

A mathematical text has to obey the rules of the language that is it written in
(e.g. English, French, German or Russian). It also means that a mathematical formula
is part of the surrounding text, and has to be treated as such.
3.2 Proper punctuation

3.2.1 A sentence ends with punctuation
Section 3.1 has an important consequence: If a sentence ends with a formula, then
this formula need to be followed by some kind of punctuation (in most cases by a full
point). The following example does it very wrong:
Example 3.1: Very wrong punctuation in an equation
1
It follows that
It follows that
2 \[
3 a^2 + b^2 = c^2
a2 + b2 = c2
4 \]
5
. This formula is important.
.
6 This formula is important.
The next example is also wrong:
47
Example 3.2: Wrong punctuation in an equation I
1 It follows that It follows that

2 \[
3 a^2 + b^2 = c^2 a2 + b2 = c2
4 \]
5 This formula is important. This formula is important.
The following is still wrong:

Example 3.3: Wrong punctuation in an equation II
1 It follows that: It follows that:

2 \[
3 a^2 + b^2 = c^2 a2 + b2 = c2
4 \]
The following example finally does it right:

Example 3.4: Right punctuation in an equation

2 \[
3 a^2 + b^2 = c^2. a2 + b2 = c2 .
4 \]
But it is even better if we add some slight spacing between the formula and the period.
Example 3.5: Best punctuation in an equation

2 \[
3 a^2 + b^2 = c^2 \,. a2 + b2 = c2 .
4 \]
This last approach is taken from [TSE14], which we strongly encourage the reader to
check out.
3.2.2 Punctuation in commutative diagrams?

There is some disagreement in the mathematical community about whether a commu-
tative diagram is allowed to include punctuation coming from the surrounding text.
The author is of the opinion that a commutative diagram should not contain any such
punctuation.
48
There are two standard ways to achieve this: One can finish up the sentence that
precedes a commutative diagram beforehand:
Example 3.6: Finishing the sentence before the commutative diagram
1 We consider the following commutative diagram:

2 \[
3 \begin{tikzcd}
4 A
5 \arrow{r}
6 \arrow{d}
7 &
8 A'
9 \arrow{d}
10 \\
11 C
12 \arrow{r}
13 &
14 C'
15 \end{tikzcd}
16 \]
17 The horizontal arrows in this diagram are isomorphisms.
We consider the following commutative diagram:
A A0
C C0
The horizontal arrows in this diagram are isomorphisms.
But it often better to incorporate the diagram in the surrounding sentence in such a
way that it contains no punctuation:
Example 3.7: Incorporating the commutative diagram in the sentence
1 In the commutative diagram

2 \[
3 \begin{tikzcd}
4 A
5 \arrow{r}
6 \arrow{d}
7 &
8 A'
9 \arrow{d}
10 \\
49
11 C
12 \arrow{r}
13 &
14 C'
15 \end{tikzcd}
16 \]
17 both horizontal arrows are isomorphisms.
In the commutative diagram
A A0
C C0
both horizontal arrows are isomorphisms.
3.2.3 Lists contain punctuation

Text that is organized using list environments still obeys the rules of punctuation.
Consider the following counterexample:
Example 3.8: Wrong punctuation in lists
1 A set $B$ is a basis of $V$ if

2 \begin{enumerate}
3 \item
4 $B$ is linearly independent
5 \item
6 $B$ is a generating set
7 \end{enumerate}
A set B is a basis of V if
1. B is linearly independent
2. B is a generating set
To figure out the correct punctuation simply remove the surrounding list and consider
the resulting text. In the above example this gives the following:
A set B is a basis of V if B is linearly independent B is a generating set
This is not a proper sentence, and should be the following instead:
A set B is a basis of V if B is linearly independent and B is a generating set.
The above counterexample should hence read as follows:
50
Example 3.9: Right punctuation in lists I
1 A set $B$ is a basis of $V$ if

2 \begin{enumerate}
3 \item
4 $B$ is linearly independent and
5 \item
6 $B$ is a generating set.
7 \end{enumerate}
A set B is a basis of V if
1. B is linearly independent and
2. B is a generating set.
There are also some other acceptable versions:

Example 3.10: Right punctuation in lists II
1 A set $B$ is a basis of $V$ if it satisfies the following conditions:

2 \begin{enumerate}
3 \item
4 $B$ is linearly independent.
5 \item
6 $B$ is a generating set.
7 \end{enumerate}
A set B is a basis of V if it satisfies the following conditions:

1. B is linearly independent.
2. B is a generating set.
3.2.4 Hyphen and dashes

Know your lines
In LATEX there are (at least) four line-like symbols which need to be distinguished,
the hyphen, the en dash, the em dash and the minus sign. See Table 3.1 for their
LATEX code and look. Each of these symbols plays its own role, and these roles depend
both on language and on convention. For mathematical writing in English the following
usages are important:
• For “n-dimensional” and “3-manifold”, use the hyphen.
• For combining names as in “Cauchy–Schwarz inequality” use the en dash.
51
name code output

hyphen - -
en dash -- –
em dash --- —
minus sign $-$ −
Table 3.1: Kinds of hyphen and dashes.
language name dash used

English Clebsch–Gordan coefficients en dash
German Clebsch-Gordan-Koeffizienten hyphen
French coefficients de Clebsch-Gordan hyphen
Dutch Clebsch-Gordan-coëfficienten hyphen
Russian Коэффициенты Клебша — Гордана em dash
Japanese クレブシュ–ゴルダン係数 en dash
Table 3.2: The Clebsch–Gordan coefficients in different languages.
• To interrupt a sentence–as done here–use the em dash without surrounding space,

or – if you’re feeling British – use the en dash surrounded by space.
For more information about the use of hyphen and dashes in the English language we
refer to [CMS17, 6.75–6.94].
To underline how language dependent the usage of hyphen and dashes is we’re
considering in Table 3.2 how the the Clebsch–Gordan are called in different languages,
as taken from Wikipedia. (We use the same font1 for all examples to allow a better
comparison.)
Don’t put a hyphen in math mode

A non-trivial amount of people makes the mistake of putting a hypen into math mode.
Example 3.11: A hyphen in math mode
1 Let~$A$ be a $\mathbb{Z}-$module.
Let A be a Z−module.
We can see that we don’t get a hyphen but a minus sign. We hence need to place the
symbol - outside of the math environment.
1 Namely “Noto Serif CJK JP”.
52
right aligned left aligned
x. x x. X x. x X. x
X. x X. X x. X X. X
Table 3.3: Spacings after a period (with a zoom of factor 3.5).
Example 3.12: Proper placement of the hyphen
1 Let~$A$ be a $\mathbb{Z}$-module.
Let A be a Z-module.
3.2.5 Correct spacing after a period

The problem
Not all periods are treated equally by LATEX: If a period is both followed by whitespace
and not preceded by an uppercase letter then LATEX will assume that this period is
supposed to end a sentence. It will some introduce some additional space following
this period. The effect of this can be seen in Table 3.3.
We can see from the first two columns of Table 3.3 (by considering the positions
of the periods) that the spacing in “x. *” is larger then the corresponding spacing
in “X. *”. This happens because LATEX thinks that the period in “x. *” is supposed to
end a sentence. We can also see from the first two tables that this difference of spacing
occurs both if the following letter is lower case and if it is upper case. We can see from
the third and fourth column that the amount of additional spacing does not depenend
on whether the following letter is in lower case or upper case.
The solution
We can use the command “\ ” – which we will write for readability as \(space) – to
tell LATEX that a preceding period is not meant to end a sentence. We can similarly use
the command \@ to tell LATEX that the following period is ending a sentence. Consider
the following example:
Example 3.13: Proper spacing after periods
1 Imagine now some old and long-dead English kings, e.g.\ Henry III and Henry
IV\@.
53
2 (Kings not named Henry are also okay.)
Imagine now some old and long-dead English kings, e.g. Henry III and Henry
IV. (Kings not named Henry are also okay.)
Note that in the above example one should also use ties ~ (as explained in Section 3.3.1)
to ensure that the Henries don’t lose their number.
Example 3.14: Proper spacing after periods, and using ~
1 Imagine now some old and long-dead English kings, e.g.\ Henry~III and
Henry~IV\@.
2 (Kings not named Henry are also okay.)
Imagine now some old and long-dead English kings, e.g. Henry III and Henry IV.
(Kings not named Henry are also okay.)
It should be pointed out that if a problematic period is followed by a closing

parenthesis, a closing brecket or quotations marks then then the problematic spacing
will still occur after these symbols. Lets consider the following example (where we use
the forbidden \\ to compare the versions).
Example 3.15: Spacings after periods followed by parentheses or quotation marks
1 Many colours (e.g. red, blue, etc.) are supported by \enquote{ColorX}. Even
purple! \\
2 Many colours (e.g.\ red, blue, etc.) are supported by \enquote{ColorX}. Even
purple! \\
3 Many colours (e.g.\ red, blue, etc.)\ are supported by \enquote{ColorX}. Even
purple! \\
4 Many colours (e.g.\ red, blue, etc.)\ are supported by \enquote{ColorX}\@. Even
purple!
Many colours (e.g. red, blue, etc.) are supported by “ColorX”. Even purple!
The problem of a sentence ending with an upper case letter does not occur if this
letter is in math mode.
Example 3.16: Upper case mathematics followed by a period
1 Let $f$ be an endomorphism of $X$.

2 Then $f$ is an isomorphism or $f = 0$.\\
54
3 Let $f$ be an endomorphism of $X$\@.

5 Let $f$ be an endomorphism of $X$.\
6 Then $f$ is an isomorphism or $f = 0$.
7 \\
8 Let $f$ be an endomorphism of $\operatorname{X}$.
10 Let $f$ be an endomorphism of $\operatorname{X}$\@.
12 Let $f$ be an endomorphism of $\operatorname{X}$.\
13 Then $f$ is an isomorphism or $f = 0$.
Let f be an endomorphism of X. Then f is an isomorphism or f = 0.

In view of Section 3.3.1 we get a problem: What if we want the space after an
abbreviation, as in “i.e. word”, not to be broken? Should we then write i.e.\ word
or i.e.~word?
In this case we write i.e.~word, since the tie ~ already contains the effect of \(space).
Consider for this the following example:
Example 3.17: Spacing of ~ and “\ ”
1 Let $M$ be a $2$-manifold, e.g. $M = S^2$-dimensional. \\

2 Let $M$ be a $2$-manifold, e.g.\ $M = S^2$-dimensional. \\
3 Let $M$ be a $2$-manifold, e.g.~$M = S^2$-dimensional.
Let M be a 2-manifold, e.g. M = S 2 -dimensional.

To summarize this section: As long as words are written, no problem occur. If an

abbreviation or some non-word things appear one should pay heed to periods which may
occur. The most of these special cases are covered by paying attention to “i.e. words”
and “e.g. words”.
We refer to [L2M19, 19.5.1, 19.6] for more information on this topic.
55
3.3 Text layout
3.3 Text layout

3.3.1 Use non-breakable space
If two expressions in text mode are separated by a space or a hyphen then LATEX may
separate these expressions in the output pdf-file by a line break. But such line breaks
are often undesirable and can sometimes be downright wrong.
• In abbreviations like “Prof. Dr. John Doe”.
• In combinations of words and numbers like “Lemma 5”.
• In combinations of mathematics and words like “3-dimensional”.
In such cases one needs to prevent the line break:
Line breaks at a space

No line break will occur if the tie symbol ~ is used instead of a space. Consider for this
the following example:
Example 3.18: An unwanted line break at a space
1 It follows from the above Lemma 5 It follows from the above Lemma 5
that the integer $2$ is prime. that the integer 2 is prime.
We don’t want a line break to occur in the term “Lemma 5” and thus introduce a
tie. In the above example it is also a good idea to prevent a possible line break in the
term “integer 2”. We therefore get the following code:
Example 3.19: Using ~ to prevent unwanted line breaks at spaces
1 It follows from the above Lemma~5 It follows from the above Lemma 5
that the integer~$2$ is prime. that the integer 2 is prime.
Line breaks at a hypen

To prevent a line break at a hyphen the command \nobreakdash needs to be placed in
front of it can be used. Consider the following example:
Example 3.20: An unwanted line break at a hyphen
1 The sphere $S^2$ will be our very first non-trivial example of a smooth
$2$-manifold (without boundary).
The sphere S 2 will be our very first non-trivial example of a smooth 2-manifold
(without boundary).
To prevent a line break at the hyphen we put \nobreakdash before it:
56
3.3 Text layout
Example 3.21: Using \nobreakdash to prevent unwanted line breaks at hyphens
1 The sphere $S^2$ will be our very first non-trivial example of a smooth
$2${\nbd}manifold (without boundary).
The sphere S 2 will be our very first non-trivial example of a smooth 2-manifold
(without boundary).
The command \nobreakdash works not only with hyphens but also with en dashes and
em dashes.
3.3.2 Don’t use \\ or \newline

The proper way to separate two succeeding paragraphs in LATEX is to leave an empty
line between them.
Example 3.22: New paragraphs done right
1 This is the first paragraph.

2 It consists of multiples lines to This is the first paragraph. It con-
make this example better. sists of multiples lines to make this
3
example better.
4 This is a second paragraph.
This is a second paragraph. It also
5
consists of multiple lines to make this
It also consists of multiple lines
to make this example even more
better.
example even more better.
6
This one is the third and last para-
7 This one is the third and last graph in this example. It also con-
paragraph in this example. sists of multiple lines.
8 It also consists of multiple lines.
Note that new paragraphs start indented.

The use of \\ or \newline doesn’t actually end a paragraph but instead forces LATEX
to do a line break.
Example 3.23: New “paragraphs” done wrong
1 This is the first paragraph in a This is the first paragraph in a coun-

counterexample.\\
terexample.
2 This is not a new paragraph.
This is not a new paragraph. The
3
text was just forced to start a new
The text was just forced to start a
new line.\newline
4 This one is also not a new
line.
paragraph. This one is also not a new paragraph.
5 But this whole thing looks ugly. But this whole thing looks ugly.
The use of \\ and \newline is only for starting new rows in matrices, tables and arrays.
Don’t use is to separate paragraphs.
57
3.3 Text layout
3.3.3 Text consisting of lists

The list environments enumerate and itemize do by default contain certain indentations,
that can be problematic in mathematical writing.
Identation compared to the left margin

There is by default some indentation between the left margin of the text area and the
item labels. This is meant to visually separate the list from the surrounding text. But
in mathematical writing there often occur lists that don’t have a surrounding text. The
most prominent example for this are propositions (and their proofs) that consist of
multiple statements:
Example 3.24: A list without surrounding text
1 \begin{theorem}
2 Let $G$ be a finite group.
3 \begin{enumerate}
4 \item
5 If $H$ is subgroup of $G$ then $\operatorname{ord}(H)$ divides
$\operatorname{ord}(G)$.
6 \item
7 If $a$ is any element of $G$ then $\operatorname{ord}(a)$ divides
8 \item
9 Every group of prime order is cyclic.
10 \item
11 If $p$ is a prime divisor of $\operatorname{ord}(G)$ then $G$ contains an
element of order $p$.
12 \end{enumerate}
13 \end{theorem}
Theorem 1. Let G be a finite group.
1. If H is subgroup of G then ord(H) divides ord(G).

2. If a is any element of G then ord(a) divides ord(G).
3. Every group of prime order is cyclic.
4. If p is a prime divisor of ord(G) then G contains an element of order p.
In such cases one should remove the identation between the left margin of the text
area and the item labels, as explained in Section 2.7.3.
58
3.3 Text layout
Example 3.25: A list without surrounding text
1 \begin{theorem}
2 Let $G$ be a finite group.
3 \begin{enumerate}[wide = 0pt, leftmargin=*]
4 \item
5 If $H$ is subgroup of $G$ then $\operatorname{ord}(H)$ divides
6 \item
7 If $a$ is any element of $G$ then $\operatorname{ord}(a)$ divides
8 \item
9 Every group of prime order is cyclic.
10 \item
11 If $p$ is a prime divisor of $\operatorname{ord}(G)$ then $G$ contains an
element of order $p$.
12 \end{enumerate}
13 \end{theorem}
Theorem 2. Let G be a finite group.

1. If H is subgroup of G then ord(H) divides ord(G).
2. If a is any element of G then ord(a) divides ord(G).
3. Every group of prime order is cyclic.

4. If p is a prime divisor of ord(G) then G contains an element of order p.
Theorems and proofs beginning with a list

If a theorem-like environment or a the environment proof begins with a list environment
then the first item of the list will not be aligned with the other list items:
Example 3.26: Improper aligned of list items
1 \begin{lemma}
2 \begin{enumerate}
3 \item
4 Every group~$G$ is isomorphic to its opposite group~$G^{\mathrm{op}}$.
5 \item
6 A group~$G$ equals its opposite group~$G^{\mathrm{op}}$ if and only if it
is abelian.
7 \end{enumerate}
8 \end{lemma}
59
3.3 Text layout
Lemma 3. 1. Every group G is isomorphic to its opposite group Gop .

2. A group G equals its opposite group Gop if and only if it is abelian.
One way to circumvent this problem is by inserting the command \leavevmode before
the beginning of the list.
Example 3.27: Proper aligned of list items with \leavevmode
1 \begin{lemma}
2 \leavevmode
3 \begin{enumerate}
4 \item
5 Every group~$G$ is isomorphic to its opposite group~$G^{\mathrm{op}}$.
6 \item
7 A group~$G$ equals its opposite group~$G^{\mathrm{op}}$ if and only if it
is abelian.
8 \end{enumerate}
9 \end{lemma}
Lemma 4.
1. Every group G is isomorphic to its opposite group Gop .

2. A group G equals its opposite group Gop if and only if it is abelian.
The problem with this approach is that it allows a line break to occur at the position
of \leavevmode. So it may happen that the header of the theorem-like environment of
proof environment occurs at the very bottom of the page while the list appears only on
the next page. (This looks quite horrible.)
As far as the author is aware there exists no good solution for this problem.2 A good
way of coping with this problem is to circumvent it by introducing some text before
the list environment:
Example 3.28: Proper aligned of list items by introducing text
1 \begin{lemma}
2 Let $G$ be a group with opposite group $G^{\mathrm{op}}$.
3 \begin{enumerate}
4 \item
5 The groups $G$ and $G^{\mathrm{op}}$ are isomorphic.
6 \item
2 Thepackage ntheorem is allegedly able to address this problem, at the price of introducing new
problems.
60
3.4 Mathematics in text
7 The groups $G$ and $G^{\mathrm{op}}$ are equal if and only if $G$ is
abelian.
8 \end{enumerate}
9 \end{lemma}
Lemma 5. Let G be a group with opposite group Gop .

1. The groups G and Gop are isomorphic.
2. The groups G and Gop are equal if and only if G is abelian.
An seen above we should also eliminate the indentation between the left margin of the
text area and the item labels:
Example 3.29: Proper aligned of list items by introducing text
1 \begin{lemma}
2 Let $G$ be a group with opposite group $G^{\mathrm{op}}$.
3 \begin{enumerate}[wide = 0pt, leftmargin=*]
4 \item
5 The groups $G$ and $G^{\mathrm{op}}$ are isomorphic.
6 \item
7 The groups $G$ and $G^{\mathrm{op}}$ are equal if and only if $G$ is
abelian.
8 \end{enumerate}
9 \end{lemma}
Lemma 6. Let G be a group with opposite group Gop .

1. The groups G and Gop are isomorphic.
2. The groups G and Gop are equal if and only if G is abelian.

3.4.1 Don’t break inline mathematics
When inline formulas or equations are too long or badly placed then it can happen
that a line break occurs, tearing the formula apart:
Example 3.30: Line break in inline math
1 The most important formula of all times is without a doubt is my mind $1 + 2 +

3 + 4 = 10 - 5$.
61
2 Truly a work of genius!
The most important formula of all times is without a doubt is my mind 1 + 2 +

3 + 4 = 10 − 5. Truly a work of genius!
This horrible affront to nature can thankfully be completely stopped by setting the
penalties \binoppenalty and \relpenalty to their maximum value, which can be accessed
via the command \maxdimen.
Example 3.31: Preventing inline math in inline math
1 % in the preamble:
2 \binoppenalty = \maxdimen
3 \relpenalty = \maxdimen
4 % in the main text:
5 The most important formula of all times is without a doubt is my mind $1 + 2 +
3 + 4 = 10 - 5$.
6 Truly a work of genius!
The most important formula of all times is without a doubt is my mind

1 + 2 + 3 + 4 = 10 − 5. Truly a work of genius!
3.4.2 Don’t begin a sentence with a mathematical symbol

Don’t begin a sentence with a mathematical symbol. Consider the following example:
Example 3.32: Mathematics at the beginning of a sentence
1 \begin{theorem}
2 $\mathbf{Mod}(R)$ is an abelian category.
3 \end{theorem}
Theorem 1. Mod(R) is an abelian category.

Example 3.33: No mathematics at the beginning of a sentence
1 \begin{theorem}
2 The category $\mathbf{Mod}(R)$ is abelian.
3 \end{theorem}
Theorem 2. The category Mod(R) is abelian.
62
An exception to the above rule are lists, in which short entries may start with a
math symbol. Consider for this the following example.
Example 3.34: List items beginning with mathematics
1 Let $V$ be a finite dimensional vector space.

2 Let $U$ and $W$ be two linear subspaces of $V$ of complementary dimensions,
i.e.\ such that $\dim V = \dim U + \dim W$.
3 Then the following conditions are equivalent:
4 \begin{enumerate}[label = \roman*)]
5 \item
6 $V = U \oplus W$,
7 \item
8 $V = U + W$,
9 \item
10 $V = U \cap W$.
11 \end{enumerate}
Let V be a finite dimensional vector space. Let U and W be two linear subspaces
of V of complementary dimensions, i.e. such that dim V = dim U + dim W .
Then the following conditions are equivalent:
i) V = U ⊕ W ,
ii) V = U + W ,
iii) V = U ∩ W .
3.4.3 Don’t put two formulas next to each other

One should not put two mathematical formulas right next to each other. Consider the
following example:
Example 3.35: Not properly separating mathematics
1 Then $f^2 = g$,$f$ is differentiable and $f = \exp(h)$.
Then f 2 = g,f is differentiable and f = exp(h).
Thu human eye (and brain) parses this sentence as follows:

Then f 2 = g,f is differentiable and f = exp(h).
But this betrays the actual content of the sentence. This problem can be fixed by
introducing some text. One solution is the following:
63
3.5 Text in mathematics
Example 3.36: Properly separating mathematics I
1 Then $f^2 = g$, the function $f$ is differentiable and $f = \exp(h)$.
Then f 2 = g, the function f is differentiable and f = exp(h).
This sentence is parsed as follows:

Then f 2 = g, the function f is differentiable and f = exp(h).
The following is another solution:
Example 3.37: Properly separating mathematics II
1 The function $f$ satisfies $f^2 = g$, it is differentiable and $f = \exp(h)$.
The function f satisfies f 2 = g, it is differentiable and f = exp(h).
This sentence is parsed as follows:

The function f satisfies f 2 = g, it is differentiable and f = exp(h).
3.4.4 Don’t begin a line with a mathematical symbol (optional)

One interesting idea for inline mathematics is to never start a line with a mathematical
symbol. This can be achieved by prefixing every inline mathematical formula by a tie ~.
The author has taken this approach from [Bel18].

3.5.1 Use \text
When text needs to be used in math mode the command \text is to be used.
Example 3.38: Syntax of \text
1 \text{normal text in math~mode}
Let’s start with the following horrible example:

Example 3.39: Missing text in math mode
1 Let $A \coloneqq \{ x \in [0,1] : x \; is \; rational \}$.
Let A := {x ∈ [0, 1] : x is rational}.
This next one looks a bit better, but is still not okay:
64
definition
name set
range of elements condition on elements
Figure 3.1: Structure tree of the definition.
:=
math
A {:}
math math
x ∈ [0, 1] x is rational
math text
Figure 3.2: Structure tree of the mathematical expression.
Example 3.40: Wrong way of using text in math mode
1 Let $A \coloneqq \{ x \in [0,1] : x \text{ is rational} \}$.
Instead do this:
Example 3.41: Right way of using text in math mode
1 Let $A \coloneqq \{ x \in [0,1] : \text{$x$ is rational} \}$.
Note that the right code reflects the structure of the displayed expression: Indeed, the
diagram in Figure 3.1 shows the structure of the given expression. To get the right
kind of code we assign to each node of this diagram the desired content, and whether
this node is mathematics or text. If a node contains both mathematics and text then
it is primary a text, which just happens to contain some mathematics. We arrive
at the diagram in Figure 3.2. By transforming this diagram node-wise into code we
arrive at Example 41. Note that both the nodes “x ∈ [0, 1]” and “x is rational” both
65
contain mathematics, but that the second is actually a text. It is therefore not proper
to combine these two math modes, as done in Example 40.
3.5.2 Put space around text in display mathematics

Sometimes a single line of displayed mathematics end with some text, as in the following
example.
Example 3.42: Wrong spacing around text in math mode I
1 Therefore
2 \[
3 g h g^{-1}
4 =
5 h
6 \text{for all $g, h \in G$}.
7 \]
Therefore
ghg −1 = hfor all g, h ∈ G.
In such a case one needs to put space between the mathematics and the text. This
space should not be part of the text:
Example 3.43: Wrong spacing around text in math mode II
1 Therefore
2 \[
3 g h g^{-1}
4 =
5 h
6 \text{ for all $g, h \in G$.}
7 \]
Therefore
ghg −1 = h for all g, h ∈ G.
Instead it should be put between the formula and the text. In the above example,
where a single block of mathematics is follows by a single block of text which contains
quantifiers for the previous formula, one should use the spacing \qquad:
Example 3.44: Right spacing around text in math mode I
1 Therefore
2 \[
3 g h g^{-1}
4 =
66
5 h
6 \qquad
7 \text{for all $g, h \in G$.}
8 \]
Therefore
ghg −1 = h for all g, h ∈ G.
Sometimes two blocks of mathematics are separated by a short text. In this case one
should use the spacing \quad on both sides of the text:
Example 3.45: Right spacing around text in math mode II
1 That $gh = hg$ can equivalently be expressed as

2 \[
3 g h g^{-1} = h
4 \quad\text{and}\quad
5 h g h^{-1} = g \,.
6 \]
That gh = hg can equivalently be expressed as
ghg −1 = h and hgh−1 = g .
It may even be appropriate to use \qquad on both sides when more space is needed.
3.5.3 Use \intertext and \shortintertext

Multi-line display mode environments like gather* and align* (see Section 5.1) can be
interrupted by inserting some text via the commands \intersect and \shortintertext
to insert some text between different lines of mathematics. This is particularly useful
can combine multiple align* environments into a single one, which then allows for a
common alignment of all lines.
The following is an example for what we don’t want.
Example 3.46: Two non-aligned blocks
1 We consider the equalities

2 \begin{align*}
3 H
4 &= a_1 + a_2 + a_3 + a_4 \\
5 &= b_1 + b_2 + b_3 + b_4 + b_5 + b_6 \\
6 &= c_1 + c_2 + c_3 + c_4 + c_5
7 \end{align*}
8 and
67
9 \begin{align*}
10 I
11 &= d_1 + d_2 + d_3 + d_4 + d_5 + d_6 \\
12 &= e_1 + e_2 + e_3 \\
13 &= f_1 + f_2 + f_3 + f_4 + f_5 \,.
14 \end{align*}
We consider the equalities
H = a1 + a2 + a3 + a4
= b1 + b2 + b3 + b4 + b5 + b6
= c1 + c2 + c3 + c4 + c5
and
I = d1 + d2 + d3 + d4 + d5 + d6
= e1 + e2 + e3
= f1 + f2 + f3 + f4 + f5 .
Note the annoying misalignment of the equality signs of the two align* environments.
We can solve this problem by using only one align* environment and inserting the
text “and” by using the command \intertext:
Example 3.47: Two aligned blocks
1 We consider now the equalities

2 \begin{align*} We consider now the equalities
3 H
4 &= a_1 + a_2 + a_3 + a_4 \\ H = a1 + a2 + a3 + a4
5 &= b_1 + b_2 + b_3 + b_4 + b_5 + = b1 + b2 + b3 + b4 + b5 + b6
b_6 \\
= c1 + c2 + c3 + c4 + c5
6 &= c_1 + c_2 + c_3 + c_4 + c_5
7 \intertext{and}
and
8 I
9 &= d_1 + d_2 + d_3 + d_4 + d_5 +
I = d1 + d2 + d3 + d4 + d5 + d6
d_6 \\
10 &= e_1 + e_2 + e_3 \\ = e1 + e2 + e3
11 &= f_1 + f_2 + f_3 + f_4 + f_5 \,. = f1 + f2 + f3 + f4 + f5 .
12 \end{align*}
The command \shortintertext works in the same way as \intertext but inserts less
vertical space around it. To see the effects of \shortintertext let us start with the
following example:
68
Example 3.48: Two non-aligned equalities
1 First we have some lengthy text above the mathematical formulas, followed by
the formulas
2 \[
3 a = b
4 \]
5 and
6 \[
7 b = a \,.
8 \]
9 We finish off by writing some more text after the above mathematical two
formulas.
First we have some lengthy text above the mathematical formulas, followed by
the formulas
a=b
and
b = a.
We finish off by writing some more text after the above mathematical two
formulas.
We naturally want to align the two equality signs, which we by using a single \align* en-
vironment together with \intertext, as explained above:
Example 3.49: Two aligned equalities with too much space
the formulas
2 \begin{align*}
3 a &= b
4 \intertext{and}
5 b &= a \,.
6 \end{align*}
formulas.
the formulas
a=b
and
b = a.
69
3.6 Mathematics as text
formulas.
But we can see that the inserted \intersect puts a large space between the two formulas,
far too large for our taste. We can solve this problem by using \shortintertext instead
of \intertext, as the following example shows:
Example 3.50: Two aligned equalities with proper space
the formulas
2 \begin{align*}
3 a &= b
4 \shortintertext{and}
5 b &= a \,.
6 \end{align*}
formulas.
the formulas
a=b
and
b = a.
formulas.

3.6.1 Don’t abuse math as text
Again similar to Section 3.6.2 one should not abuse mathematical symbols as words:
A mathematical symbol can often replace parts of a sentence, in particular a verb.
Example 3.51: Right way of using mathematical symbols as words
1 Then $a = b$ and thus $b \leq c$.
Then a = b and thus b ≤ c.
But this use of mathematics should be treated with care, and not abused:
70
Example 3.52: Wrong way of using mathematical symbols as words
1 The integer $a$ is $< 0$ and thus $\notin \mathbb{N}$.
The integer a is < 0 and thus ∈

/ N.
3.6.2 Logical symbols don’t replace text

Mathematicians often abuse logical symbols like
∃, ∀, =⇒ , ⇐⇒
as replacements for written out words. This is okay in handwriting and on blackboards,
but it is not okay in a properly written out mathematical text.
Example 3.53: Abuse of quantifiers
1 It follows that $\forall y \in B$ that $\exists x \in A$ with $f(x) = y$.
It follows that ∀y ∈ B that ∃x ∈ A with f (x) = y.
One needs to instead write out the used quantifiers.

Example 3.54: Writing out quantifiers
1 It follows that for every element $y \in B$ there exists an element $x \in A$
with $f(x) = y$.
It follows that for every element y ∈ B there exists an element x ∈ A with

f (x) = y.
In this example one should also write out the element relation:
Example 3.55: Best use of quantifiers
1 It follows that for every element $y$ of $B$ there exists an element $x$ of $A$
with $f(x) = y$.
It follows that for every element y of B there exists an element x of A with

f (x) = y.
The use of the above logical symbols is of course okay if they are used in their actual
logical function:
71
Example 3.56: Okay way to use logical symbols
1 Let $f \colon X \to Y$ be a morphism in a category $\mathcal{C}$.

2 Then
3 \begin{align*}
4 {}&
5 \text{$f$ is an epimorphism}
6 \\
7 \iff{}&
8 \bigl[
9 \forall Z \in \operatorname{Ob}(\mathcal{C}):
10 \forall g, h \in \mathcal{C}(Y,Z):
11 g \circ f = h \circ f \implies g = h
12 \bigr]
13 \\
14 \iff{}&
15 \bigl[
16 \forall Z \in \operatorname{Ob}(\mathcal{C}):
17 \text{$f^* \colon \mathcal{C}(Y,Z) \to \mathcal{C}(X,Z)$ is injective}
18 \bigr]
19 \\
20 \implies{}&
21 \text{$f^* \colon \mathcal{C}(Y,-) \to \mathcal{C}(X,-)$ is a monomorphism}.
22 \end{align*}
Let f : X → Y be a morphism in a category C. Then
f is an epimorphism
⇐⇒ ∀Z ∈ Ob(C) : ∀g, h ∈ C(Y, Z) : g ◦ f = h ◦ f =⇒ g = h

⇐⇒ ∀Z ∈ Ob(C) : f ∗ : C(Y, Z) → C(X, Z) is injective

=⇒ f ∗ : C(Y, −) → C(X, −) is a monomorphism.
3.6.3 Don’t use “iff”

Similarly to Section 3.6.2 abbreviations like “iff” need to be written out in a mathe-
matical text.
72
Chapter 4
Good behaviour inside math mode
4.1 Using the right commands and symbols
4.1.1 General symbols
Use the right symbols. Table 4.1 shows some popular sources of this problem. Note
that the commands \rightarrow and \to give the same arrow. So use whichever is
more appropriate in the given situation.
4.1.2 Operations
Many mathematical operations have both a binary version and an operator version
where the operation can range over some index set. One should not confuse the two of
them. Table 4.2 shows some popular binary operations and their operator counterpart.
The command \bigsqcap requires the package stmaryrd.
4.1.3 Negations
For some mathematical symbols there also exists a negated version, which is formed
by diagonal line through the symbol. A general way of introducing such a line is the
command \not.
Example 4.1: Using \not
1 It follows that $A \not\ni x$. It follows that A 63 x.
But in practice this command should seldom be used, as it often produces very bad
looking output. Consider the following example:
Example 4.2: Why not to use \not
1 Hence $A \not\implies B$.
Hence A 6 =⇒ B.
There are two solutions to this problem:
73
symbol right commands wrong commands

element relation \in ∈ \epsilon
\varepsilon ε
\ni 3 \backepsilon
empty set \emptyset ∅ \phi φ
\varnothing ∅
set difference A \setminus B A\B A \backslash B A\B
A \smallsetminus B ArB
A - B A−B
implication \implies =⇒ \Rightarrow ⇒
=> =>
\impliedby ⇐= \Leftarrow ⇐
<= <=
equivalence \iff ⇐⇒ \Leftrightarrow ⇔
definition \coloneqq := := :=
\eqqcolon =: =: =:
norm \| x \| kxk || x || ||x||
\lVert x \rVert
pointy brackets \langle x \rangle hxi < x > <x>

infinity \infty ∞ oo oo
function colon f \colon X \to Y f: X →Y f : X \to Y f :X→Y
Table 4.1: Right symbols and wrong symbols.
Many symbols already have a predefined lined-out version available. A few of

them are collected in Table 4.3. If the required symbol is not predefined then the
command \centernot from the package centernot often produces better looking output
then \not:
Example 4.3: Using the command \centernot
1 Consider
2 \[
3 A \centernot\ni x
4 \quad\text{versus}\quad
5 \quad
6 A \not\ni x \,.
7 \]
8 Consider also
74
operation binary generalized
sum
P
+ a+b \sum i xi
multiplication
Q
\cdot a·b \prod i xi
direct sum
L
\oplus A⊕B \bigoplus i Xi
tensor product
N
\otimes A⊗B \bigotimes i Xi
wedge
V
\wedge a∧b \bigwedge i Xi
union
S
\cup A∪B \bigcup i Xi
intersection
T
\cap A∩B \bigcap i Xi
product
Q
\times A×B \prod i Xi
d
\sqcap AuB \bigsqcap i Xi
coproduct
`
\amalg AqB \coprod i Xi
F
\sqcup AtB \bigsqcup i Xi
Table 4.2: Binary operation and operator version.
right wrong
\notin ∈
/ \not\in 6∈
\nexists @ \not\exists 6∃
\neq 6 = \not = 6=
\nleq \not\leq 6≤
\nrightarrow 9 \not\rightarrow 6 →
Table 4.3: Negated versions of popular symbols.
9 \[
10 A \centernot\implies B
11 \quad\text{versus}\quad
12 A \not\implies B \,.
13 \]
Consider
A 63 x versus A 63 x .
Consider also
A =⇒
6 B versus A 6 =⇒ B .
75
4.1.4 Don’t use \limits

People seem to think that they have to put\limits after a command to add limits to it:
Example 4.4: Using limits
1 \[
2 \sum\limits_{k=1}^n k n
X n(n + 1)
3 = k=
4 \frac{n(n+1)}{2}
2
k=1
5 \]
But this is not only unnecessary, but also dangerous. It is unnecessary because (most
of) the commands in questions already have this functionality built in:
Example 4.5: Using built-in limits
1 \[
2 \sum_{k=1}^n k n
X n(n + 1)
3 = k=
4 \frac{n(n+1)}{2}
2
k=1
5 \]
This built-in limits have the advantage that they can distinguish between inline math
and display math math:
Example 4.6: Inline math vs. display math with built-in limits
The sum k=0 2k = 2n+1 −1 is inline

1 The sum $\sum_{k=0}^n 2^k = 2^{n+1} Pn
while the sum
- 1$ is inline while the sum
2 \[
3 \sum_{k=0}^n 3^k n
X
4 \neq 3k 6= 3n+1 − 1
5 3^{n+1} - 1 k=0
6 \]
7 is in display mode. is in display mode.
We can see that the inline version does not only use the smaller summation sign but
also sets the limits to the right of the summation sign. This is a feature which \limits
version is missing:
76
Example 4.7: Inline math with \limits
1 So here is some text which will

generate some lines. So here is some text which will gen-
2 The text itself ’isnt important,
erate some lines. The text itself isn’t
but we really want it to fill some
important, but we really want it to
lines.
fill some lines. The important thing
3 The important thing is the sum n
is the sum k. Well, not really the
P
$\sum\limits_{k=0}^n k$.
4 Well, not really the sum itself, k=0
but the use of sum itself, but the use of \limits for
\texttt{{\textbackslash}limits} for typesetting it..
typesetting it..
The limits are still placed above the top and below the summation sign. This has its
price: The line in which the sum resides breaks the usual vertical space between lines,
which gives the text an inconsistent and unorganized look. Compare this to the version
without \limits:
Example 4.8: Inline with built-in limits
1 So here is some text which will

generate some lines. So here is some text which will gen-
2 The text itself ’isnt important,
erate some lines. The text itself isn’t
but we really want it to fill some
important, but we really want it to
fill some lines.
Pn The important thing
lines.
3 The important thing is the sum
$\sum_{k=0}^n k$.
is the sum k=0 k. Well, not really
4 Well, not really the sum itself, the sum itself, but use of built-in lim-
but use of built-in limits for its for typesetting it.
typesetting it.
Here the line distance is nicely consistent and pleasing to the eye.
The usual predefined commands on which one would expect limits already have them
defined, e.g. \sum, \prod or \lim, as the following example shows:
Example 4.9: Inline vs. display for different commands with built-in limits
1 Compare the inline versions $\sum_{k=0}^n k$ and $\prod_{k=1}^n k$ and $\lim_{n

\to \infty} a_n$ with the display versions
2 \[
3 \sum_{k=0}^n k \,,
4 \quad
5 \prod_{k=1}^n k \,,
6 \quad
7 \lim_{n \to \infty} a_n \,.
8 \]
77
Compare the inline versions k=0 k and k=1 k and limn→∞ an with the display
Pn Qn
versions
n n
lim an .
X Y
k, k,
n→∞
k=0 k=1
When now commands are defined with \DeclareMathOperator or \operatorname then

one can make them support limits by using \DeclareMathOperator* and \operatorname*
instead. Suppose for example that we have made in the preamble the following definition:
Example 4.10: Using \DeclareMathOperator* to define \colim
1 \DeclareMathOperator*{\colim}{colim}
We can then do the following:

Example 4.11: Using \colim
1 Inline we have $\colim_{X' \leq X} F(X')$ and in display mode we get

2 \[
3 \colim_{X' \leq X} F(X') \,.
4 \]
Inline we have colimX 0 ≤X F (X 0 ) and in display mode we get
colim
0
F (X 0 ) .
X ≤X
The command \operatorname* was named \operatornamewithlimits in the past, but

this name is deprecated.
4.1.5 Use extensible arrows instead of \overset and \underset

Some people put text above or under arrows by wrongly using the commands \overset
or \underset:
Example 4.12: Using \overset and \underset to put text above or below an
arrow
1 \[
2 X
3 \overset{f}{\longrightarrow}
4 Y
5 \underset{g \circ h \circ k}{\longrightarrow}
6 Z
7 \]
78
f
X −→ Y −→ Z
g◦h◦k
We can see above that the length of the arrow does not adjust to the size of the text
above it or below it. The proper way to put text on top of an arrow or bellow an arrow
of the form “→” is therefore the command \xrightarrow:
Example 4.13: Using \xrightarrow to put text above or below an arrow
1 \[
2 X
3 \xrightarrow{f}
4 Y
5 \xrightarrow[g \circ h \circ k]{}
6 Z \,.
7 \]
f
X−
→ Y −−−−→ Z .
g◦h◦k
The package amsmath defines only the two most basic extensible arrows. Many more
kinds of extensible arrows are provided by the package mathtools and some more are
contained in the package extarrows. An overview of the various kinds of extensible
arrows can be found in Table 4.4.
The author recommends to define custom commands as shortcuts for the most used
arrows.
Example 4.14: Defining arrow commands as shortcuts
1 \newcommand{\xto}{\xrightarrow}
2 \newcommand{\xlongto}[1]{\xlongrightarrow{\;#1\;}}
3 The function $A \xto{f} B$ is the same as
4 \[
5 A \xlongto{f} B \,.
6 \]
The function A −
→ B is the same as
f
f
A−
−→ B .
4.1.6 Know your ellipses

An ellipsis should never by written as ... (three single dots). LATEX instead provides
different kinds of ellipses to use in math mode, namely
79
f f
amsmath A−
→B A←
−B
\xrightarrow \xleftarrow
A=
f f f
mathtools A 7−
→B A←
→B ⇒B
\xmapsto \xleftrightarrow \xRightarrow
f
=B
f f
A⇐ A⇐
⇒B A←
−- B
\xLeftarrow \xLeftrightarrow \xhookleftarrow
f f f
→B
A ,− A−
+B A−
*B
\xhookrightarrow \xrightharpoondown \xrightharpoonup
f f f
−
*
A)
−B −B
A) −B
A(
\xrightleftharpoons \xleftharpoondown \xleftharpoonup
f
−B
A(
−
+
\xleftrightharpoons
A == B
f f f
extarrows A←
→B A⇐
⇒B
\xlongequal \xleftrightarrow \xLeftrightarrow
f f f
A ←−
−B A −→ B A ←−
−→ B
\xlongleftarrow \xlongrightarrow \xlongleftrightarrow
=B A =⇒ B =⇒ B
f f f
A ⇐= A ⇐=
\xLongleftarrow \xLongrightarrow \xLongleftrightarrow
Table 4.4: Extensible arrows in amsmath, mathtools and extarrows.
\dotsb, \dotsc, \dotsm, \dotsi, \dotso,

\cdots, \ddots, \vdots, \ldots.
Each of these have they own role and (mostly) distinct look and feel. There are two
groups of ellipses. The first group consists of \dotsb, \dotsc, \dotsm, \dotsi,\dotso,
whereas the second group consists of \cdots, \ddots, \vdots, \ldots.
The ellipses in the first group are named after their function, see Table 4.5. Note
that the members of this group all have names of the form \dots*, where * is a letter
that specifies the semantic use of the dots.
The ellipses in the second group are not named after their function but after their
orientation, see Table 4.6 The ellipses \cdots, \ddots and \vdots are typically used in
matrices.
80
command where to use example
\dotsb Between binary relations and operations. x1 + · · · + xn

x1 ≤ · · · ≤ xn
\dotsc Between commata. x1 , . . . , x n
\dotsm Abbreviating multiplication. x1 · · · xn
Iterated integrals.
R R
\dotsi X1
· · · Xn
\dotso Others.
Table 4.5: Ellipses with a specific functions.
command description look
\cdots Horizontal dots, vertically centered. ···

..
\ddots Diagonal dots. .
..
\vdots Vertical dots, horizontally centered. .
\ldots Lowered dots. ...
Table 4.6: Ellipses with a specific orientation.
Example 4.15: Typical usage of the ellipses \cdots, \vdots, \ddots
1 \[
2 \begin{pmatrix}
3 a_{11} & \cdots & a_{1n} \\
4 \vdots & \ddots & \vdots \\
5 a_{n1} & \cdots & a_{nn}
6 \end{pmatrix}
7 \]
 
a11 ··· a1n
 .. .. .. 
 . . . 
an1 ··· ann
The command \ldots may be used to denote left out digits or symbols.
81
Example 4.16: Using the ellipsis \ldots
1 We find that
2 \[
3 x = 0.1234567891011\ldots
4 \]
5 Consider the word $w = a_1 \ldots a_n$ where $a_1, \dotsc, a_n$ are letters in
an alphabet $\Sigma$.
We find that
x = 0.1234567891011 . . .
Consider the word w = a1 . . . an where a1 , . . . , an are letters in an alphabet Σ.
Note that the members of this second group of dots all have names of the form \*dots,
where * is a letter that specified the positioning of these dots.
So overall one should use the ellipses \cdots, \ddots and \vdots for matrices, and
otherwise the ellipses \dotsb, \dotsc, \dotsm, \dotsi and occasionally \dotso.
There also exists the generic command \dots which tries to automagically use the
right kind of positioning and spacing. But the author recommends no using this
command as it can lead to inconsistent results:
Example 4.17: Inconsistent results with \dots
1 The ellipses in $x_1 \leq x_2 \leq \dots \leq x_n$ and $y_1 \leq y_2 \leq
\dots$ should look the same.
The ellipses in x1 ≤ x2 ≤ · · · ≤ xn and y1 ≤ y2 ≤ . . . should look the same.
4.1.7 Know your matrices

The package amsmath provides various environments for matrices, the most basic of
which is the environment matrix.
Example 4.18: The basic matrix environment matrix
1 \[
2 \begin{matrix}
3 a & b \\ a b
4 c & d c d
5 \end{matrix}
6 \]
There are five more variants of this basic matrix environment, that put different
delimiters around the matrix. The package mathtools also provides small versions of
all six kinds of matrices. See Table 4.7 for a table of all these kinds of matrices.
82
size matrices

a b a b a b
normal c d c d c d
matrix pmatrix bmatrix

a b a b

a b

c d c d c d
Bmatrix vmatrix Vmatrix
a b a b
a b
small c d c d c d
smallmatrix psmallmatrix bsmallmatrix
a b
a b a b
c d c d c d
Bsmallmatrix vsmallmatrix Vsmallmatrix
Table 4.7: Kinds of matrices.
The package mathtools also provides starred versions of the matrices in Table 4.7.
These starred versions allow to specify as an optional argument the alignment of the
columns. The three possible arguments are l for left-alignment, c for centered alignment
and r for right-alignment. The standard alignment is c.
Example 4.19: Aligning matrix entries
1 \[
2 \begin{bmatrix*}[l]
3 a & -b \\
4 -c & d
5 \end{bmatrix*}
6 \qquad
7 \begin{bmatrix*}[c]
8 a & -b \\
9 -c & d
10 \end{bmatrix*}
11 \qquad
12 \begin{bmatrix*}[r]
13 a & -b \\
14 -c & d
15 \end{bmatrix*}
16 \qquad
17 \begin{bmatrix}
18 a & -b \\
19 -c & d
83
4.2 Avoid bad notation
20 \end{bmatrix}
21 \]

a −b a −b a −b a −b
−c d −c d −c d −c d
One should choose the kind of matrix, its size and alignment dependent on convention,
usage and the matrix entries.
Example 4.20: Using the right kind of matrix
1 We consider the morphism

2 \[
3 X \oplus Y
4 \xrightarrow{\,
5 \begin{bsmallmatrix*}[r] We consider the morphism
6 f & -d \\
f −d
h i
7 0 & g
0 g
8 \end{bsmallmatrix*} X ⊕ Y −−−−−→ X 0 ⊕ Y 0 .
9 \,
10 }
11 X' \oplus Y' \,.
12 \]

4.2.1 Don’t use \subset
Some people use ⊂ (\subset) to denote inclusion and ( (\subsetneq) to denote proper
inclusion, while some other people use ⊆ (\subseteq) to denote inclusion and ⊂ to
denote proper inclusion. The second convention has the advantage of making sense
and being consistent with the usual use of ≤ and <, whereas the first convention has
the non-advantage of existing. The problem is that both conventions are wide-spread
but use the symbol ⊂ is different ways.
The conflict between the above two conventions has abused the symbol ⊂ to a point
that it should simply never be used. Instead one should always use ⊆ for inclusion, (
for proper inclusion and * (\nsubseteq) for non-inclusion.
Some people prefer the symbols j (\subseteqq) and $ (\subsetneqq) instead. The
author thinks that these symbols are unnecessary large and recommends not to use
them. (But they do at least leave poor old ⊂ alone.)
84
4.2.2 Don’t underline

Underlining mathematics works well on the blackboard, in handwriting and it was
done in the age of typewriters. Don’t do it in LATEX.
4.2.3 Use dx instead of dx

A differential is written as \mathrm{d}x, not as dx. When it occurs at the end of an
integral then a slight spacing \ is also introduced in front of it. So don’t do the following:
Example 4.21: Wrong kind of differentials
1 \[
2 \int_a^b f(x) dx Z b Z b
3 \quad f (x)dx f (x)dx
4 \int_a^b f(x) \mathrm{d}x a a
5 \]

Example 4.22: Right kind of differential
1 \[ Z b
2 \int_a^b f(x) \,\mathrm{d}x f (x) dx
3 \] a
4.2.4 Don’t force fancy fractions

When fractions are placed inline, as an exponent or in an index then they should be of
the form a/b. The notation
a
b
is reserved for display style. So don’t do the following:
Example 4.23: Full fractions in exponent, index and inline
1 Consider $e^{\frac{1}{x}}$ and $x_{\frac{1}{n}}$ and $\frac{2}{3}$.
Consider e x and x n1 and 23 .

1
Do the following instead:

Example 4.24: Flat fractions in exponent, index and inline
1 Consider $e^{1/x}$ and $x_{1/n}$ and $2/3$.
85
4.3 Defining new commands
Consider e1/x and x1/n and 2/3.
Don’t use funky fractions like ab. They’ll make you go blind and burn down your
house.

Many mathematical operators have predefined commands, which should be used.
Example 4.25: Using commands vs. not using them
1 Use $\sin(x)$ instead of $sin x$, use $\dim V$ instead of $dim V$ and use
$\lim_{n \to \infty} a_n$ instead of $lim_{n \to \infty} a_n$.
Use sin(x) instead of sinx, use dim V instead of dimV and use limn→∞ an
instead of limn→∞ an .
New commands can be defined in various ways:
4.3.1 \DeclareMathOperator
Commands of the form \Word that are supposed to give the output Word can easily be
defined using the command \DeclareMathOperator.
Example 4.26: Syntax of \DeclareMathOperator
1 \DeclareMathOperator{\Word}{Word}
The command \DeclareMathOperator can only be used in the preamble.

To define the command \Hom we us the following text in the preamble:
Example 4.27: Declaring a mathematical operator with \DeclareMathOperator
2 \DeclareMathOperator{\End}{End}
The command \Hom can then be used in the usual way:

Example 4.28: Using a declared mathematical operator
1 Thus $\End(V) = \End_k(V)$ becomes a vector space.
Thus End(V ) = Endk (V ) becomes a vector space.
86
When a command \Word is defined with \DeclareMathOperator then LATEX will auto-
matically insert some space around \Word when needed:
Example 4.29: Automatic spacing of \DeclareMathOperator
1 \begin{align*}
2 &x \End V
3 \\ x End V
4 &x \End(V)
x End(V )
5 \\
6 &x \End {(V)} x End (V )
7 \end{align*}
Note that in the first expression LATEX inserts some spacing both to the left and to the
right of End. In the second expression LATEX observes that the used math operator is
follows by a parenthesis and thus inserts no additional spacing. For the third expression
we prevent LATEX from making such an observation by using a pair of curly brackets.
This behavior leads to a bad looking output when \DeclareMathoperator is abused.
Suppose that we want a command \Complex that inserts the code \mathbb{C}.
Example 4.30: Wrong way of using \DeclareMathOperator
2 \DeclareMathOperator{\Complex}{\mathbb{C}}
This will lead to the following problem:

Example 4.31: Wrong output when \DeclareMathOperator is abused
1 The span of $x_1, \dotsc, x_n \in \Complex^m$ equals $\Complex x_1 + \dotsb +
\Complex x_n$.
The span of x1 , . . . , xn ∈ Cm equals C x1 + · · · + C xn .
We expect the output Cx1 + · · · + Cxn but get some unwanted spacing instead.
4.3.2 \operatorname
The command \operatorname can be used to give the formatting of a mathemical
operator without defining a new command.
Example 4.32: Using \operatorname
1 Thus $\operatorname{Hom}(V,W) = \operatorname{Hom}_k(V,W)$ becomes a vector

space.
Thus Hom(V, W ) = Homk (V, W ) becomes a vector space.
87
If the same command is used multiple times then one should use \DeclareMathOperator
instead of \operatorname, to keep the code clean.
4.3.3 Don’t abuse \mathrm

The commands \mathrm and \operatorname do not give the same formatting. With
\operatornamewe get the necessary spacing when not using parentheses, which does
not happen when using \mathrm.
Example 4.33: Missing spacing after \mathrm
1 Compare $\operatorname{End} V$ to $\mathrm{End} V$.
Compare End V to EndV .
4.3.4 \newcommand
A very general way of defining new commands is given by \newcommand. Its syntax is as
follows:
Example 4.34: Syntax of \newcommand
1 \newcommand{\name}[number of arguments n]{ definition including #1, ..., #n }

Example 4.35: Using the command \newcommand
1 \newcommand{\bimodule}[2]{#1-#2-bimodule}
2 Let $M$ be an \bimodule{$A$}{$B$}.
Let M be an A-B-bimodule.
One may think about \DeclareMathOperator as a combination of \newcommand and

\operatorname:
Example 4.36: \DeclareMathOperator = \newcommand plus \operatorname
1 \newcommand{\Ouv}{\operatorname{Ouv}}
2 $\Ouv X$
Ouv X
Trying to define an already existing command with \newcommand will lead to an error.
To overwrite an already existing command one can use \renewcommand instead. But
this shouldn’t really be used (unless you really, really know what you’re doing): Even
88
if you don’t like a particular command there is a chance that some package that you’re
using relies on it. So overwriting a command can easily surprise you with some new
problems.
4.3.5 \DeclarePairedDelimiter
Mathematical operations like the absolute value |·| and a norm k·k are denoted by putting
certain delimiters to the left and right of the argument. To define a corresponding LATEX
command like \abs or \norm one should use the command \DeclarePairedDelimiter
(which is provided by the package mathtools):
Example 4.37: Syntax of \DeclarePairedDelimiter
1 \DeclarePairedDelimiter{\name}{left delimiter}{right delimiter}
As an example we use \DeclarePairedDelimiter to define a command \abs for absolute

value:
Example 4.38: Defining the command \abs with \DeclarePairedDelimiter
1 \DeclarePairedDelimiter{\abs}{\lvert}{\rvert}
The defined command can now be used as expected:

Example 4.39: Using a declared delimiter
1 \[
2 \abs{-5} = 5
3 \]
|−5| = 5
Declaring a command \delim via \DeclarePairedDelimiter will automatically also

define starred version \delim* that scales the surrounding delimiters according to the
given content between them. One can also specify a scaling size like \big, \bigg, etc.
to scale the delimiters.
Example 4.40: Scaling of declared delimiters
1 \begin{align*}
2 \abs{-\frac{1}{2}}
3 &=
4 \frac{1}{2} \,,
5 \\
6 \abs*{-\frac{1}{2}}
7 &=
8 \frac{1}{2} \,,
89
9 \\
10 \abs[\bigg]{-\frac{1}{2}}
11 &=
12 \frac{1}{2}
13 \end{align*}
1 1
|− | = ,
2 2
1 1
− = ,
2 2

1 1
− =
2 2
The variant \DeclarePairedDelimiterX allows building more sophisticated commands:

Example 4.41: Syntax of \DeclarePairedDelimiterX
1 \DeclarePairedDelimiterX{\name}
2 [number of arguments n]
3 {left delimiter}{right delimiter}
4 {expression build from #1, #2, ..., #n}
If the built up expression contains a delimiter, and by prefixing this delimiter with the
command \delimsize will ensure that the delimiter scales in the same way as the two
surrounding delimiters. As an example we define a command \inner for inner product.
Example 4.42: Using \DeclarePairedDelimiterX for more advanced delimiters
1 \DeclarePairedDelimiterX{\inner}[2]{\langle}{\rangle}{#1 \,\delimsize\vert\, #2}
The defined command \inner takes two arguments and inserts a line (\vert) with some
surrounding space (\,) between them.
Example 4.43: Using more advanced delimiters
1 \[
2 \inner{\psi_1}{\psi_2}
3 \quad
4 \inner*{\frac{f}{g}}{\frac{h}{k}}
5 \]

f h
hψ1 | ψ2 i
gk
90
There is also the variant \DeclarePairedDelimiterXPP that is even more flexible than
\DeclarePairedDelimiterX.
4.3.6 The package xparse

A useful way for defining commands – in particular more involved ones – is provided by
the package xparse. This package provides the command \NewDocumentCommand, which
can be used to define new commands.
Example 4.44: Syntax of \NewDocumentCommand
1 \NewDocumentCommand{\name}{arguments}{definition}
The field arguments specifies what kinds of arguments the command \name will take:
Mandatory arguments
Mandatory arguments can be declared with the option m:
Example 4.45: Mandatory arguments with \NewDocumentCommand II
1 % in the preamble
2 \NewDocumentCommand{\double}{m}{#1 #1}
4 \double{sometext}
sometext sometext
One can specify multiple arguments:

Example 4.46: Mandatory arguments with \NewDocumentCommand II
1 % in the preamble
2 \NewDocumentCommand{\swap}{m m}{#2 #1}
4 \swap{first}{second}
second first
Optional arguments without default value

Optional arguments can be declared with the option o. It then needs to be checked
with the command \IfNoValueTF
91
Example 4.47: Syntax of \IfValueTF
1 \IfValueTF{#number of argument}
2 {if the argument has been set}
3 {if the argument has not been set}
One can similarly use \IfNoValueTF instead of \IfNoValueTF. This has the same effect
as switching the cases in Example 47.
Example 4.48: Optional arguments with \NewDocumentCommand, I
1 % in the preamble
2 \NewDocumentCommand{\module}{m o}{\IfNoValueTF{#2}{#1-module}{#1-#2-bimodule}}
4 Let $M$ be an \module{$R$} and let $N$ be an \module{$R$}[$S$].
Let M be an R-module and let N be an R-S-bimodule.
Optional arguments with default value

One can use O{default value} instead of o to declare an optional argument that has a
default value.
Example 4.49: Optional arguments with \NewDocumentCommand, II
1 % in the preamble
2 \NewDocumentCommand{\restrict}{m m O{}}{#1|_{#2}^{#3}}
4 Let $f = \restrict{g}{X}$ and $f' = \restrict{g'}{X}[Y]$
Let f = g|X and f 0 = g 0 |YX
Starred versions
One can use the argument * together with the command \IfBooleanTF to check for the
occurrence of a star. This can be used to also define a starred version of a command.
Example 4.50: Starred versions with \NewDocumentCommand
1 % in the preamble
2 \NewDocumentCommand{\choice}{s m m}{\IfBooleanTF{#1}{#3}{#2}}
3 % in the main text’
4 Dont confuse \choice{first}{second} with \choice*{first}{second}.
Don’t confuse first with second.
92
4.4 Stretch your arrows
name tiny small scriptsize normal large huge

distance 0.6em 1.2em 1.8em 2.4em 3.6em 4.8em
Table 4.8: Standard distances for column sep.
4.4 Stretch your arrows

If some expression occurs atop or below an arrow, then this arrow must be stretched
suffciently long to accommodate these expression. The extensible arrows introduced in
Section 4.1.5 automatically do so. If an arrow in a commutative diagram isn’t long
enough then this arrow must also be stretched:
Example 4.51: Commutative diagram with an arrow too short
1 \[
2 \begin{tikzcd}
3 X \arrow{r}{f \circ g - g \circ f} \arrow{d}
4 &
5 Y \arrow{r}{k} \arrow{d}
6 &
7 Z \arrow[equal]{d}
8 \\
9 X' \arrow[dashed]{r}{h'}
10 &
11 Y' \arrow{r}{k'}
12 &
13 Z'
14 \end{tikzcd}
15 \]
f ◦g−g◦f k
X Y Z
h0 k0
X0 Y0 Z0
The column distance of a commutative diagram is governed by the option column sep.
The value of column sep is expected to be a distance, e.g. 4em. Some standard distances
are predefined, see Table 4.8. With column sep one can fix the above diagram.
Example 4.52: Using column sep
1 \[
2 \begin{tikzcd}[column sep = huge]
3 X \arrow{r}{f \circ g - g \circ f} \arrow{d}
93
4.5 Beware of spacings
name tiny small scriptsize normal large huge

distance 0.45em 0.9em 1.35em 1.8em 2.7em 3.6em
Table 4.9: Standard distances for row sep.
4 &
5 Y \arrow{r}{k} \arrow{d}
6 &
7 Z \arrow[equal]{d}
8 \\
9 X' \arrow[dashed]{r}{h'}
10 &
11 Y' \arrow{r}{k'}
12 &
13 Z'
14 \end{tikzcd}
15 \]
f ◦g−g◦f k
X Y Z
h0 k0
X0 Y0 Z0
One can similarly use the option row sep to change the distance of the rows. Some
predefined distances can be found in Table 4.9. They are the same as for column sep
but scaled down by a factor of 0.75.

When typesetting a document LATEX groups the appearing symbols and expressions
into certain groups and then adds spacing around these symbols and expressions
depending on which group they belong to. Three of these groups are “operators”,
“relation symbols” and “binary operations”. The symbols = and < are for example
treated as relations symbols, and the symbols + and \cdot as binary operations. We
can see in the following example how some space is automatically added around these
symbols:
Example 4.53: Standard spacing around relation symbols and binary operators
1 \[
2 a = b \qquad a < b \qquad a + b \qquad a \cdot b
94
3 \]
a=b a<b a+b a·b
To compare this to a version without spacing we can surround the symbols by a

pair of curly brackets. This circumvents LATEX from taking the surround code into
consideration. This leads to the following result:
Example 4.54: Disabling the standard spacing around a symbol
1 \[
2 a {=} b \qquad a {<} b \qquad a {+} b \qquad a {\cdot} b
3 \]
a=b a<b a+b a·b
The automatic spacing can become a problem, as the following examples illustrate:
Example 4.55: Clashing spacings around symbols
1 \[
2 X/\sim
3 \quad
4 R/\operatorname{J}(R) X/ ∼ R/ J(R) id ⊗h
5 \quad
6 \operatorname{id} \otimes h
7 \]
This problem can be fixed by surround the respective symbols in curly brackets.
Example 4.56: Preventing a clash of spacings
1 \[
2 X/{\sim}
3 \quad
4 R/{\operatorname{J}(R)} X/∼ R/J(R) id ⊗ h
5 \quad
6 {\operatorname{id}} \otimes h
7 \]
One can tell LATEX how to treat a certain symbol:
95
Example 4.57: Specifying the role (and thus spacing) of a symbol
1 \[
2 a | b
3 \quad
4 a \mathop{|} b
5 \quad a|b a | b a|b a|b
6 a \mathrel{|} b
7 \quad
8 a \mathbin{|} b
9 \]
To define a command \divides to express that a number n divides a number m we do

therefore do the following:
Example 4.58: Defining and using \divides
1 \newcommand{\divides}{\mathrel{|}}
2 \[
3 n \divides m n|m
4 \]
For more on this topic see [TSE11].
96
Chapter 5
Mathematical layout
5.1 Environments for display mathematics
5.1.1 Don’t use $$ $$ or eqnarray
There are many good ways to put mathematical content into display mode, but
$$ $$ and eqnarray are none of them. The method $$ $$ is too low level, and the
environment eqnarray has too many problems and has been deprecated since forever.
One can use any of the following, depending on the planned usage.
5.1.2 The environments \[ \] and equation*

The environments \[ \] and equation* can be used for a single line of display math
mode. Both commands do the same thing (when the package amsmath is loaded).
Example 5.1: Using the environments \[ \] and equation*
1
Suppose that both the formula
Suppose that both the formula
2 \[
3 a + b = c
a+b=c
4 \]
5
and the formula
and the formula
6 \begin{equation*}
7 2a - b = c \,.
2a − b = c .
8 \end{equation*}
9
hold. Then a and b are unique.
hold.
10 Then $a$ and $b$ are unique.
The non-starred version equation numbers the equation.

Example 5.2: Using the environment equation
1 The formula The formula

2 \begin{equation}
3 a^2 - b^2 = (a + b)(a - b) a2 − b2 = (a + b)(a − b) (5.1)
4 \end{equation}
5 is one of the binomial formulas. is one of the binomial formulas.
97
5.1.3 The environment gather*

The environment gather* is meant for multiple lines that are non-aligned.
Example 5.3: Using the environment gather*
1 We consider for every integer $n \geq 0$ the polynomial

2 \[
3 p_n
4 =
5 \sum_{k=0}^n x^k \,.
6 \]
7 In particular
8 \begin{gather*}
9 p_0 = 1 \,,
10 \qquad
11 p_1 = 1 + x \,,
12 \qquad
13 p_2 = 1 + x + x^2 \,,
14 \qquad
15 p_3 = 1 + x + x^2 + x^3 \,,
16 \\
17 p_4 = 1 + x + x^2 + x^3 + x^4 \,,
18 \qquad
19 p_5 = 1 + x + x^2 + x^3 + x^4 + x^5 \,.
20 \end{gather*}
We consider for every integer n ≥ 0 the polynomial

n
X
pn = xk .
k=0
In particular
p0 = 1 , p1 = 1 + x , p 2 = 1 + x + x2 , p 3 = 1 + x + x2 + x3 ,
p4 = 1 + x + x2 + x3 + x4 , p5 = 1 + x + x2 + x3 + x4 + x5 .
The non-starred version gather numbers the lines.

Example 5.4: Using the environment gather
1 We have the polynomials

2 \begin{gather}
3 p_0 = 1 \,,
4 \qquad
5 p_1 = 1 + x \,,
6 \qquad
98
7 p_2 = 1 + x + x^2 \,,

8 \qquad
9 p_3 = 1 + x + x^2 + x^3 \,,
10 \\
11 p_4 = 1 + x + x^2 + x^3 + x^4 \,,
12 \qquad
13 p_5 = 1 + x + x^2 + x^3 + x^4 + x^5 \,.
14 \end{gather}
We have the polynomials
p0 = 1 , p1 = 1 + x , p 2 = 1 + x + x2 , p 3 = 1 + x + x2 + x3 ,
(5.2)
p 4 = 1 + x + x2 + x3 + x4 , p 5 = 1 + x + x2 + x3 + x4 + x5 . (5.3)
5.1.4 The environments align* and alignat*

The environment align* takes multiple lines which can aligned. Each line contains the
symbol & once, and the occurrences of this symbol are then aligned.
Example 5.5: Using the environment align*
1 We find that
2 \begin{align*}
3 a + b + c
4 &= We find that
5 d + e + f + g
6 \\ a+b+c=d+e+f +g
7 &=
=h+i+j
8 h + i + j
9 \\ = k + l + m + n.
10 &=
11 k + l + m + n \,.
12 \end{align*}
The unstarred version align numbers the lines.
99
Example 5.6: Using the environment align
1 We find again that

2 \begin{align}
3 a + b + c
4 &= We find again that
5 d + e + f + g
6 \\ a+b+c=d+e+f +g (5.4)
(5.5)
7 &=
=h+i+j
8 h + i + j
9 \\ = k + l + m + n. (5.6)
10 &=
11 k + l + m + n \,.
12 \end{align}
One can also use multiple aligned columns, which then need to be separated by an
additional &. For n aligned columns we hence need 2n + 1 occurrences of & per line.
Example 5.7: Using align* with multiple columns
1 We consider the values

2 \begin{align*}
3 x_1 &= 1 \,, & x_2 &= 2 \,, & x_3 &= 3 \,, \\
4 x_4 &= 4 \,, & x_5 &= 5 \,, & x_6 &= 6 \,, \\
5 x_7 &= 7 \,, & x_8 &= 8 \,, & x_9 &= 9 \,.
6 \end{align*}
We consider the values
x1 = 1 , x2 = 2 , x3 = 3 ,
x4 = 4 , x5 = 5 , x6 = 6 ,
x7 = 7 , x8 = 8 , x9 = 9 .
The environment alignat* is similar to the environment align* but doesn’t add any
built-in spacing between the aligned columns. Any such spacing must therefore by
added by hand. One also has to specify the number of columns beforehand.
Example 5.8: Using the environment alignat*
1 We also consider the values

2 \begin{alignat*}{3}
3 y_1 &= 9 \,, &\qquad y_2 &= 8 \,, &\qquad y_3 &= 7 \,, \\
4 y_4 &= 6 \,, & y_5 &= 5 \,, & y_6 &= 4 \,, \\
5 y_7 &= 3 \,, & y_8 &= 2 \,, & y_9 &= 1 \,.
6 \end{alignat*}
100
We also consider the values
y1 = 9 , y2 = 8 , y3 = 7 ,
y4 = 6 , y5 = 5 , y6 = 4 ,
y7 = 3 , y8 = 2 , y9 = 1 .
For aligning multiple columns one should use alignat* instead of align* to get (manu-
ally) a good looking distance between the aligned columns. The environment alignat
works the same way as alignat* but automatically numbers the lines.
5.1.5 Don’t use center plus $ $

By all that is holy, don’t do the following:
Example 5.9: Using \center with $ $
1 Some text that ends in Some text that ends in

2 \begin{center}
3 $a + b = c$. a + b = c.
4 \end{center}
5 Now some more text. Now some more text.
5.1.6 The environments gathered, aligned and alignedat

The environments gathered, aligned and alignedat are variations of the environ-
ments gather*, align* and alignat* that can be used inside an already existing math
environment.
Example 5.10: Using aligned
1 \[
2 \left\{
3 \begin{aligned} ( )
4 a + b &= c \\ a+b=c
5 d &= e + f d=e+f
6 \end{aligned}
7 \right\}
8 \]
5.1.7 Overview
One should always use the most basic environment that does the job: Using overpowered
environments can lead to unexpected problems. Consider the following example:
101
5.2 Where to break and align formulas
Example 5.11: Improper use of align*
1 \begin{align*}
2 ABCD = EF = GHI = JKL = M = NOP
3 \\
4 QRS = TUV = WX = Y = Z
5 \end{align*}
ABCD = EF = GHI = JKL = M = N OP

QRS = T U V = W X = Y = Z
The environment align* automatically alignes both lines on the right since no infor-
mation about alignment was given. In the above situation one should use gather*
instead.
Example 5.12: Using gather* for non-aligned equations
1 \begin{gather*}
2 ABCD = EF = GHI = JKL = M = NOP
3 \\
4 QRS = TUV = WX = Y = Z
5 \end{gather*}
ABCD = EF = GHI = JKL = M = N OP

QRS = T U V = W X = Y = Z
If a multi-line display mode environment is used for a single line then one can get
spacing issues, see Section 5.4.
The flowchart in Figure 5.1 explains how to choose the correct display environment.
(This flowchart is partly inspired by [Kot15].)

Oftentimes a formula is broken among multiple lines. This is done for at least two
reasons:
• To improve the readability of the given formula.
• To prevent that the formula goes over the margins of the text area.
If a formula is broken among multiple lines then one has to choose at which places the
formula is broken, and how the resulting parts of the formula will be aligned. In this
section we present some standard ways of doing so.
102
formula
no \[ \]
multi-line? equation*
yes
no
alignment? gather*
yes
no
multiple columns? align*
yes
alignat*
If numbering of the line(s) is needed then the unstarred version is to be used.
Figure 5.1: Deciding on a math environment.
5.2.1 When to break a formula

We first discuss when a formula needs to be broken.
When the formula is too long

If a formula is too long to physically fit into the text area, i.e. if it goes over the margins
of the text area, then it must be broken up. These occurrences are easy to spot since
they give will give an “overfull hbox” warning. Consider the following example:
Example 5.13: An overfull hbox
1 It follows from
2 \[
3 aaaaaaa
4 =
5 bbbbbbb
6 =
7 ccccccc
8 \leq
9 ddddddd
10 =
103
11 eeeeeee
12 =
13 fffffff
14 \leq
15 ggggggg
16 =
17 hhhhhhh
18 \]
19 that $a \leq h$.
It follows from
aaaaaaa = bbbbbbb = ccccccc ≤ ddddddd = eeeeeee = f f f f f f f ≤ ggggggg = hhhhhhh
that a ≤ h.
When the formula is visually too long

Sometimes a formula does fit into a single line, but barely so. Consider the following
example:
Example 5.14: Visually overfull hbox
1 Here is some text.

2 \[
3 aaaaaaaaa = bbbbbbbbbbbbbb = cccccccccccccc = dddddddddddddd = eeeeeeeeee
4 \]
5 Here is some more text.
Here is some text.
aaaaaaaaa = bbbbbbbbbbbbbb = cccccccccccccc = dddddddddddddd = eeeeeeeeee
Here is some more text.
The formula in the above example does actually not go over the margin. But it has
stretched the displaymode over its visual limits. The formula does therefore need to be
broken up.
Expressing structure
Often the breaking up of a formula is done to better express the structure–and thus
content–of the displayed formula. Consider the following example:
104
Example 5.15: A formula that should be broken up for readability
1 If $k$ is algebraically closed and $i$ is a square root of $-1$ then

2 \[
3 k[x]/(x^2 + 1)
4 =
5 k[x]/( (x - i) (x + i) )
6 \cong
7 k[x]/(x - i) \times k[x]/(x + i)
8 \cong
9 k \times k
10 \]
11 by the Chinese remainder theorem.
If k is algebraically closed and i is a square root of −1 then
k[x]/(x2 + 1) = k[x]/((x − i)(x + i)) ∼

= k[x]/(x − i) × k[x]/(x + i) ∼
=k×k
by the Chinese remainder theorem.
The above output still has an appropriate length to be put into a single line, and if
space is spare then this is an acceptable solution. But this single-line approach to the
formula does not help to display its internal structure. This can be done by splitting
up the formula as done in the next example:
Example 5.16: Broken up version of Example 15
1 If $k$ is algebraically closed and $i$ is a square root of $-1$ then

2 \begin{align*}
3 k[x]/(x^2 + 1)
4 &=
5 k[x]/( (x - i) (x + i) )
6 \\
7 &\cong
8 k[x]/(x - i) \times k[x]/(x + i)
9 \\
10 &\cong
11 k \times k
12 \end{align*}
13 by the Chinese remainder theorem.
If k is algebraically closed and i is a square root of −1 then
k[x]/(x2 + 1) = k[x]/((x − i)(x + i))

∼
= k[x]/(x − i) × k[x]/(x + i)
∼
=k×k
105
by the Chinese remainder theorem.
This form makes it clear where the equalities and isomorphisms occur.
5.2.2 Where to break and align a formula

We now discuss at which points a formula can be broken, and how these broken parts
can then be aligned.
Aligning at relation symbols I

A first approach is to all relation symbols under each other,
Example 5.17: Aligning relation symbols I
1 Let $x$,$y$ be two nilpotent elements in a commutative ring $R$.

2 Then
3 \begin{align*}
4 \exp(x) \exp(y)
5 &=
6 \left( \sum_{k=0}^\infty \frac{x^k}{k!} \right)
7 \left( \sum_{l=0}^\infty \frac{y^l}{l!} \right)
8 \\
9 &=
10 \sum_{k,l=0}^\infty \frac{x^k y^l}{k! \, l!}
11 \\
12 &=
13 \sum_{n=0}^\infty \, \sum_{k+l = n} \frac{x^k y^l}{k! \, l!}
14 \\
15 &=
16 \sum_{n=0}^\infty \frac{1}{n!} \sum_{k=0}^n \binom{n}{k} x^k y^{n-k}
17 \\
18 &=
19 \sum_{n=0}^\infty \frac{1}{n!} (x + y)^n
20 \\
21 &=
22 \exp(x + y) \,.
23 \end{align*}
106
Let x,y be two nilpotent elements in a commutative ring R. Then

∞
! ∞ !
xk X yl
exp(x) exp(y) =
X
k! l!
k=0 l=0
∞
X xk y l
=
k! l!
k,l=0
X∞ X xk y l
=
n=0 k+l=n
k! l!
∞ n
X 1 X n
= xk y n−k
n=0
n! k
k=0
∞
X 1
= (x + y)n
n=0
n!
= exp(x + y) .
Aligning at relation symbols II

One can also align all relation symbols to the left, so that the broken up parts of the
formula all lie on top of each other.
Example 5.18: Aligning relation symbols II
1 Let~$R$ and~$S$ be two commutative rings.

2 Then for every~$(r,s) \in R \times S$,
3 \begin{align*}
4 {}&
5 (r, s) \in (R \times S)^\times
6 \\
7 \iff{}&
8 \text{there exist $(r', s') \in R \times S$ with $(r,s)(r',s') = (1,1)$}
9 \\
10 \iff{}&
11 \text{there exist $r' \in R$ and $s' \in S$ with $rr' = r$ and $ss' = 1$}
12 \\
13 \iff{}&
14 \text{$r \in R^\times$ and $s \in S^\times$}
15 \\
16 \iff{}&
17 (r,s) \in R^\times \times S^\times \,,
18 \end{align*}
19 and therefore $(R \times S)^{\times} = R^\times \times S^\times$.
107
Let R and S be two commutative rings. Then for every (r, s) ∈ R × S,
(r, s) ∈ (R × S)×
⇐⇒ there exist (r0 , s0 ) ∈ R × S with (r, s)(r0 , s0 ) = (1, 1)
⇐⇒ there exist r0 ∈ R and s0 ∈ S with rr0 = r and ss0 = 1
⇐⇒ r ∈ R× and s ∈ S ×
⇐⇒ (r, s) ∈ R× × S × ,
and therefore (R × S)× = R× × S × .
The empty pair of curly brackets in \iff{}& ensures that the spacings coming from
\iffand & do not interfere with each other. Otherwise something like this happens:
Example 5.19: Wrong spacing when alignment points are set wrong
1 \begin{align*}
2 &\text{some stuff} \\
3 =&\text{some other stuff}
4 \end{align*}
some stuff
=some other stuff
Breaking at a binary operator

Sometimes it is also useful to break a long term of a formula at a binary operator. In
this case this operator needs to occur in the line after the break. The following example
does it wrong:
Example 5.20: Wrong aligning at a binary operator I
1 \begin{align*}
2 & a + b + c + \\ a + b + c+
3 & d + e
d+e
4 \end{align*}
The following should be done instead:
108
Example 5.21: Right breaking at a binary operator I
1 \begin{align*}
2 & a + b + c \\ a+b+c
3 & + d + e
+d+e
4 \end{align*}
If a formula is broken at relation symbols and one of the resulting terms is broken at a
binary operator, then the operator is not aligned together with the relation symbols.
Instead the binary operator appears after the relation symbols. The following example
does it wrong:
Example 5.22: Wrong breaking at a binary operator II
1 \begin{align*}
2 a + a
3 &=
4 b + b + b + b
5 \\ a+a=b+b+b+b
6 &+
+b+b+b+b
7 b + b + b + b
8 \\ =c+c+c+c+c
9 &=
10 c + c + c + c + c
11 \end{align*}
Instead the following has to be done:

Example 5.23: Right breaking at a binary operator II
1 \begin{align*}
2 a + a
3 ={}&
4 b + b + b + b
5 \\ a+a=b+b+b+b
6 {}&
+b+b+b+b
7 +b + b + b + b
8 \\ =c+c+c+c+c
9 ={}&
10 c + c + c + c + c
11 \end{align*}
A single term in another line

If a single line equation is too long, then it is sometimes appropriate to put the last
term in a new line, such that the last term occurs below the second to last term.
109
Example 5.24: Single term in new line
1 It follows that
2 \begin{align*}
3 aaaaaaaaaaaa
4 =
5 bbbbbbbbbbb
6 =
7 cccccccccc
8 =
9 ddddddddd
10 &=
11 eeeeeeee
12 \\
13 &=
14 ffffff
15 \end{align*}
16 and hence $2 + 2 = 5$.
It follows that
aaaaaaaaaaaa = bbbbbbbbbbb = cccccccccc = ddddddddd = eeeeeeee

= ffffff
and hence 2 + 2 = 5.
5.2.3 Don’t break formulas badly

One should always keep in mind that breaking up a formula isn’t just meant to prevent
technical problems, but more importantly to let the resulting output display the
structure–and thus part of the content–of the formula. A badly broken up formula
is harder to understand for both the reader and the author. Consider the following
example:
Example 5.25: Badly broken formula
1 It follows that
2 \begin{align*}
3 aaaaaaaaa
4 &=
5 bbbbbbbbbbb
6 =
7 cccccccc
8 \leq
9 dddddd
110
10 =
11 eeeeeee
12 \\
13 &=
14 eee
15 \leq
16 fffffff
17 =
18 gggggggg
19 \leq
20 hhhhhhhhh
21 \\
22 &<
23 kkkkkkk
24 =
25 llll
26 \leq
27 mmmma
28 =
29 nnnnnnn
30 =
31 pp \,.
32 \end{align*}
It follows that
aaaaaaaaa = bbbbbbbbbbb = cccccccc ≤ dddddd = eeeeeee

= eee ≤ f f f f f f f = gggggggg ≤ hhhhhhhhh
< kkkkkkk = llll ≤ mmmma = nnnnnnn = pp .
In such a case the breaking of the formula should be done in a consistent way. There
seems to be two sensible approaches in this example:
Align everything
One can align all occurring relation symbols:
Example 5.26: Aligning all relation symbols
1 It follows that
2 \begin{align*}
3 aaaaaaaaa
4 &= bbbbbbbbbbb \\
5 &= cccccccc \\
6 &\leq dddddd \\
111
7 &= eeeeeee \\
8 &= eee \\
9 &\leq fffffff \\
10 &= gggggggg \\
11 &\leq hhhhhhhhh \\
12 &< kkkkkkk \\
13 &= llll \\
14 &\leq mmmma \\
15 &= nnnnnnn \\
16 &= pp \,.
17 \end{align*}
It follows that
aaaaaaaaa = bbbbbbbbbbb
= cccccccc
≤ dddddd
= eeeeeee
= eee
≤ fffffff
= gggggggg
≤ hhhhhhhhh
< kkkkkkk
= llll
≤ mmmma
= nnnnnnn
= pp .
This approach has the advantage of being very consistent. But it has the disadvantage
of taking a lot of space. It also may not reflect the structure of the formula particularly
well, as this layout gives all (in)equalities the same importance.
Align at inequalities
On could align all the inequality symbols, to make it clear where these occur:
Example 5.27: Aligning all inequalities
1 It follows that
2 \begin{align*}
3 aaaaaaaaa
4 &= bbbbbbbbbbb
112
5.3 Aligning nearly aligned formulas
5 = cccccccc
6 \\
7 &\leq
8 dddddd
9 = eeeeeee
10 = eee
11 \\
12 &\leq
13 fffffff
14 = gggggggg
15 \\
16 &\leq
17 hhhhhhhhh
18 \\
19 &<
20 kkkkkkk
21 = llll \\
22 &\leq mmmma
23 = nnnnnnn
24 \\
25 &= pp \,.
26 \end{align*}
It follows that
aaaaaaaaa = bbbbbbbbbbb = cccccccc

≤ dddddd = eeeeeee = eee
≤ f f f f f f f = gggggggg
≤ hhhhhhhhh
< kkkkkkk = llll
≤ mmmma = nnnnnnn
= pp .
This layout emphasizes the importance of the inequalities, while relegating the equalities
to a less important position. Note that we have also aligned the first and last equality
signs to make it clear where the manipulations begin and end. If some other equalities
are also particularly important (e.g. if they follows from some previously hard-earned
proposition) then they too should be aligned

Sometimes formulas turn out to look nearly aligned in the compiled output, even
though this wasn’t planned. But the formulas may still be non-aligned enough to look
113
jarring. In such a case it is often best to align these formulas.

Example 5.28: Accidental jarringly non-aligned expressions
1 \begin{gather*}
2 KK^{-1} = 1 = K^{-1}K \,,
3 \quad
4 EF - FE = \frac{ K - K^{-1} }{ q - q^{-1} } \,,
5 \\
6 KE = q^2 EK \,,
7 \quad
8 KF = q^{-2} FK \,.
9 \end{gather*}
K − K −1
KK −1 = 1 = K −1 K , EF − F E = ,
q − q −1
KE = q 2 EK , KF = q −2 F K .
Note that the first line of the output looks slightly shifted to the left when compared to
the second line. But properly aligning both lines this impression vanishes and instead
gives rise to more coherent look and fell.
Example 5.29: Intentional well-aligned expressions
1 \begin{align*}
2 KK^{-1} = 1 = K^{-1}K \,,
3 \quad
4 &EF - FE = \frac{ K - K^{-1} }{ q - q^{-1} } \,,
5 \\
6 KE = q^2 EK \,,
7 \quad
8 &KF = q^{-2} FK \,.
9 \end{align*}
K − K −1
KK −1 = 1 = K −1 K , EF − F E = ,
q − q −1
KE = q 2 EK , KF = q −2 F K .
114
5.4 Proper spacing before multi-line display mode environments
5.4 Proper spacing before multi-line display mode

environments
If mathematical content is put into display mode then this content will not only be
horizontally centered but also receive some vertical spacing around it to separate it
from the surrounding text. Observe in the following example that both above and
below the formula there is an additional spacing of roughly (exactly?) one line:
Example 5.30: Vertical space around display mode

consectetur adipiscing elit, sed do Lorem ipsum dolor sit amet, con-
eiusmod tempor incididunt ut labore sectetur adipiscing elit, sed do eius-
et mod tempor incididunt ut labore et
2 \[
3 a = b \,. a = b.
4 \]
5 Nunc aliquet bibendum enim Nunc aliquet bibendum enim facilisis
facilisis gravida. Nisl nunc mi gravida. Nisl nunc mi ipsum faucibus
ipsum faucibus vitae aliquet nec vitae aliquet nec ullamcorper.
ullamcorper.
If the line before a display mode environment is too short then this vertical spacing
may be visually too large. The one-line display environments \[ \] and do therefore
automatically adjust this spacing. Observe that in the following example the distance
between the first two text lines is far shorter than the distance between the second and
third text lines.
Example 5.31: Adjusted spacing before one-line display mode
1 Lorem
2 \[
Lorem
3 a = b a=b
4 \]
ipsum dolor sit amet, consectetur
5 ipsum dolor sit amet, consectetur
6 \[
a = b.
7 a = b \,.
8 \]
ipsum elit.
9 ipsum elit.
We can say more precisely that the additional spacing inserted by LATEX is either
\abovedisplayskip or its shorter version \abovedisplayshortskip.
The multi-line display mode environments (like align*) don’t have this feature. It
always uses the spacing \abovedisplayskip (for technical reasons). This can lead to
some unpleasant spacing:
115
5.5 Use the cases environment
Example 5.32: Unpleasant spacing around a multi-line environment
1 text text text text text text text text text text text text text text text
text text text
2 \begin{alignat*}{1}
3 x &= y_i x = yi
4 \shortintertext{text}
5 x &= z_{i+1}
text
6 \end{alignat*} x = zi+1
7 text text text text text text text
text text text text text text text text text
The package mathools provides a partial fix to this problem: By putting the com-
mand \SwapAboveDisplaySkip at the beginning of a multi-line display mode environment
we force LATEX to use the shorter spacing \abovedisplayshortskip.
Example 5.33: Pleasant spacing around a multi-line environment
1 text text text text text text text

text text text text text text text text text text
2 \begin{alignat*}{1} text
3 \SwapAboveDisplaySkip x = yi
4 x &= y_i
5 \shortintertext{text}
text
6 x &= z_{i+1} x = zi+1
7 \end{alignat*}
8 text text text text text text text text text text text text text text text
text
5.5 Use the cases environment

Use the environment cases for case distinctions:
Example 5.34: Using cases
1 It follows that
2 \[
3 A(x)
4 =
5 \begin{cases}
6 x^2 & \text{if $x \leq 0$,} \\
7 3x & \text{if $x = 0$.}
8 \end{cases}
9 \]
116
5.6 Proper placement of the qed-symbol
It follows that
if x ≤ 0,
(
x2
A(x) =
3x if x = 0.
In most cases one should actually use the environment cases*, which ensures that the
second column will be treated as text.
Example 5.35: Using cases*
1 It follows that
2 \[
3 A(x)
4 =
5 \begin{cases*}
6 x^2 & if $x \leq 0$, \\
7 3x & if $x = 0$.
8 \end{cases*}
9 \]
It follows that
if x ≤ 0,
(
x2
A(x) =
3x if x = 0.
The package mathtools defines some more useful variants of the environment cases.
We refer to [CTN19b, 3.4.3] for more information on this.

The proof environment automatically places a qed-symbol at its end.
Example 5.36: Using the proof environment
1 \begin{proof}
2 This is obviously a proof.
3 \end{proof}
Proof. This is obviously a proof.
It can happen that this automatic placement of the qed-symbol gives a bad looking
result. The general rule is that the qed-symbol should not occur at the end of an
otherwise empty line. To fix such a bad placement one can use the command \qedhere
117
qed-symbol after lists

If a proof consists of a list then the qed-symbol will by default be placed after the list,
and thus in a new line.
Example 5.37: Improper placement of the qed-symbol after a list I
1 \begin{proof}
2 This proof consists of a list.
3 \begin{enumerate}
4 \item
5 Some part of the proof.
6 \item
7 Another part of the proof.
8 \end{enumerate}
9 \end{proof}
Proof. This proof consists of a list.
1. Some part of the proof.

2. Another part of the proof.
This can be fixed by placing the command \qedhere
Example 5.38: Proper placement of the qed-symbol after a list
1 \begin{proof}
3 \begin{enumerate}
4 \item
6 \item
8 \qedhere
9 \end{enumerate}
10 \end{proof}

But one has to be careful no to introduce a new line before the command \qedhere.
118
Example 5.39: Improper placement of the qed-symbol after a list II
1 \begin{proof}
3 \begin{enumerate}
4 \item
6 \item
8
9 \qedhere
10 \end{enumerate}
11 \end{proof}

qed-symbol after displaystyle

If a proof ends with display style mathematics then the qed-symbol will by default be
placed after this display mode.
Example 5.40: Improper placement of the qed-symbol after display mode I
1 \begin{proof}
2 This proof ends with a some display~mode mathematics, namely
3 \[
4 \sin(x+y)
5 =
6 \cos(x)\sin(y) + \sin(x)\cos(y) \,.
7 \]
8 \end{proof}
Proof. This proof ends with a some display mode mathematics, namely
sin(x + y) = cos(x) sin(y) + sin(x) cos(y) .
There are at least different approaches to fixing this situation:

Some authors prever to bring the qed-symbol to the height of the display environment:
119
Example 5.41: Somewhat proper placement of the qed-symbol after display mode
1 \begin{proof}
3 \[
4 \sin(x+y)
5 =
6 \cos(x)\sin(y) + \sin(x)\cos(y) \,.
7 \qedhere
8 \]
9 \end{proof}
sin(x + y) = cos(x) sin(y) + sin(x) cos(y) .
But this approach does not work if the (last line) of the display environment has
additional height. Then the mathematical formula go below the line on which the
qed-symbol rests.
Example 5.42: Improper placement of the qed-symbol after display mode II
1 \begin{proof}
3 \[
4 1
5 + \frac{a}{b - a}
6 =
7 \frac{(b - a) + a}{b - a}
8 =
9 \frac{b}{b - a} \,.
10 \qedhere
11 \]
12 \end{proof}
a (b − a) + a b
1+ = = .
b−a b−a b−a
The author is the opinion that in such a cases there is no good placement for the
qed-symbol.
Aside the placement of the qed-symbol there is another problem to ending a proof
with a display environment: One of the functions of display mode is to put an emphasis
on the displayed mathematical content. Putting such an emphasis directly before the
120
5.7 Tagging and numbering
end of a proof can lead to the proof missing a sense of closure.

This problem leads to the second solution for the placement of the qed-symbol.
Simply never end a proof with a display environment. This can be done by rewriting
the end of proof, or by adding a closing sentence to it.
qed-symbol at the end of a long text

It can also happen that the qed-symbol is pushed to a new line if the previous line is
completely filled with text. (Although LATEX will actually try quite hard to prevent
this from happening.) If this happens then one should (slightly) rewrite the text to
circumvent this problem.

A finer control of tags can be achieved using \tag and \notag.
5.7.1 Don’t autonumber all formulas

Don’t indiscriminately number every occurring formula. Instead, an equation should
be numbered only if it will be referred to later on. This numbering should then be done
automatically by using a suitable environment like equation, gather, align or alignat.
5.7.2 \tag
With the command \tag a custom tag can be set. This is useful for marking selected
equations by special symbols:
Example 5.43: Using \tag for marking a line
1 Consider the equation

2 \begin{equation}
3 \label{important equation}
4 2 + 2 = 5 \,.
5 \tag{\ast}
6 \end{equation}
7 Note that \cref{important equation} can equivalently be expressed as~$5 = 2 +
2$.
Consider the equation

2 + 2 = 5. (∗)
Note that equation (∗) can equivalently be expressed as 5 = 2 + 2.
The argument of \tag is in text mode, and the resulting tag is automatically enclosed
in parentheses. These parentheses can be removed by using the starred command \tag*
instead.
121
The command \tag should not be used for regular numbering of equations. It should
be used to tag only certain (often a single) equations in a special way.
It can also be used to express that certain transformations have been used, as the
following example demonstrates:
Example 5.44: Using \tag to explain steps
1 It follows from the Chinese remainder theorem that

2 \begin{align*}
3 \mathbb{R}[x] / ( x^3 + x^2 + x + 1 )
4 &=
5 \mathbb{R}[x] / ( (x^2 + 1) (x + 1) )
6 \\
7 &\cong
8 \mathbb{R}[x] / ( x^2 + 1 ) \times \mathbb{R}[x] / ( x + 1 )
9 \tag{CRT}
10 \\
11 &\cong
12 \mathbb{C} \times \mathbb{R}
13 \end{align*}
It follows from the Chinese remainder theorem that
R[x]/(x3 + x2 + x + 1) = R[x]/((x2 + 1)(x + 1))

∼
= R[x]/(x2 + 1) × R[x]/(x + 1) (CRT)
∼
=C×R
5.7.3 \notag
According to Section 5.7.1 a formula should be numbered only if it needs to be referred to.
But if this formula occurs in a multi-line environment like align* then by switching align
all lines will be numbered. The prevent the numbering of the unrequired lines the
command \notag can then be used:
Example 5.45: Using \notag to prevent selected line numbers
1 \begin{align}
2 a a=b
3 &= b \notag \\
4 &= c \\
=c (5.7)
5 &= d \notag \\ =d
6 &= e
=e (5.8)
7 \end{align}
122
5.8 Multi-line set descriptions
5.8 Multi-line set descriptions

Multi-line set descriptions of the form
 
 e1 , . . . , e n ∈ R 
is a complete set of 

 
(e1 , . . . , en )
pairwise orthogonal


idempotents

 


can be typeset by using a tabular environment for the right hand side of the set
description:
Example 5.46: Multi-line set descriptions with tabular
1 \[
2 \left\{
3 x \in X
4 \,\middle|\,
x satisfies

5 \begin{tabular}{@{}c@{}}
x∈X
certain conditions

6 $x$ satisfies \\
7 certain conditions
8 \end{tabular}
9 \right\}
10 \]
Note that the entries of the environment tabular are automatically in text mode. The
argument {} ensure that the environment tabular does not insert additional spacing to
its left and right.
123
Bibliography
[Bel18] Pieter Belmans. My good practice concerning inline mathematical formulas.
December 18, 2018. url: https://pbelmans.ncag.info/blog/2010/12/18/
my-good-practice-concerning-inline-mathematical-formulas (visited on
September 4, 2019).
[CMS17] The Chicago Manual of Style. The Essential Guide for Writers, Editors and
Publishers. 17th ed. The University of Chicago Press, 2017. xvii+1144 pp.
isbn: 978-0-226-28705-8. doi: 10.7208/cmos17.
[CTN16] Simon Fear. Publication quality tables in LATEX. Version 1.618033. April 29,
2016. 17 pp. url: https://www.ctan.org/pkg/booktabs (visited on Septem-
ber 1, 2019).
[CTN18] Augusto Stoffel. {tikzcd}. Commutative diagrams with TikZ. Version 0.9f.
November 19, 2018. 17 pp. url: https://ctan.org/pkg/tikz-cd (visited
on September 4, 2019).
[CTN19a] Javier Bezos. Customizing lists with the enumitem package. Version 3.9.
June 20, 2019. 23 pp. url: https://ctan.org/pkg/enumitem (visited on
September 14, 2019).
[CTN19b] Morten Høgholm and Lars Madsen. The mathtools package. Version 1.22.
July 31, 2019. 37 pp. url: https://ctan.org/pkg/mathtools (visited on
September 12, 2019).
[CTN19c] Philip Kime, Moritz Wemheuer, and Philipp Lehman. The biblatex Pack-
age. Programmable Bibliographies and Citations. Version 3.13. August 17,
2019. 336 pp. url: https://ctan.org/pkg/biblatex (visited on Septem-
ber 6, 2019).
[Kot15] Stefan Kottwitz. Example: Flowchart. August 29, 2015. url: http : / /
www.texample.net/tikz/examples/math-flowchart/(visited on August 27,
2019).
[L2M19] LATEX2e. An unofficial reference manual. June 2019. ix + 262 pp. url:
https://latexref.xyz (visited on September 4, 2019).
[TSE11] Nobby. What is the difference between \mathbin vs. \mathrel? December 21,
2011. url: https://tex.stackexchange.com/questions/38982/what-is-the-
difference-between-mathbin-vs-mathrel (visited on August 19, 2019).
[TSE14] Lupino. For formal articles, should a displayed equation be followed by a
punctuation to conform to the language grammar ? April 10, 2014. url:
https://tex.stackexchange.com/a/170691/108717 (visited on August 18,
2019).
124
List of examples
1.1 Code is too cramped . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Code is less cramped . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 Code isn’t cramped . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.4 Unrelated symbols need to be separated . . . . . . . . . . . . . . . . . 2
1.5 Unrelated symbols are separated . . . . . . . . . . . . . . . . . . . . . 3
1.6 Unrelated symbols are separated even better . . . . . . . . . . . . . . 3
1.7 Wrong spacing before a footnote . . . . . . . . . . . . . . . . . . . . . 3
1.8 Right spacing before a footnote . . . . . . . . . . . . . . . . . . . . . 4
1.9 Layout of main.tex . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.10 Layout of mystyle.sty . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.11 Layout of chapter1.tex . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.12 Layout of section1.tex . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.13 Basic example of main.tex . . . . . . . . . . . . . . . . . . . . . . . . 6
1.14 Using the command \includeonly . . . . . . . . . . . . . . . . . . . . 7
1.15 Syntax of \includeonly . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.16 Using the option chapterprefix . . . . . . . . . . . . . . . . . . . . . 8
1.17 Changing KOMA fonts . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.18 Using the option egregdoesnotlikesansseriftitles . . . . . . . . . . 8
1.19 Using the environment abstract . . . . . . . . . . . . . . . . . . . . . 9
1.20 Using the option titlepage . . . . . . . . . . . . . . . . . . . . . . . . 9
1.21 Using the option oneside . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.22 Using the option openany . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.23 Using the option openany . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.24 General document layout with scrbook . . . . . . . . . . . . . . . . . 11
1.25 Inconsistency looks bad . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.1 Syntax of \newtheorem) . . . . . . . . . . . . . . . . . . . . . . . . . . 12

2.2 Using the command \newtheorem . . . . . . . . . . . . . . . . . . . . . 12
2.3 Using the command \newtheorem* . . . . . . . . . . . . . . . . . . . . 13
2.4 Theorem-like environment use different counters by default . . . . . . 13
2.5 Syntax of \newcounter . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.6 Setting up a common counter . . . . . . . . . . . . . . . . . . . . . . . 14
2.7 Binding a new counter to the section level . . . . . . . . . . . . . . . 14
2.8 Syntax of \label . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.9 Syntax of \ref . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.10 Using the commands \label and \ref . . . . . . . . . . . . . . . . . . 16
2.11 Specifying the kind of reference . . . . . . . . . . . . . . . . . . . . . 17
2.12 Syntax of \eqref . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
125
List of examples
2.13 Using the command \eqref . . . . . . . . . . . . . . . . . . . . . . . . 17

2.14 Loading the package cleveref . . . . . . . . . . . . . . . . . . . . . . 18
2.15 Syntax of \cref . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.16 Using the command \cref . . . . . . . . . . . . . . . . . . . . . . . . 18
2.17 Using \lcnamecref . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.18 Order of preamble with cleveref and hyperref . . . . . . . . . . . . . 20
2.19 Syntax of \enquote . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.20 Using the command \enquote . . . . . . . . . . . . . . . . . . . . . . . 20
2.21 \enquote chooses the right kind of quotation marks . . . . . . . . . . 21
2.22 Nested quotes with \enquote . . . . . . . . . . . . . . . . . . . . . . . 21
2.23 Using the environment enumerate . . . . . . . . . . . . . . . . . . . . 22
2.24 Using the environment itemize . . . . . . . . . . . . . . . . . . . . . . 22
2.25 Using the environment description . . . . . . . . . . . . . . . . . . . 22
2.26 Changing the numbering style of enumerate lists . . . . . . . . . . . . 23
2.27 Changing the symbol for itemize lists . . . . . . . . . . . . . . . . . . 23
2.28 Resuming lists with resume* . . . . . . . . . . . . . . . . . . . . . . . 24
2.29 Indentation and distance of pararaphs in lists, default behavior . . . . 25
2.30 Indentation and distance of pararaphs in lists, changed behavior . . . 25
2.31 Distance between list items with itemsep . . . . . . . . . . . . . . . . 26
2.32 enumerate environment with standard indentation . . . . . . . . . . . 26
2.33 enumerate environment without indentation . . . . . . . . . . . . . . . 27
2.34 enumerate environments with only identitation between label and text 27
2.35 enumerate environment with indentation between margin and label, and
label and text . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
label and text . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
label and text . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
2.38 description environment with standard indentation . . . . . . . . . . 29
2.39 description environment with no indentation . . . . . . . . . . . . . . 29
2.40 Syntax of \setlist . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
2.41 Global settings for list environments . . . . . . . . . . . . . . . . . . . 30
2.42 Different settings for nested lists . . . . . . . . . . . . . . . . . . . . . 30
2.43 Global settings depending on level . . . . . . . . . . . . . . . . . . . . 31
2.44 Accessing level counters in settings for list environments . . . . . . . . 31
2.45 Syntax of \newlist . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
2.46 Custom clones of list environments . . . . . . . . . . . . . . . . . . . . 32
2.47 Loading the package biblatex . . . . . . . . . . . . . . . . . . . . . . 34
2.48 Syntax for an entry of the bib-file . . . . . . . . . . . . . . . . . . . . 34
2.49 bib-file entry for a single book . . . . . . . . . . . . . . . . . . . . . . 35
2.50 Adjusting the formatting of pagestotal . . . . . . . . . . . . . . . . . 36
2.51 bib-file entry for a book with multiple volumes . . . . . . . . . . . . . 37
2.52 bib-file entry for a single book . . . . . . . . . . . . . . . . . . . . . . 38
2.53 bib-file entry for an online resource . . . . . . . . . . . . . . . . . . . 39
2.54 bib-file entry for an online resource . . . . . . . . . . . . . . . . . . . 39
126
List of examples
2.55 Syntax of \cite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

2.56 Using the command \cite . . . . . . . . . . . . . . . . . . . . . . . . 40
2.57 Syntax of \bibliography . . . . . . . . . . . . . . . . . . . . . . . . . 40
2.58 Using the command \printbibliography . . . . . . . . . . . . . . . . . 41
2.59 Compiling in the console (with bibliography) . . . . . . . . . . . . . . 42
2.60 A basic table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
2.61 Full horizontal lines in tables . . . . . . . . . . . . . . . . . . . . . . . 43
2.62 Syntax of \cline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
2.63 Partial horizontal lines in tables . . . . . . . . . . . . . . . . . . . . . 43
2.64 Vertical lines in tables . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
2.65 A table with all kinds of lines in it . . . . . . . . . . . . . . . . . . . . 44
2.66 Using the rules of booktabs . . . . . . . . . . . . . . . . . . . . . . . . 45
3.1 Very wrong punctuation in an equation . . . . . . . . . . . . . . . . . 47

3.2 Wrong punctuation in an equation I . . . . . . . . . . . . . . . . . . . 48
3.3 Wrong punctuation in an equation II . . . . . . . . . . . . . . . . . . 48
3.4 Right punctuation in an equation . . . . . . . . . . . . . . . . . . . . 48
3.5 Best punctuation in an equation . . . . . . . . . . . . . . . . . . . . . 48
3.6 Finishing the sentence before the commutative diagram . . . . . . . . 49
3.7 Incorporating the commutative diagram in the sentence . . . . . . . . 49
3.8 Wrong punctuation in lists . . . . . . . . . . . . . . . . . . . . . . . . 50
3.9 Right punctuation in lists I . . . . . . . . . . . . . . . . . . . . . . . . 51
3.10 Right punctuation in lists II . . . . . . . . . . . . . . . . . . . . . . . 51
3.11 A hyphen in math mode . . . . . . . . . . . . . . . . . . . . . . . . . 52
3.12 Proper placement of the hyphen . . . . . . . . . . . . . . . . . . . . . 53
3.13 Proper spacing after periods . . . . . . . . . . . . . . . . . . . . . . . 53
3.14 Proper spacing after periods, and using ~ . . . . . . . . . . . . . . . . 54
3.15 Spacings after periods followed by parentheses or quotation marks . . 54
3.16 Upper case mathematics followed by a period . . . . . . . . . . . . . . 54
3.17 Spacing of ~ and “\ ” . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
3.18 An unwanted line break at a space . . . . . . . . . . . . . . . . . . . . 56
3.19 Using ~ to prevent unwanted line breaks at spaces . . . . . . . . . . . 56
3.20 An unwanted line break at a hyphen . . . . . . . . . . . . . . . . . . . 56
3.21 Using \nobreakdash to prevent unwanted line breaks at hyphens . . . 57
3.22 New paragraphs done right . . . . . . . . . . . . . . . . . . . . . . . . 57
3.23 New “paragraphs” done wrong . . . . . . . . . . . . . . . . . . . . . . 57
3.24 A list without surrounding text . . . . . . . . . . . . . . . . . . . . . 58
3.25 A list without surrounding text . . . . . . . . . . . . . . . . . . . . . 59
3.26 Improper aligned of list items . . . . . . . . . . . . . . . . . . . . . . 59
3.27 Proper aligned of list items with \leavevmode . . . . . . . . . . . . . . 60
3.28 Proper aligned of list items by introducing text . . . . . . . . . . . . . 60
3.29 Proper aligned of list items by introducing text . . . . . . . . . . . . . 61
3.30 Line break in inline math . . . . . . . . . . . . . . . . . . . . . . . . . 61
3.31 Preventing inline math in inline math . . . . . . . . . . . . . . . . . . 62
3.32 Mathematics at the beginning of a sentence . . . . . . . . . . . . . . . 62
127
List of examples
3.33 No mathematics at the beginning of a sentence . . . . . . . . . . . . . 62

3.34 List items beginning with mathematics . . . . . . . . . . . . . . . . . 63
3.35 Not properly separating mathematics . . . . . . . . . . . . . . . . . . 63
3.36 Properly separating mathematics I . . . . . . . . . . . . . . . . . . . . 64
3.37 Properly separating mathematics II . . . . . . . . . . . . . . . . . . . 64
3.38 Syntax of \text . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
3.39 Missing text in math mode . . . . . . . . . . . . . . . . . . . . . . . . 64
3.40 Wrong way of using text in math mode . . . . . . . . . . . . . . . . . 65
3.41 Right way of using text in math mode . . . . . . . . . . . . . . . . . . 65
3.42 Wrong spacing around text in math mode I . . . . . . . . . . . . . . . 66
3.43 Wrong spacing around text in math mode II . . . . . . . . . . . . . . 66
3.44 Right spacing around text in math mode I . . . . . . . . . . . . . . . 66
3.45 Right spacing around text in math mode II . . . . . . . . . . . . . . . 67
3.46 Two non-aligned blocks . . . . . . . . . . . . . . . . . . . . . . . . . . 67
3.47 Two aligned blocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
3.48 Two non-aligned equalities . . . . . . . . . . . . . . . . . . . . . . . . 69
3.49 Two aligned equalities with too much space . . . . . . . . . . . . . . . 69
3.50 Two aligned equalities with proper space . . . . . . . . . . . . . . . . 70
3.51 Right way of using mathematical symbols as words . . . . . . . . . . 70
3.52 Wrong way of using mathematical symbols as words . . . . . . . . . . 71
3.53 Abuse of quantifiers . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
3.54 Writing out quantifiers . . . . . . . . . . . . . . . . . . . . . . . . . . 71
3.55 Best use of quantifiers . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
3.56 Okay way to use logical symbols . . . . . . . . . . . . . . . . . . . . . 72
4.1 Using \not . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73

4.2 Why not to use \not . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
4.3 Using the command \centernot . . . . . . . . . . . . . . . . . . . . . 74
4.4 Using limits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
4.5 Using built-in limits . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
4.6 Inline math vs. display math with built-in limits . . . . . . . . . . . . 76
4.7 Inline math with \limits . . . . . . . . . . . . . . . . . . . . . . . . . 77
4.8 Inline with built-in limits . . . . . . . . . . . . . . . . . . . . . . . . . 77
4.9 Inline vs. display for different commands with built-in limits . . . . . 77
4.10 Using \DeclareMathOperator* to define \colim . . . . . . . . . . . . . 78
4.11 Using \colim . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
4.12 Using \overset and \underset to put text above or below an arrow . 78
4.13 Using \xrightarrow to put text above or below an arrow . . . . . . . 79
4.14 Defining arrow commands as shortcuts . . . . . . . . . . . . . . . . . 79
4.15 Typical usage of the ellipses \cdots, \vdots, \ddots . . . . . . . . . . 81
4.16 Using the ellipsis \ldots . . . . . . . . . . . . . . . . . . . . . . . . . . 82
4.17 Inconsistent results with \dots . . . . . . . . . . . . . . . . . . . . . . 82
4.18 The basic matrix environment matrix . . . . . . . . . . . . . . . . . . 82
4.19 Aligning matrix entries . . . . . . . . . . . . . . . . . . . . . . . . . . 83
4.20 Using the right kind of matrix . . . . . . . . . . . . . . . . . . . . . . 84
128
List of examples
4.21 Wrong kind of differentials . . . . . . . . . . . . . . . . . . . . . . . . 85

4.22 Right kind of differential . . . . . . . . . . . . . . . . . . . . . . . . . 85
4.23 Full fractions in exponent, index and inline . . . . . . . . . . . . . . . 85
4.24 Flat fractions in exponent, index and inline . . . . . . . . . . . . . . . 85
4.25 Using commands vs. not using them . . . . . . . . . . . . . . . . . . . 86
4.26 Syntax of \DeclareMathOperator . . . . . . . . . . . . . . . . . . . . . 86
4.27 Declaring a mathematical operator with \DeclareMathOperator . . . . 86
4.28 Using a declared mathematical operator . . . . . . . . . . . . . . . . . 86
4.29 Automatic spacing of \DeclareMathOperator . . . . . . . . . . . . . . . 87
4.30 Wrong way of using \DeclareMathOperator . . . . . . . . . . . . . . . 87
4.31 Wrong output when \DeclareMathOperator is abused . . . . . . . . . . 87
4.32 Using \operatorname . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
4.33 Missing spacing after \mathrm . . . . . . . . . . . . . . . . . . . . . . . 88
4.34 Syntax of \newcommand . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
4.35 Using the command \newcommand . . . . . . . . . . . . . . . . . . . . . 88
4.36 \DeclareMathOperator = \newcommand plus \operatorname . . . . . . . 88
4.37 Syntax of \DeclarePairedDelimiter . . . . . . . . . . . . . . . . . . . 89
4.38 Defining the command \abs with \DeclarePairedDelimiter . . . . . . 89
4.39 Using a declared delimiter . . . . . . . . . . . . . . . . . . . . . . . . 89
4.40 Scaling of declared delimiters . . . . . . . . . . . . . . . . . . . . . . . 89
4.41 Syntax of \DeclarePairedDelimiterX . . . . . . . . . . . . . . . . . . . 90
4.42 Using \DeclarePairedDelimiterX for more advanced delimiters . . . . 90
4.43 Using more advanced delimiters . . . . . . . . . . . . . . . . . . . . . 90
4.44 Syntax of \NewDocumentCommand . . . . . . . . . . . . . . . . . . . . . . 91
4.45 Mandatory arguments with \NewDocumentCommand II . . . . . . . . . . 91
4.46 Mandatory arguments with \NewDocumentCommand II . . . . . . . . . . 91
4.47 Syntax of \IfValueTF . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
4.48 Optional arguments with \NewDocumentCommand, I . . . . . . . . . . . . 92
4.49 Optional arguments with \NewDocumentCommand, II . . . . . . . . . . . 92
4.50 Starred versions with \NewDocumentCommand . . . . . . . . . . . . . . . 92
4.51 Commutative diagram with an arrow too short . . . . . . . . . . . . . 93
4.52 Using column sep . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
4.53 Standard spacing around relation symbols and binary operators . . . 94
4.54 Disabling the standard spacing around a symbol . . . . . . . . . . . . 95
4.55 Clashing spacings around symbols . . . . . . . . . . . . . . . . . . . . 95
4.56 Preventing a clash of spacings . . . . . . . . . . . . . . . . . . . . . . 95
4.57 Specifying the role (and thus spacing) of a symbol . . . . . . . . . . . 96
4.58 Defining and using \divides . . . . . . . . . . . . . . . . . . . . . . . 96
5.1 Using the environments \[ \] and equation* . . . . . . . . . . . . . . 97

5.2 Using the environment equation . . . . . . . . . . . . . . . . . . . . . 97
5.3 Using the environment gather* . . . . . . . . . . . . . . . . . . . . . . 98
5.4 Using the environment gather . . . . . . . . . . . . . . . . . . . . . . 98
5.5 Using the environment align* . . . . . . . . . . . . . . . . . . . . . . 99
5.6 Using the environment align . . . . . . . . . . . . . . . . . . . . . . . 100
129
List of examples
5.7 Using align* with multiple columns . . . . . . . . . . . . . . . . . . . 100

5.8 Using the environment alignat* . . . . . . . . . . . . . . . . . . . . . 100
5.9 Using \center with $ $ . . . . . . . . . . . . . . . . . . . . . . . . . . 101
5.10 Using aligned . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
5.11 Improper use of align* . . . . . . . . . . . . . . . . . . . . . . . . . . 102
5.12 Using gather* for non-aligned equations . . . . . . . . . . . . . . . . . 102
5.13 An overfull hbox . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
5.14 Visually overfull hbox . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
5.15 A formula that should be broken up for readability . . . . . . . . . . 105
5.16 Broken up version of Example 15 . . . . . . . . . . . . . . . . . . . . 105
5.17 Aligning relation symbols I . . . . . . . . . . . . . . . . . . . . . . . . 106
5.18 Aligning relation symbols II . . . . . . . . . . . . . . . . . . . . . . . 107
5.19 Wrong spacing when alignment points are set wrong . . . . . . . . . . 108
5.20 Wrong aligning at a binary operator I . . . . . . . . . . . . . . . . . . 108
5.21 Right breaking at a binary operator I . . . . . . . . . . . . . . . . . . 109
5.22 Wrong breaking at a binary operator II . . . . . . . . . . . . . . . . . 109
5.23 Right breaking at a binary operator II . . . . . . . . . . . . . . . . . . 109
5.24 Single term in new line . . . . . . . . . . . . . . . . . . . . . . . . . . 110
5.25 Badly broken formula . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
5.26 Aligning all relation symbols . . . . . . . . . . . . . . . . . . . . . . . 111
5.27 Aligning all inequalities . . . . . . . . . . . . . . . . . . . . . . . . . . 112
5.28 Accidental jarringly non-aligned expressions . . . . . . . . . . . . . . 114
5.29 Intentional well-aligned expressions . . . . . . . . . . . . . . . . . . . 114
5.30 Vertical space around display mode . . . . . . . . . . . . . . . . . . . 115
5.31 Adjusted spacing before one-line display mode . . . . . . . . . . . . . 115
5.32 Unpleasant spacing around a multi-line environment . . . . . . . . . . 116
5.33 Pleasant spacing around a multi-line environment . . . . . . . . . . . 116
5.34 Using cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
5.35 Using cases* . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
5.36 Using the proof environment . . . . . . . . . . . . . . . . . . . . . . . 117
5.37 Improper placement of the qed-symbol after a list I . . . . . . . . . . 118
5.38 Proper placement of the qed-symbol after a list . . . . . . . . . . . . . 118
5.39 Improper placement of the qed-symbol after a list II . . . . . . . . . . 119
5.40 Improper placement of the qed-symbol after display mode I . . . . . . 119
5.41 Somewhat proper placement of the qed-symbol after display mode . . 120
5.42 Improper placement of the qed-symbol after display mode II . . . . . 120
5.43 Using \tag for marking a line . . . . . . . . . . . . . . . . . . . . . . . 121
5.44 Using \tag to explain steps . . . . . . . . . . . . . . . . . . . . . . . . 122
5.45 Using \notag to prevent selected line numbers . . . . . . . . . . . . . 122
5.46 Multi-line set descriptions with tabular . . . . . . . . . . . . . . . . . 123
130
Index
\(space), 53 \Leftrightarrow, 74
\@, 53 \notrightarrow, 75
$ $, 101 \Rightarrow, 74
$$ $$, 97 \rightarrow, 73
%, 3 \to, 73
&, 99, 100 \xhookleftarrow, 80
~, 56 \xhookrightarrow, 80
\|, 74 \xLeftarrow, 80
\\, 57 \xleftarrow, 80
\[ \], 97, 115 \xleftharpoondown, 80
\xleftharpoonup, 80
abstract, 8 \xLeftrightarrow, 80
\addtokomafont, 8 \xleftrightarrow, 80
align, 99 \xleftrightharpoons, 80
align*, 67, 99, 115 \xlongequal, 80
alignat, 101 \xLongleftarrow, 80
alignat*, 100 \xlongleftarrow, 80
aligned, 101 \xlongleftrightarrow, 80
alignedat, 101 \xLongrightarrow, 80
aligning formulas, 67, 106–110, 113–114 \xlongrightarrow, 80
see also \align* \xmapsto, 80
see also \alignat* \xRightarrow, 80
Alph*, 23 \xrightarrow, 79, 80
alph*, 23 \xrightharpoondown, 80
amsmath, 12, 79, 82, 97 \xrightharpoonup, 80
amssymb, 12 \xrightleftharpoons, 80
amsthm, 12 @article, 38
\newtheorem, 12 article, 8
proof, 59, 117 arXiv, 39
\appendix, 10 author, 35
Arabic numeral, 23
Arabic numerals, 10, 36 babel,21
arabic*, 23 babel,21
arrows backend, 33
extensible, 79 \backepsilon, 74
\Leftarrow, 74 \backmatter, 9, 10
131
Index
\backslash, 74 bibliography,
10, 41
before a footnote, 3 \cite,
40
bib-file, 33–39 MathSciNet, 35
biber, 33, 42 \printbibliography, 41
biblatex, 33 \bibliography, 40
\cite, 40 bibliography, 10, 41
keys \big, 89
author, 35 \bigg, 89
crossref, 37 \binoppenality, 62
date, 36, 39 Bmatrix, 83
doi, 36 bmatrix, 83
edition, 36 @book, 35
isbn, 36 book, 8
issn, 38 booktabs, 45
journaltitle, 38 \bottomrule, 45
number,36, 38 \cmidrule, 45
pages, 38 \midrule, 45
pagetotal, 36 \toprule, 45
publisher, 36 \bottomrule, 45
series, 36 Bsmallmatrix, 83
subtitle, 35 bsmallmatrix, 83
title, 35
url, 39 c, 42, 83
urldate, 39 capitalise, 18
volume, 37, 38 cases, 116
volumes, 37 cases*, 117
year, 36 \cdots, 81
options center, 101
backend, 33 \centernot, 74
dateabbrev, 34 centernot, 74
style, 34 \chapter, 9
urldate, 34 chapterprefix, 8
\printbibliography, 41 \cite, 40
types cleveref, 18
@article, 38 babel, 21
@book, 35 capitalise, 18
@mvbook, 37 \Cref, 19
@online, 38 \cref, 18
bibliography, 10, 33–42 \lcnamecref, 19
arXiv, 39 \lcnamecrefs, 19
back end, 33 \nameCref, 19
bib-file, 33–39 \namecref, 19
biber, 33, 42 \nameCrefs, 19
biblatex, 33 \namecrefs, 19
\bibliography, 40 noabbrev, 18
132
Index
\cline, 43 $$ $$, 97
\cmidrule, 45 \[ \], 97, 115
\colon, 74 align, 99
\coloneqq, 74 align*, 67, 99, 115
column sep, 93 alignat, 101
commutative diagrams, 15, 48–50 alignat*, 100
consistency, 11 eqnarray, 97
corollary, 12 equation, 97
counter, 13 equation*, 97, 115
\Cref, 19 gather, 98
\cref, 18 gather*, 67, 98
crossref, 37 document class, 8
csquotes, 20 classes
\enquote, 20 article, 8
book, 8
date, 36, 39 report, 8
dateabbrev, 34 scrartcl, 8
\ddots, 81 scrbook, 8
\DeclareMathOperator, 78, 86 scrreprt, 8
\DeclareMathOperator*, 78 options
\DeclarePairedDelimiter, 89 abstract, 8
\DeclarePairedDelimiterX, 90 bibliography, 10, 41
\DeclarePairedDelimiterXPP, 91 chapterprefix, 8
defining commands egregdoesnotlikesansserif-
\DeclareMathOperator, 78, 86 titles,
\DeclareMathOperator*, 78 8
\DeclarePairedDelimiter, 89 oneside, 9
\DeclarePairedDelimiterX, 90 openany, 9, 10
\DeclarePairedDelimiterXPP, 91 titlepage, 9
\newcommand, 88 doi, 36
\NewDocumentCommand, 91 dots
\operatorname, 78, 87 see ellipsis
\operatorname*, 78 \dots, 82
\renewcommand, 88 \dotsb, 81
delimiters \dotsc, 81
\|, 74 \dotsi, 81
\langle, 74 \dotsm, 81
\lVert, 74 \dotso, 81
\rangle, 74
\rVert, 74 edition, 36
see also \DeclarePairedDelimiter egregdoesnotlikesansseriftitles, 8
\delimsize, 90 ellipses, 79–82
description, 22, 28 \cdots, 81
differential, 85 \ddots, 81
display math environment \dots, 82
133
Index
\dotsb, 81 \include, 4
\dotsc, 81 \includeonly, 6
\dotsi, 81 \input, 4
\dotsm, 81 \usepackage, 4
\dotso, 81 fonts
\ldots, 81 \addtokomafont, 8
\vdots, 81 footnote, 3
\emptyset, 74 fractions, 85–86
em dash, 51, 57 \frontmatter, 9, 10
\enquote, 20
enumerate, 22, 26 gather,98
see also enumitem gather*,67, 98
enumi, 31 gathered, 101
enumii, 31 Git, 1
enumitem, 22
\hline, 43
Alph*, 23
huge,93, 94
alph*, 23
arabic*, 23
hyperlink, 19
hyperref, 19
enumi, 31
enumii, 31
hyphen, 51, 52, 56
itemsep, 25
\IfBooleanTF, 92
label, 22, 23
iff, 72
leftmargin, 27
\iff, 74
listparindent, 25
\impliedby, 74
\newlist, 32
\implies, 74
parsep, 25
\in, 74
resume, 24
\include, 4
resume*, 24
\includeonly, 6
Roman*, 23
including files
roman*, 23
see file inclusion
setlist, 29
indent, 26
wide, 26, 29
indentation
en dash, 51, 57
for list environments, 26–29
\epsilon, 74
in source code, 3
eqnarray, 97
of paragraphs, 25, 57
\eqqcolon, 74
\infty, 74
\eqref, 17
\input, 4
equation, 97
integral, 81, 85
equation*, 97, 115
\intertext, 67
error, 1, 3
introduction, 10
extarrows, 12, 79
isbn, 36
extensible arrows, 79
issn, 38
itemize, 22, 28
Facebook, 1
itemsep, 25
file inclusion
134
Index
journal, 38 logical symbols, 71

journaltitle, 38 LuaLATEX, 1
\lVert, 74
kile, 42
KOMA-Script classes magic
\addtokomafont, 8 see microtype
bibliography, 10, 41 \mainmatter, 9, 10
chapterprefix, 8 \mathrm, 88
egregdoesnotlikesansseriftitles, MathSciNet, 35
8 mathtools, 12, 79, 82, 83, 117
oneside, 9 \SwapAboveDisplaySkip, 116
openany, 9, 10 matrices, 82–84
scrartcl, 8 normal sized
scrbook, 8 Bmatrix, 83
scrreprt, 8 bmatrix, 83
titlepage, 9 matrix, 82, 83
totoc, 10 pmatrix, 83
Vmatrix, 83
l, 42, 83 vmatrix, 83
\label, 16 small sized
label, 22, 23 Bsmallmatrix, 83
\langle, 74 bsmallmatrix, 83
large, 93, 94 psmallmatrix, 83
\lcnamecref, 19 smallmatrix, 83
\lcnamecrefs, 19 Vsmallmatrix, 83
\ldots, 81 vsmallmatrix, 83
\leavevmode, 60 matrix, 82, 83
\Leftarrow, 74 \maxdimen, 62
leftmargin, 27 Mercurial, 1
\Leftrightarrow, 74 microtype, 1, 12
lemma, 12 \midrule, 45
limits minus sign, 51, 74
\DeclareMathOperator*, 78 multi-line mathematics
\limits, 76 align, 99
\operatorname*, 78 align*, 67, 99, 115
\limits, 76 alignat, 101
line breaks alignat*, 100
in math mode, 103–113 aligned, 101
in text mode, 56, 57, 61 alignedat, 101
list environments cases, 116
description, 22 cases*, 117
enumerate, 22 gather, 98
itemize, 22 gather*, 67, 98
listparindent, 25 gathered, 101
lists, 50, 63 @mvbook, 37
135
Index
\nameCref,19 @online, 38
\namecref,19 openany, 9, 10
\nameCrefs, 19 \operatorname, 78, 87
\namecrefs, 19 \operatorname*, 78
negation orthography, 47
\centernot, 74 \overset, 78
\neq, 75
\nexists, 75 packages
\nleq, 75 amsmath, 12, 79, 82, 97
\not, 73 amssymb, 12
\notin, 75 amsthm, 12
\notrightarrow, 75 babel, 21
\neq, 75 biblatex, 33
\newcommand, 88 booktabs, 45
\newcounter, 14 centernot, 74
\NewDocumentCommand, 91 cleveref, 18
\IfBooleanTF, 92 csquotes, 20
mandatory argument, 91 enumitem, 22
optional argument extarrows, 12, 79
with default value, 92 hyperref, 19
without default value, 91 mathtools, 12, 79, 82, 83, 117
\newline, 57 microtype, 12
\newlist, 32 ntheorem, 60
\newtheorem, 12 stmaryrd, 12
\newtheorem*, 13 tikz-cd, 15
\nexists, 75 xparse, 91
\ni, 74 pages, 38
\nleq, 75 pagetotal, 36
noabbrev, 18 paragraphs, 25, 57
\nobreakdash, 56 parsep, 25
normal, 93, 94 \part, 9
\not, 73 penalties
\notin, 75 \binoppenality, 62
\notrightarrow, 75 \relpenalty, 62
\nsubseteq, 84 \phi, 74
ntheorem, 60 pmatrix, 83
number, 36, 38 preface, 10
numerals \printbibliography, 41
Arabic, 10, 23, 36 proof, 59, 117
Roman, 23, 36 proposition, 12
roman, 10 psmallmatrix, 83
publisher, 36
o, 91 punctuation
O{ }, 92 em dash, 51, 57
oneside, 9 en dash, 51, 57
136
Index
hyphen, 51, 52, 56 \big, 89

in commutative diagrams, 48–50 \bigg, 89
in formulas, 47–48 \delimsize, 90
in lists, 50–51 scrartcl, 8
scrbook, 8
qed-symbol, 117–121 scrbook, 8
see also \qedhere scriptsize, 93, 94
\qedhere, 117
scrreprt, 8
\qquad, 66
\section, 8
\quad, 67
sectioning
quotation marks, 20 \appendix, 10
nested, 21 \backmatter, 9, 10
\chapter, 9
r, 42, 83
\frontmatter, 9, 10
\rangle, 74
\mainmatter, 9, 10
\ref, 16
\part, 9
referencing
\section, 8
capitalise, 18
\subsection, 8
cleveref, 18
\subsubsection, 8
\Cref,
19
series, 36
\cref,
18
setlist, 29
\eqref, 17
\setminus, 74
hyperref, 19
\shortintertext, 67
\label, 16
small, 93, 94
\lcnamecref, 19
smallmatrix, 83
\lcnamecrefs, 19
\smallsetminus, 74
\nameCref, 19
\namecref, 19
spacing, 25, 68
\nameCrefs, 19
after a period, 53–55
\(space), 53
\namecrefs, 19
\@, 53
noabbrev, 18
\ref, 16
around display mode, 115
\SwapAboveDisplaySkip, 116
\relpenalty, 62
\renewcommand, 88
in lists, 25
report, 8
in math mode, 85, 87, 88, 94–96,
resume, 24
123
resume*, 24
in math mode, 66
\qquad, 66
\Rightarrow, 74
\quad, 67
\rightarrow, 73
Roman numerals, 23, 36 standard classes, 8
article, 8
roman numerals, 10
book, 8
Roman*, 23
report, 8
roman*, 23
stmaryrd, 12
\rVert, 74
structure of an expression, 65
scaling style, 34
137
Index
\subsection, 8 column sep

\subset,84 huge,93
\subseteq, 84 large,93
\subseteqq, 84 normal, 93
\subsetneq, 84 scriptsize, 93
\subsetneqq, 84 small, 93
\subsubsection, 8 tiny, 93
subtitle, 35 row sep
\SwapAboveDisplaySkip, 116 huge,94
symbols, 73 large,94
normal, 94
table of contents, 10 scriptsize, 94
tables, 42–46 small, 94
booktabs, 45 tiny, 94
\bottomrule, 45 column sep, 93
c, 42 tiny, 93, 94
\cline, 43 title, 35
\cmidrule, 45 title page, 10
\hline, 43 titlepage, 9
horizontal lines, 43 \to, 73
l, 42 \toprule, 45
\midrule, 45 totoc, 10
r, 42
tabular, 42 unbreakable
\toprule, 45 dash, 57
vertical lines, 44 hyphen, 56
tabular, 42 space, 56
\tag, 121, 122 underlining, 85
tagging and numbering, 121 \underset, 78
align, 99 Unicode, 1
equation, 97 URL, 39
gather, 98 url, 39
\tag, 121, 122 urldate, 34, 39
TEXStudio, 42 \usepackage, 4
text in math mode
\intertext, 67 \varepsilon, 74
\shortintertext, 67 \varnothing, 74
theorem, 12 \vdots, 81
theorem-like environments, 12, 59 version control, 1, 3
corollary, 12 Vmatrix, 83
lemma, 12 vmatrix, 83
proposition, 12 volume, 37, 38
theorem, 12 volumes, 37
tie, 56 Vsmallmatrix, 83
tikz-cd, 15 vsmallmatrix, 83
138
Index
warning, 1, 3
wide, 26, 29
XƎLATEX, 1
\xhookleftarrow, 80
\xhookrightarrow, 80
\xLeftarrow, 80
\xleftarrow, 80
\xleftharpoondown, 80
\xleftharpoonup, 80
\xLeftrightarrow, 80
\xleftrightarrow, 80
\xleftrightharpoons, 80
\xlongequal, 80
\xLongleftarrow, 80
\xlongleftarrow, 80
\xlongleftrightarrow, 80
\xLongrightarrow, 80
\xlongrightarrow, 80
\xmapsto, 80
xparse, 91
\xRightarrow, 80
\xrightarrow, 79, 80
\xrightharpoondown, 80
\xrightharpoonup, 80
\xrightleftharpoons, 80
year, 36
139

Writing Latex

Uploaded by

Document Informationclick to expand document information

Copyright:

Available Formats

Writing Latex

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Writing Latex

Uploaded by

Copyright:

Available Formats

Writing LATEX

Some do’s and don’ts

September 15, 2019

List of tables vii

List of figures viii

1 Before you start writing 1

2 Useful and important packages 12

2.7.2 Resuming lists . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

4 Good behaviour inside math mode 73

4.1.4 Don’t use \limits . . . . . . . . . . . . . . . . . . . . . . . . . 76

List of examples 125

3.1 Kinds of hyphen and dashes. . . . . . . . . . . . . . . . . . . . . . . . 52

4.1 Right symbols and wrong symbols. . . . . . . . . . . . . . . . . . . . . 74

3.1 Structure tree of the definition. . . . . . . . . . . . . . . . . . . . . . . 65

5.1 Deciding on a math environment. . . . . . . . . . . . . . . . . . . . . . 103

1.1 Use version control

1.2 The compilation process

1.2.2 Don’t ignore warnings

1.3 Code quality

Instead do the following:

You can (and should) even do the following:

This last one is easiest to write and easiest to navigate.

Instead do the following:

Example 1.5: Unrelated symbols are separated

1.3.2 Write one sentence per line

1.3.3 Properly indent your source code

1.3.4 Use % for problematic line breaks

Example 1.8: Right spacing before a footnote

1 Here is some text.%

1.4 Splitting up a project into multiple files

1.4.1 The commands \usepackage, \include, \input

mystyle chapter1 chapter2

section1 section2 section3 section4 section5 section6

Figure 1.1: Splitting up a project into files.

Example 1.9: Layout of main.tex

The file mystyle.sty may look as follows:

10 %%%%% NEW COMMANDS

The file chapter1.tex should look roughly as follows:

1 \chapter{Name of the first chapter}

3 Some introduction to this chapter before the first section appears.

The file section1.tex should look roughly as follows:

1 \section{Name of the first section}

3 Text of the first section.

1.4.2 Use \includeonly

By putting \includeonly{chapter1} in the preamble we tell LATEX to only compile this

Example 1.15: Syntax of \includeonly

1 \includeonly{file_1, ..., file_n}

1.5 Choosing the document class

1 \documentclass[10pt, a4paper, chapterprefix = true]{scrbook}

One can also use the undocumented option egregdoesnotlikesansseriftitles.

Example 1.18: Using the option egregdoesnotlikesansseriftitles

1 \documentclass[10pt, a4paper, egregdoesnotlikesansseriftitles]{scrbook}

1.5.2 Deciding on a document class

Example 1.19: Using the environment abstract

1 \documentclass[10pt, a4paper, abstract = on]{scrartcl}

1 \documentclass[10pt, a4paper, titlepage = false]{scrreprt}

1 \documentclass[10pt, a4paper, oneside]{scrbook}

Table 2.1: The commands \nameref.

The commands \nameref