% arara: pdflatex % arara: bibtex % arara: makeglossaries % arara: pdflatex % arara: makeglossaries % arara: pdflatex % arara: makeindex: { style: thesis-index.ist} % arara: pdflatex % arara: pdflatex % arara: pdflatex % arara: pdflatex % arara: makeindex: { style: thesis-index.ist} % arara: pdflatex \documentclass[report]{dickimawthesis} \usepackage{scrhack} \usepackage{amsmath} \usepackage{amsfonts} \usepackage{longtable} \usepackage{listings} \usepackage{booktabs} \usepackage{alltt} \usepackage{siunitx} \usepackage{array} \usepackage{cmap} \usepackage[utf8]{inputenc} \usepackage[math]{anttor} \usepackage{libris} \renewcommand*{\ttdefault}{txtt} \showdowfalse \newcommand*{\bookdocdate}{\formatdate{16}{03}{2013}} \renewcommand{\glossaryname}{Summary of Commands and Environments} % document information \authordetails{1970}{Talbot}{Nicola L.~C.} \title[Using LaTeX to Write a PhD Thesis]{Using \LaTeX\ to Write a~PhD Thesis} \version{1.3} \volume{2} \series{The Dickimaw \LaTeX\ Series} \seriesurl{http://www.dickimaw-books.com/latex/} \date{\bookdocdate} \keywords{LaTeX,typesetting,thesis,tutorial} \affiliation {Dickimaw Books} {www.dickimaw-books.com} {Saxlingham Nethergate} \subject{Thesis writing} % this is for LaTeX2HTML's benefit: %\manscreenpdf{thesis_screen.pdf} %\manscreenlog{thesis_screen.log} %\manpaperpdf{thesis_a4.pdf} %\manpaperlog{thesis_a4.log} % ensure that the index and glossary information is written to the appropriate files \makeindex \makeglossaries % Acronyms \newacr{ctan}{CTAN} {the Comprehensive \TeX\ Archive Network} {http://mirror.ctan.org/} \newacr{ukfaq}{UK FAQ} {UK List of \TeX\ Frequently Asked Questions} {http://www.tex.ac.uk/faq} \newcommand{\koma}{KOMA-Script} \input{glsentries} % shortcuts to other documents \newcommand*{\latexbook}{the \LaTeX\ user's guide~\cite{lamport94}} \newcommand*{\Latexbook}{The \LaTeX\ user's guide~\cite{lamport94}} \newcommand*{\latexcomp}{\emph{The \LaTeX\ Companion}~\cite{goossens94}} \newcommand*{\latexguide}{\emph{A Guide to \LaTeX}~\cite{kopka95}} \newcommand*{\baseurl}{http://www.dickimaw-books.com} \newcommand*{\packageurl}{\baseurl/latex/packages} \newcommand*{\packagecls}[1]{% \htmladdnormallink{\cls{#1}}{\packageurl/index.html\##1}} \newcommand*{\packagesty}[1]{% \htmladdnormallink{\sty{#1}}{\packageurl/index.html\##1}} \newcommand*{\downloadurl}{\baseurl/latex/thesis/examples} \newcommand*{\texdoc}{\htmladdnormallink{texdoc}{\baseurl/latex/novices/html/texdoc.html}} \indexpreamble{\latexhtml{Page numbers}{Locations} in \textbf{bold} indicate the entry definition in the summary.} \newcommand{\xtableref}[1]{\objectref{Table}{#1}} % \figconts[graphics opts]{image}{caption}{label} \newcommand{\figconts}[4][]{% \begin{makeimage}\end{makeimage}\figuretop{#4} \centering \incPgfOrGraphics[#1]{#2}% #3% caption \label{#4}% } \newcommand{\fboxfigconts}[4][]{% \begin{makeimage}\end{makeimage}\figuretop{#4} \centering \incFboxPgfOrGraphics[#1]{#2}% #3% caption \label{#4}% } \newcommand*{\backcovertext}{If you choose to buy a~copy of this book, Dickimaw Books asks for your support through buying the Dickimaw Books edition to help cover costs.} \copyrighttext{% Copyright \textcopyright\ 2007 Nicola L.~C. Talbot Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation; with no Invariant Sections, no Front-Cover Texts, and one Back-Cover Text: \dq{\backcovertext} A copy of the license is included in the section entitled \htmlref{\dq{GNU Free Documentation License}}{sec:fdl}. \doifnotbook {% The base URL for this document is: \url{\baseurl/latex/thesis/} }% } \html{\input{htmlonly}} % Need a long table for the required and optional fields (for screen format), but it's % confusing LaTeX2HTML. \newenvironment{FieldTabA4} {\begin{table}[htbp] \caption{Required and Optional Fields}\label{tab:reqopt} \centering \begin{tabular}{lp{0.4\textwidth}>{\raggedright}p{0.4\textwidth}} \bfseries Entry Type & \bfseries Required Fields & \bfseries Optional Fields\tabularnewline } { \end{tabular} \end{table} } \newenvironment{FieldTabScr} {\newpage\begin{longtable}{lp{0.4\textwidth}>{\raggedright}p{0.4\textwidth}} \caption{Required and Optional Fields}\label{tab:reqopt}\tabularnewline \bfseries Entry Type & \bfseries Required Fields & \bfseries Optional Fields\tabularnewline \endfirsthead \caption*{Required and Optional Fields Cont.}\tabularnewline \bfseries Entry Type & \bfseries Required Fields & \bfseries Optional Fields\tabularnewline \endhead \endfoot } {\end{longtable}} \newenvironment{FieldTab} {\ifscreen\begin{FieldTabScr}\else\begin{FieldTabA4}\fi} {\ifscreen\end{FieldTabScr}\else\end{FieldTabA4}\fi} \newenvironment{fieldtab}{\latexhtml{\begin{FieldTab}}{\begin{FieldTabA4}}}{\latexhtml{\end{FieldTab}}{\end{FieldTabA4}}} \begin{document} \maketitle \frontmatter \setcounter{tocdepth}{2} \setnode{contents} \tableofcontents \setnode{listoffigures} \listoffigures \setnode{listoftables} \listoftables \listofexamples \chapter{Abstract} This book is aimed at PhD students who want to use \LaTeX\ to typeset their PhD thesis. If you are unfamiliar with \LaTeX\, I~recommend that you first read Volume~1: \latexnovices. \begin{htmlonly} The examples given in this document can be download from the \htmladdnormallink{examples directory}{examples/index.html}. \end{htmlonly} \mainmatter %%%%%%%%%%%%%%%% INTRODUCTION %%%%%%%%%%%%%%%%%%%%%%%%%%%% \setnode{introduction} \chapter{Introduction} \label{ch:intro} Many PhD students in the sciences are encouraged to produce their PhD thesis in \LaTeX, particularly if their work involves a~lot of mathematics. In addition, these days, \LaTeX\ is no longer the sole province of mathematicians and computer scientists and is now starting to be used in the arts and social sciences (see, for example, some of the topics listed in the \TeX\ online catalogue~\cite{texcattopic}). This book is intended as a~brief guide on how to typeset the various components that are usually required for a~thesis. If you have never used \LaTeX\ before, I recommend that you first read Volume~1: \latexnovices, as this book assumes you have a~basic knowledge of \LaTeX. As with Volume~1, I'll be using PDF\LaTeX\ and TeXWorks. If you are creating a~DVI file or you are using a~different editor, you'll have to adapt the instructions. \warning If you are unfamiliar with terms such as \dq{preamble}, read \novices[ch:def]{definitions}. If you don't know how to find package documentation, read \novices{texdoc}. Throughout this document there are pointers to related topics in the \gls{ukfaq}. \latexhtml{These are displayed in the margin in square brackets, as illustrated on the \ifbookorother{\ifthispageodd{right}{left}}{right}.}{% These are displayed in the text like this:}% \faq{What is LaTeX?}{latex} You may find these resources useful in answering related questions that are not covered in this book. \doifbook{% To find the resources, go to \url{http://www.tex.ac.uk/faq} and either look for the question title in the list, or enter a~keyword in the search field.% } \latex{% On-line versions of this book, along with associated files, are available at: \url{\baseurl/latex/thesis/}. \doifnotbook {% The links in this document are colour-coded: internal links are blue, external links are magenta. }% } \latexhtml {% To refresh your memory or for those who haven't read Volume~1, throughout this book source code is illustrated in a~typewriter font with the word \inputlabelformat{Input} placed in the margin, and the corresponding output (how it will appear in the PDF document) is typeset with the word \outputlabelformat{Output} in the margin. \xminisec{Example:} A single line of code is displayed like this: }% {% To refresh your memory or for those who haven't read \htmladdnormallink{Volume~1}{\baseurl/latex/novices/}, throughout this book, source code is illustrated in the form:} \begin{codeS} This is an \gls{textbf}\marg{example}. \end{codeS}% The corresponding output\html{ (how it will appear in the PDF document)\footnote{This HTML version of the book uses bitmaps to illustrate the output, which doesn't look as good as the actual PDF version.}}\ is illustrated like this: \begin{resultS}[exampleoutput.html] This is an \textbf{example}. \end{resultS}% \begin{latexonly} Segments of code that are longer than one line are bounded above and below, illustrated as follows: \begin{code} \begin{alltt} Line one\gls{par} Line two\glsni{par} Line three. \end{alltt} \end{code}% with corresponding output: \begin{result} Line one\par Line two\par Line three. \end{result}% \end{latexonly}% \doifnotbook {% (Commands typeset in blue, such as \glsni{par}, indicate a hyperlink to the command definition in the \latexhtml{\htmlref{summary}{ch:glossary}}{\htmladdnormallink{summary}{summary.html}}.) } Command definitions are shown in a~typewriter font in the form: \begin{definition} \glsni{documentclass}\gls{opt.opensq}\meta{options}\gls{opt.closesq}\gls{leftbracechar}\meta{class file}\gls{rightbracechar} \end{definition}% In this case the command being defined is called \cmdname{documentclass} and text typed \meta{like this} (such as \meta{options} and \meta{class file}) indicates the type of thing you need to substitute. (Don't type the angle brackets!) For example, if you want the \icls{scrbook} class file you would substitute \meta{class file} with \texttt{scrbook} and if you want the \clsopt{letterpaper} option you would substitute \meta{options} with \texttt{letterpaper}, like this: \begin{codeS} \glsnl{documentclass}\oarg*{letterpaper}\marg*{scrbook} \end{codeS} When it's important to indicate a~space, the visible space symbol~\gls{visiblespace} is used. For example: \begin{codeS} A\glsni{visiblespace}sentence\glsni{visiblespace}consisting\glsni{visiblespace}of\glsni{visiblespace}six\glsni{visiblespace}words. \end{codeS} When you type up the code, replace any occurrences of \glsni{visiblespace} with a~space. \xminisec{Note:} \warning Be careful of the dangers of obsolete code propagation. It often happens that students pass on their \LaTeX\ code to new students who, in their turn, pass it on to the next lot of students, and so on. You're told \dq{use this magic bit of code to format your thesis} without knowing what it does. Ancient buggy code that's 20~years out-of-date festers in university departments refusing to die. But if it worked for previous students, what's the problem? The problem is that it may stop working a~week before your submission date and when you go for help, you may be told you're using obsolete packages and there's nothing for it but to rewrite your thesis using the modern alternatives. How do you know if a~package is obsolete? Some of the obsolete packages and commands are listed in l2tabu~\cite{l2tabu}, or you can check to see if a~package is listed in \gls{ctan}['s] obsolete tree (\url{http://mirror.ctan.org/obsolete/}). Stefan~Kottwitz also has a~list of obsolete classes and packages in his \htmladdnormallinkfoot{TeXblog}{http://texblog.net/latex-articles/packages/}. The other thing to do is check the package's entry on \gls{ctan} to see if it has been deprecated. For example, suppose someone tells you to use the \sty{glossary} package. If you go to \url{http://ctan.org/pkg/glossary} it will tell you that the \sty{glossary} package is no longer supported and that it's been replaced by the \sty{glossaries} package. Similarly, if you go to \url{http://ctan.org/pkg/epsfig} it will tell you that the \sty{epsfig} package is obsolete and you should use \sty{graphicx} instead. \setnode{build} \section{Building Your Document} \label{sec:build} To \dq{typeset}, \dq{build}, \dq{compile} or \dq{LaTeX} your document means to run the \iappname{pdflatex} (or \iappname{latex}) executable on your document source code. If you are using a front-end, such as TeXworks, WinEdt, TeXstudio, or TeXnicCenter, this usually just means clicking on the appropriate button or selecting the appropriate menu item. (See \novices[ch:tex2pdf]{fromsource2output} for further details.) It's important to remember that a~front-end is an \keyword{interface}. It's not, for example, TeXworks that is creating your PDF. When you click on the \dq{typeset} button, TeXworks tells the operating system to run the required executable. This is usually \appname{pdflatex}, but there are other executables that may need to be used to help create your document, such as \iappname{bibtex} or \iappname{biber} (discussed in \chapterref{ch:citations}) and \iappname{makeindex} or \iappname{xindy} (discussed in \chapterref{ch:indgloss}). For example, if your document has a~bibliography and you are using TeXworks, you first need to make sure the drop-down menu is set to \dq{pdfLaTeX} (see \figureref{fig:texworks-pdflatex}) and click on the green \dq{Typeset} button. Then you need to select \dq{BibTeX} from the drop-down menu (see \figureref{fig:texworks-bibtex}) and click on the green \dq{Typeset} button. Then again select \dq{pdfLaTeX} (\figureref{fig:texworks-pdflatex}) and click the \dq{Typeset} button. Finally, to ensure your cross-references are all up-to-date, you need to click on the \dq{Typeset} button again. If you are using \iappname{biber} instead of \iappname{bibtex} (see \sectionref{sec:biblatex}), then you have to replace the above \dq{BibTeX} step with \dq{Biber} instead. \begin{figure}[htbp] \figconts {pictures/texworks-pdflatex} {\caption{Selecting pdfLaTeX from the Drop-Down Menu}} {fig:texworks-pdflatex} \end{figure} \begin{figure}[htbp] \figconts {pictures/texworks-bibtex} {\caption{Selecting BibTeX from the Drop-Down Menu}} {fig:texworks-bibtex} \end{figure} If the tool you require isn't listed in the drop-down box, you will have to add it. For example, to add \iappname{makeglossaries} to the list of available tools in TeXworks, you need to select \menu{Edit}\menuto\menu{Preferences}, which will open the \dq{TeXworks Preferences} dialog. Make sure the \dq{Typesetting} tab is selected and click on the lower \incGraphics[alt=+,height=2ex]{pictures/texworks-addbutton} button next to the \dq{Processing tools} list. This will open the \dq{Tool Configuration} dialog. Set the \dq{Name} field to the name of the application, as you want it to appear in the tool list (for example \dq{MakeGlossaries}). Then click on the \dq{Browse} button to find the application on your computer. Next you need to click on the \incGraphics[alt=+,height=2ex]{pictures/texworks-addbutton} button next to the \dq{Arguments} list. Set the argument to \verb|$basename|. Since \appname{makeglossaries} doesn't modify the PDF, uncheck the \dq{View PDF after running} box (see \figureref{fig:texworks-makeglossaries}). \begin{figure}[htbp] \figconts {pictures/texworks-makeglossaries} {\caption{Adding Makeglossaries to the list of tools in TeXworks}} {fig:texworks-makeglossaries} \end{figure} This is a~bit of a~hassle (if not downright confusing for a beginner) and even more so when you have glossaries and an index in your document as well as a~bibliography. Fortunately there are ways of automating this process so that you only need one button press to perform all those different steps. There are several applications available to do this for you, and I~strongly recommend you try one of them, if possible, to reduce the complexity involved in building a~document. \novices{crossref} mentioned \iappname{latexmk}, which is available on \gls{ctan}. This is a~Perl script, so it will run on any operating system that has Perl installed (see \novices{perl}). Since Volume~1 was published, a~Java alternative called \iappname{arara} has arrived on \gls*{ctan}. Java applications will run on any operating system that has the \htmladdnormallinkfoot{Java Runtime Environment}{http://www.java.com/getjava/} installed, so both \appname{latexmk} and \appname{arara} are multi-platform solutions to automated document compilation. \sectionref{sec:latexmk} gives a~brief introduction to \appname{latexmk}, and \sectionref{sec:arara} gives a~brief introduction to \appname{arara}. \setnode{latexmk} \subsection{LaTeXmk} \label{sec:latexmk} As mentioned \latexhtml{above}{\htmlref{in the previous section}{sec:build}}, \iappname{latexmk} is a~Perl script that automates the process of building a~\LaTeX\ document. In order to use \appname{latexmk}, you must have Perl installed (see \novices{perl}). Both TeX~Live and MikTeX come with \appname{latexmk} but, if for some reason you don't have it installed, you can use the TeX~Live or MikTeX update manager to install it. Alternatively, you can download \url{http://mirror.ctan.org/support/latexmk.zip} and install it manually. Once \appname{latexmk} is installed, you then need to add it to the list of available tools in TeXworks\footnote{If you are using a different front-end, you will have to consult your front-end's manual.}. This is done via the \menu{Edit}\menuto\menu{Preferences} menu item. This opens TeXwork's Preferences dialog box. Make sure the \dq{Typesetting} tab is selected (\figureref{fig:texworks-pref}). \begin{figure}[htbp] \figconts {pictures/texworks-pref} {\caption{TeXwork's Preferences Dialog Box}} {fig:texworks-pref} \end{figure} To add a~new tool, click on the lower \incGraphics[alt=+,height=2ex]{pictures/texworks-addbutton} button next to the list of processing tools. This opens the tool configuration dialog box (\figureref{fig:texworks-latexmkbibtex}). \begin{figure} \figconts {pictures/texworks-latexmkbibtex} {\caption{Adding LaTeXmk in the TeXWorks Tool Configuration Dialog}} {fig:texworks-latexmkbibtex} \end{figure} Type \dq{LaTeXmk} in the \dq{Name} box, then use the \dq{Browse} button to locate \appname{latexmk} on your computer. Next you need to click on the \incGraphics[alt=+,height=2ex]{pictures/texworks-addbutton} button to add each argument. The argument list should consist of the following (in the order listed): \begin{verbatim} -e $pdflatex=q/pdflatex $synctexoption %O %S/ -pdf -bibtex $fullname \end{verbatim} Once you've done this, click \dq{Okay} to close the tool configuration dialog, and click \dq{Okay} to close the Preferences dialog box. LaTeXmk should now be listed in the drop-down menu next to the green \dq{Typeset} button. Now, if you have LaTeXmk selected and you click on the \dq{Typeset} button \appname{pdflatex} and \appname{bibtex}\slash\appname{biber} will be run as necessary to create an up-to-date PDF. Unfortunately, adding \iappname{makeindex}, \iappname{texindy} or \iappname{makeglossaries} to LaTeXmk's set of rules is more complicated. For this you need to create a~configuration\slash initialisation (RC) file\footnote{There are some example RC files available at: \url{http://mirror.ctan.org/support/latexmk/example_rcfiles/}.}. The name and location of this file depends on your operating system. For example, on a~Unix-like operating system, this may be \verb|$HOME/.latexmkrc|. You will need to consult the \appname{latexmk} manual~\cite{latexmk} for further details. Once you've found out the name and location of the RC file for your operating system, you can use the text editor of your choice to create this file. To add \iappname{makeglossaries}, you need to type the following in the RC file: \begin{verbatim} add_cus_dep('glo', 'gls', 0, 'makeglossaries'); add_cus_dep('acn', 'acr', 0, 'makeglossaries'); sub makeglossaries{ system( "makeglossaries \"$_[0]\"" ); } \end{verbatim} To add \iappname{makeindex}, you need to type the following: \begin{verbatim} add_cus_dep('idx', 'ind', 0, 'makeindex'); sub makeindex{ system("makeindex \"$_[0].idx\""); } \end{verbatim} If you prefer to use \iappname{texindy} instead of \appname{makeindex}, you will need to replace the above lines with (change the language as appropriate): \begin{verbatim} add_cus_dep('idx', 'ind', 0, 'texindy'); sub texindy{ system("texindy -L english \"$_[0].idx\""); } \end{verbatim} Now select \dq{LaTeXmk} from the drop-down menu next to the green \dq{Typeset} button in TeXworks (\figureref{fig:texworks-latexmk}), and you're ready to build your documents. \begin{figure} \figconts {pictures/texworks-latexmk} {\caption{LaTeXmk Tool Selected in TeXworks}} {fig:texworks-latexmk} \end{figure} \setnode{arara} \subsection{Arara} \label{sec:arara} As mentioned in \sectionref{sec:build}, \iappname{arara} is a~Java application that automates the process of building a~\LaTeX\ document. In order to use \appname{arara}, you must have the \htmladdnormallink{Java Runtime Environment}{http://www.java.com/getjava/} installed. The latest TeX~Live distribution includes \appname{arara}, so you can install it via the TeX~Live package manager. Alternative, you can install \appname{arara} manually as follows: fetch the installer \htmladdnormallink{\texttt{arara-3.0\booklinebreak-installer.jar}}{https://github.com/cereda/arara/blob/master/releases/arara-3.0-installer.jar} (or \htmladdnormallink{\texttt{arara-3.0-installer.exe}}{https://github.com/cereda/arara/blob/master/releases/arara-3.0-installer.exe}) from \url{https://github.com/cereda/arara/tree/master/releases}. On Windows, run \reportlinebreak\screenlinebreak\texttt{arara-3.0-installer.exe}. On other operating systems run \reportlinebreak\texttt{arara-3.0-\screenlinebreak installer.jar} in privileged mode. For example, on a~Unix-based system: \begin{verbatim} sudo java -jar arara-3.0-installer.jar \end{verbatim} (If you are doing a~manual install make sure you check the box to add the predefined rules, as shown in \figureref{fig:arara-installer}.) \begin{figure} \figconts {pictures/arara-installer} {\caption{Arara Installer}} {fig:arara-installer} \end{figure} Once \appname{arara} has been installed, you can add it to the list of tools in TeXworks. As \htmlref{before}{sec:latexmk}, open the TeXwork's Preferences dialog box using \menu{Edit}\menuto\menu{Preferences} and select the \dq{Typesetting} tab (\figureref{fig:texworks-pref}). To add a~new tool, click on the lower \incGraphics[alt=+,height=2ex]{pictures/texworks-addbutton} button next to the list of processing tools. This opens the tool configuration dialog box (\figureref{fig:texworks-arara}). Type \dq{Arara} in the \dq{Name} box and use the \dq{Browse} button to find the \appname{arara} application on your computer. Use the \incGraphics[alt=+,height=2ex]{pictures/texworks-addbutton} button to add \verb|$basename| to the list of arguments, as shown in \figureref{fig:texworks-arara}. \begin{figure} \figconts {pictures/texworks-arara} {\caption{Adding Arara in the TeXWorks Tool Configuration Dialog}} {fig:texworks-arara} \end{figure} Unlike \iappname{latexmk}, \appname{arara} doesn't read the log file to determine what applications need to be run. Instead, you tell \appname{arara} how to build your document by placing special comments in your source code. For example, if your document contains the following: \begin{code} \begin{alltt} \gls{percentchar.arara} pdflatex: \marg{ synctex: on } \glsni{percentchar.arara} bibtex \glsni{percentchar.arara} pdflatex: \marg{ synctex: on } \glsni{percentchar.arara} pdflatex: \marg{ synctex: on } \glsni{documentclass}\marg{scrbook} \end{alltt} \end{code} Then running \appname{arara} on the document will run \appname{pdflatex}, \appname{bibtex}, \appname{pdflatex} and \appname{pdflatex} on your document. \appname{Arara} knows the rules \dq{pdflatex} and \dq{bibtex}. It also knows the rules \dq{biber}, \dq{makeglossaries} and \dq{makeindex}. So, if your document has a~bibliography, an index and glossaries, you need to put the following comments in your source code (replace \texttt{bibtex} with \texttt{biber} if required): \begin{code} \glsni{percentchar.arara} pdflatex: \marg{ synctex: on }\newline \glsni{percentchar.arara} bibtex\newline \glsni{percentchar.arara} makeglossaries\newline \glsni{percentchar.arara} makeindex\newline \glsni{percentchar.arara} pdflatex: \marg{ synctex: on }\newline \glsni{percentchar.arara} pdflatex: \marg{ synctex: on }\newline \glsni{documentclass}\marg{scrbook} \end{code} Now you just need to select \dq{Arara} from the drop-down list in TeXworks (\figureref{fig:texworks-arara2}) and click the green \dq{Typeset} button, and \appname{arara} will do all the work for you. \xminisec{Note:} \warning If you don't add these arara comments to your source code, nothing will happen when you run \appname{arara} on your document! You must remember to provide \appname{arara} with the rules to build your document. \begin{figure} \figconts {pictures/texworks-arara2} {\caption{Using Arara in TeXworks}} {fig:texworks-arara2} \end{figure} Unfortunately \appname{arara} (v3.0) doesn't have a~rule for \iappname{texindy}, but you can add one by creating a~file called \texttt{texindy.yaml} that contains the following:\footnote{Thanks to Paulo Cereda for supply this.}\screenpagebreak \begin{alltt} !config \# TeXindy rule for arara \# requires arara 3.0+ identifier: texindy name: TeXindy command: texindy @\{german\} @\{language\} @\{codepage\} @\{module\}\bookcontinueline\reportcontinueline\screencontinueline @\{input\} @\{options\} "@\{getBasename(file)\}.idx" arguments: - identifier: german flag: @\{isTrue(parameters.german,"-g")\} - identifier: language flag: -L @\{parameters.language\} - identifier: codepage flag: -C @\{parameters.codepage\} - identifier: module flag: -M @\{parameters.module\} - identifier: input flag: -I @\{parameters.input\} - identifier: options flag: @\{parameters.options\} \end{alltt} \latex{(The symbol \continuesymbol\ above indicates a~line wrap. Don't insert a~line break at that point.) }This file should be saved in the \texttt{rules} subdirectory of the \texttt{arara} installation directory. (For example, on Unix-like systems \texttt{/usr/local/arara/\reportlinebreak\screenlinebreak rules/texindy.yaml}.) So if you'd rather use \appname{texindy} instead of \appname{makeindex} you can replace the \begin{alltt} \gls{percentchar.arara} makeindex \end{alltt} directive with \begin{alltt} \glsni{percentchar.arara} texindy: \marg*{ language: english, codepage: latin1 } \end{alltt} (Change the language and encoding as appropriate.) \setnode{start} \chapter{Getting Started} \label{ch:start} There are many different thesis designs, varying according to university or discipline~\cite{thesistemplates}. If you have been told to use a~particular class file, use that one. If not, there are a~selection of thesis class files available on \gls{ctan} and listed in the \htmladdnormallinkfoot{OnLine \TeX\ Catalogue's Topic Index}{http://mirror.ctan.org/help/Catalogue/bytopic.html\#theses}~\cite{texcattopic}. Since there are so many to choose from, I'm just going to follow on from \htmladdnormallink{Volume~1}{\baseurl/latex/novices/} of this series and use one of the \koma\ class files. But which one? The \icls{scrreprt} class is the one usually recommended for a~report or thesis. It defaults to one-sided and has an \gls{env-abstract} environment, but it doesn't define \gls{frontmatter}, \gls{mainmatter} or \gls{backmatter}. The \icls{scrbook} class does define those commands, but it doesn't provide an \glsni{env-abstract} environment and defaults to two-sided layout. So, you can either do: \begin{code} \begin{alltt} \gls{documentclass}\marg{scrreprt} \gls{title}\marg{A Sample Thesis} \gls{author}\marg{A.N. Other} \gls{begin}\marg{document} \gls{maketitle} \gls{pagenumbering}\marg{roman} \gls{tableofcontents} \gls{chapter}*\marg{Acknowledgements} \glsni{begin}\marg{abstract} This is the abstract \gls{end}\marg{abstract} \glsni{pagenumbering}\marg{arabic} \glsni{chapter}\marg{Introduction} ... \glsni{end}\marg{document} \end{alltt} \end{code} \bookpagebreak\noindent or you can do: \begin{code} \begin{alltt} \gls{documentclass}\oarg{oneside}\marg{scrbook} \gls{title}\marg{A Sample Thesis} \gls{author}\marg{A.N. Other} \glsni{begin}\marg{document} \gls{maketitle} \gls{frontmatter} \gls{tableofcontents} \glsni{chapter}\marg{Acknowledgements} \glsni{chapter}\marg{Abstract} This is the abstract \gls{mainmatter} \glsni{chapter}\marg{Introduction} ... \glsni{end}\marg{document} \end{alltt} \end{code} I'm going to use the second approach simply out of personal preference. The \koma\ options mentioned in this book are available for both \icls{scrreprt} and \icls{scrbook}, so choose whichever class file you feel best suits your thesis. Unless you have been told otherwise, I~recommend that you start out with a~skeletal document that looks something like the following: \begin{codelisting}{thesis1.tex}\label{ex:thesis1} \begin{alltt} \gls{documentclass}\oarg{oneside}\marg{scrbook} \gls{title}\marg{A Sample Thesis} \gls{author}\marg{A.N. Other} \gls{date}\marg{July 2013} \gls{titlehead}\marg{A Thesis submitted for the degree of Doctor of Philosophy} \gls{publishers}\marg{School of Something\gls{dbbackslashchar}University of Somewhere} \glsni{begin}\marg{document} \gls{maketitle} \gls{frontmatter} \gls{tableofcontents} \gls{listoffigures} \gls{listoftables} \gls{chapter}\marg{Acknowledgements} I would like to thank my supervisor, Professor Someone. This research was funded by the Imaginary Research Council. \gls{chapter}\marg{Abstract} A brief summary of the project goes here. \gls{percentchar} A glossary and list of acronyms may go here \glsni{percentchar} or may go in the back matter. \gls{mainmatter} \gls{chapter}\marg{Introduction} \gls{label}\marg{ch:intro} \gls{chapter}\marg{Technical Introduction} \glsni{label}\marg{ch:techintro} \gls{chapter}\marg{Method} \glsni{label}\marg{ch:method} \gls{chapter}\marg{Results} \glsni{label}\marg{ch:results} \gls{chapter}\marg{Conclusions} \glsni{label}\marg{ch:conc} \gls{backmatter} \glsni{percentchar} A glossary and list of acronyms may go here \glsni{percentchar} or may go in the front matter after the abstract. \glsni{percentchar} The bibliography will go here \glsni{end}\marg{document} \end{alltt} \end{codelisting} If you do this, it will help ensure that your document has the correct structure before you begin with the actual contents of the document. (Note that the chapter titles will naturally vary depending on your subject or institution, and you may need a different paper size if you are not in Europe. I~have based the above on my own PhD thesis which I~wrote in the early to mid 1990s in the Department of Electronic Systems Engineering at the University of Essex, and it may well not fit your own requirements.) If you haven't started your thesis yet, go ahead and try this. Creating a~skeletal document can have an amazing psychological effect on some people: for very little effort it can produce a document several pages long, which can give you a~sense of achievement that can help give you sufficient momentum to get started (but of course, it's not guaranteed to work with everyone). Remember that if you want to use \iappname{arara} (see \sectionref{sec:arara}) you must add the build rules to the document: \begin{code} \begin{alltt} \gls{percentchar.arara} pdflatex: \marg{ synctex: on } \glsni{percentchar.arara} pdflatex: \marg{ synctex: on } \glsni{documentclass}\oarg{oneside}\marg{scrbook} \end{alltt} \end{code} (I'll add the \appname{arara} rules to sample listings, in the event that you want to use \appname{arara}. Since they are comments, they will be ignored if you use \iappname{pdflatex} explicitly or if you use another automation method, such as \iappname{latexmk}.) Now think about other requirements. What font size have you been told to use? \begin{description} \item[10pt] Use the \clsopt{10pt} class option: \begin{codeS} \gls{documentclass}\oarg{oneside,10pt}\marg{scrbook} \end{codeS} \item[11pt] Use the \clsopt{11pt} class option: \begin{codeS} \gls{documentclass}\oarg{oneside,11pt}\marg{scrbook} \end{codeS} \item[12pt] Use the \clsopt{12pt} class option: \begin{codeS} \gls{documentclass}\oarg{oneside,12pt}\marg{scrbook} \end{codeS} \end{description} Have you been told to have a~blank line between paragraphs and no paragraph indentation? If so, use the \scrclsopt[full]{parskip} class option: \begin{codeS} \gls{documentclass}\oarg{oneside,12pt,parskip=full}\marg{scrbook} \end{codeS} \label{geometry}% \faq{Changing the margins in \LaTeX}{changemargin}Have you been told to have certain sized margins? If so, you can use the \isty{geometry} package. For example, if you have been told you must have 1~inch margins, you can do \begin{codeS} \gls{usepackage}\oarg{margin=1in}\marg{geometry} \end{codeS} Changing the default fonts is covered in \novices[sec:changingfonts]{documentfonts}. Other possible formatting requirements, such as double-spacing, are covered in \chapterref{ch:formatting}. \setnode{include} \chapter{Splitting a~Large Document into Several Files} \label{ch:include} Some people prefer to place each chapter of a~large document in a separate file and then input the file into the main document. There are two basic ways of including the contents of an external file: \begin{definition} \gls{input}\marg{\meta{filename}} \end{definition} and \begin{definition} \gls{include}\marg{\meta{filename}} \end{definition} where \meta{filename} is the name of the file. (The \texttt{.tex} extension may be omitted in both cases.) The differences between the two commands are as follows: \begin{description} \item[\glsni{input}] acts as though the contents of the file were typed where the \glsni{input} command was. For example, suppose my main file contained the following: \begin{code} \begin{alltt} Here is a short paragraph. \glsni{input}\marg{myfile} \end{alltt} \end{code} and suppose the file \texttt{myfile.tex} contained the following lines: \begin{code} \begin{alltt} Here is some sample text. \end{alltt} \end{code} then the \glsni{input} command behaves as though you had simply typed the following in your main document file: \begin{code} \begin{alltt} Here is a short paragraph. Here is some sample text. \end{alltt} \end{code} \item[\glsni{include}] does more than just input the contents of the file. It also starts a~new page (using \gls{clearpage}) and creates an auxiliary file associated with the included file. It also issues another \glsni{clearpage} once the file has been read in. Using this approach, you can also govern which files to include using \begin{definition} \gls{includeonly}\marg{\meta{file list}} \end{definition} in the preamble, where \meta{file list} is a~comma-separated list of files you want included. This way, if you only want to work on one or two chapters, you can only include those chapters, which will speed up the document build. \LaTeX\ will still read in all the cross-referencing information for the missing chapters, but won't include those chapters in the PDF file. There is a~definite advantage to this if you have, say, a~large number of images in your results chapter, which you don't need when you're working on, say, the technical introduction. You can still reference all the figures in the omitted chapter, as long as you have previously \htmladdnormallink{\LaTeX{}ed the document}{\baseurl/latex/novices/html/fromsource2output.html\#itm:step2} without the \glsni{includeonly} command. The \sty{excludeonly} package provides the logically opposite command: \begin{definition} \gls{excludeonly}\marg{\meta{file list}} \end{definition} \end{description} The \htmlref{previous example}{ex:thesis1}\doifbook{ \vpageref{ex:thesis1}} can now be split into various files: \begin{codelisting}[\texttt{thesis.tex}]{thesis2.tex}\label{ex:thesis2} \begin{alltt} \glsni{percentchar.arara} pdflatex: \marg{ synctex: on } \glsni{percentchar.arara} pdflatex: \marg{ synctex: on } \gls{documentclass}\oarg{oneside}\marg{scrbook} \gls{title}\marg{A Sample Thesis} \gls{author}\marg{A.N. Other} \gls{date}\marg{July 2013} \gls{titlehead}\marg{A Thesis submitted for the degree of Doctor of Philosophy} \gls{publishers}\marg{School of Something\gls{dbbackslashchar}University of Somewhere} \glsni{begin}\marg{document} \gls{maketitle} \gls{frontmatter} \gls{tableofcontents} \gls{listoffigures} \gls{listoftables} \gls{chapter}\marg{Acknowledgements} I would like to thank my supervisor, Professor Someone. This research was funded by the Imaginary Research Council. \gls{chapter}\marg{Abstract} A brief summary of the project goes here. \gls{mainmatter} \glsni{include}\marg{intro} \glsni{include}\marg{techintro} \glsni{include}\marg{method} \glsni{include}\marg{results} \glsni{include}\marg{conc} \gls{backmatter} \glsni{end}\marg{document} \end{alltt} \end{codelisting} \begin{codelisting}[\texttt{intro.tex}]{intro.tex} \begin{alltt} \gls{chapter}\marg{Introduction} \gls{label}\marg{ch:intro} \end{alltt} \end{codelisting} \begin{codelisting}[\texttt{techintro.tex}]{techintro.tex} \begin{alltt} \gls{chapter}\marg{Technical Introduction} \gls{label}\marg{ch:techintro} \end{alltt} \end{codelisting} \begin{codelisting}[\texttt{method.tex}]{method.tex} \begin{alltt} \gls{chapter}\marg{Method} \gls{label}\marg{ch:method} \end{alltt} \end{codelisting} \begin{codelisting}[\texttt{results.tex}]{results.tex} \begin{alltt} \gls{chapter}\marg{Results} \gls{label}\marg{ch:results} \end{alltt} \end{codelisting} \begin{codelisting}[\texttt{conc.tex}]{conc.tex} \begin{alltt} \gls{chapter}\marg{Conclusions} \gls{label}\marg{ch:conc} \end{alltt} \end{codelisting} If you only want to work on, say, the Method and Results chapters, you can place the following command in the preamble: \begin{codeS} \glsni{includeonly}\marg{method,results} \end{codeS} % Formatting \setnode{formatting} \chapter{Formatting} \label{ch:formatting} It used to be that in order to change the format of chapter and section headings, you needed to have some understanding of the internal workings of classes such as \icls{report} or \icls{book}. Modern classes, such as \icls{memoir} and the \koma\ classes, provide a~much easier interface. However, I~recommend that you first write your thesis, and then worry about changing the document style. The ability to separate content from style is one of the advantages of using \LaTeX\ over a~word processor. Remember that writing your thesis is more important than the layout. Whilst it may be that your school or department insists on a~certain style, it should not take precedence over the actual task of writing. \setnode{docstyles} \section{Changing the Document Style} \label{sec:docstyles} If you are using a~custom thesis class file provided by your department or school, then you should stick to the styles set up in that class. If not, you may need to change the default style of your chosen class to fit the requirements. \novices[sec:section]{sectionunits} described how to change the fonts used by chapter and section headings for the \koma\ classes. For example, if the chapter headings must be set in a~large, bold, serif font you can do: \begin{codeS} \gls{addtokomafont}\marg{\gls{large}\gls{bfseries}\gls{rmfamily}} \end{codeS} The headings in the \koma\ classes default to ragged-right justification (recall \gls{raggedright} from \xrsectionref{Declarations}{declaration} of Volume~1) which is done via \begin{definition} \gls{raggedsection} \end{definition} This can be redefined as required. For example, suppose you are required to have centred headings, then you can do: \begin{codeS} \gls{renewcommand}*\marg{\glsni{raggedsection}}\marg{\gls{centering}} \end{codeS} \setnode{newpagestyle} \section{Changing the Page Style} \label{sec:newpagestyle} \novices{pagestyle} described the command \begin{definition} \gls{pagestyle}\marg{\meta{style}} \end{definition} which can be used to set the page style. The \icls{scrbook} class defaults to the \ipagestyle{headings} page style, but if this isn't appropriate, you can use the \isty{scrpage2} package, which comes with the \koma\ bundle. This package provides its own versions of the \ipagestyle{plain} and \pagestylefmt{headings} page styles, called \ipagestyle{scrplain} and \ipagestyle{scrheadings}. For simplicity, I'm assuming that your thesis is a~one-sided document. If this isn't the case and your odd and even page styles need to be different, you'll need to consult the \koma\ documentation~\cite{koma}. With the \pagestylefmt{scrheadings} page style, the page header and footer are both divided into three areas (\figureref{fig:pagestyle}): the inner (left) head/foot, the centre head/foot and the outer (right) head/foot. \begin{figure}[htbp] \figconts {pictures/pagestyle} {\caption{Page Header and Footer Elements}} {fig:pagestyle} \end{figure} These elements can be set using: \begin{definition} \gls{ihead}\oarg{\meta{scrplain inner head}}\marg{\meta{scrheadings inner head}}\\ \gls{chead}\oarg{\meta{scrplain centre head}}\marg{\meta{scrheadings centre head}}\\ \gls{ohead}\oarg{\meta{scrplain outer head}}\marg{\meta{scrheadings outer head}}\\ \gls{ifoot}\oarg{\meta{scrplain inner foot}}\marg{\meta{scrheadings inner foot}}\\ \gls{cfoot}\oarg{\meta{scrplain centre foot}}\marg{\meta{scrheadings centre foot}}\\ \gls{ofoot}\oarg{\meta{scrplain outer foot}}\marg{\meta{scrheadings outer foot}} \end{definition} In each case, the optional argument indicates what to do if the \pagestylefmt{scrplain} page style is in use and the mandatory argument indicates what to do if the \pagestylefmt{scrheadings} page style is in use. (If the optional argument is missing, no modification is made to the \pagestylefmt{scrplain} style.) Within both types of argument, you can use \begin{definition} \gls{pagemark} \end{definition} to insert the current page number and \begin{definition} \gls{headmark} \end{definition} to insert the running heading. For example, suppose you are required to put your registration number on the bottom left of each page and the page number on the bottom right, and you are also required to put the current chapter or section heading at the top left of each page, unless it's the first page of a~chapter. Then you can do: \begin{codelisting}{thesis-pagestyles.tex} \begin{alltt} \glsni{usepackage}\marg{scrpage2} \gls{pagestyle}\marg{scrheadings} \gls{newcommand}\marg{\cmdname{myregnum}}\marg{123456789}\glsni{percentchar} registration number \glsni{ihead}\marg{} \glsni{chead}\marg{} \glsni{ohead}\oarg{}\marg{\glsni{headmark}} \glsni{ifoot}\oarg{\cmdname{myregnum}}\marg{\cmdname{myregnum}}\glsni{percentchar} registration number \glsni{cfoot}\oarg{}\marg{} \glsni{ofoot}\oarg{\glsni{pagemark}}\marg{\glsni{pagemark}} \end{alltt} \end{codelisting} Note that the above don't use any font changing commands. If you want to change the font for the header and footer, you need to redefine \gls{headfont}. The page number style is given by \gls{pnumfont}. So for italic headers and footers with bold page numbers, you can redefine these commands as follows: \begin{code} \glsni{renewcommand}*\marg{\glsni{headfont}}\marg{\gls{normalfont}\gls{itshape}}\newline \glsni{renewcommand}*\marg{\glsni{pnumfont}}\marg{\glsni{normalfont}\glsni{bfseries}} \end{code} \setnode{setspace} \section{Double-Spacing} \label{sec:setspace} Whilst double-spacing is usually frowned upon in the world of modern typesetting, it is usually a~requirement for anything that may need hand-written annotations, which can include theses. This extra space gives the examiners room to write comments.\footnote{Despite the current digital age, many people still use hand-written annotations on manuscripts. It's unlikely that your examiners have pens that are incompatible with your paper.} Double-spacing can be achieved via the \isty{setspace} package. You can either set the spacing using the package options \istyopt{setspace}{singlespacing}, \istyopt{setspace}{onehalfspacing} or \istyopt{setspace}{doublespacing}, or you can switch via the declarations: \begin{definition} \gls{singlespacing}\newline \gls{onehalfspacing}\newline \gls{doublespacing} \end{definition} So, if your thesis has to be double-spaced, you can do: \begin{codelisting}{thesis-doublespaced.tex} \begin{alltt} \glsni{usepackage}\oarg{doublespacing}\marg{setspace} \end{alltt} \end{codelisting} \setnode{titlepagelayout} \section{Changing the Title Page} \label{sec:titlepage} \novices{title} described how to lay out the title page using \gls{maketitle}. If this layout isn't appropriate for your school or department's specifications, you can lay out the title page manually using the \gls{env-titlepage} environment instead of \gls{maketitle}. Within this environment, you can use \gls{hspace}\marg{\meta{length}} and \gls{vspace}\marg{\meta{length}} to insert horizontal and vertical spacing. (The unstarred versions are ignored if they occur at the start of a~line or page, respectively. The starred versions will insert the given spacing, regardless of their location.) You can also use \gls{hfill} and \gls{vfill}, which will expand to fill the available space horizontally or vertically, respectively. \bookpagebreak \xminisec{Example:} \begin{codelisting}{thesis-titlepage.tex} \begin{alltt} \glsni{begin}\marg{titlepage} \gls{centering} \glsni{vspace}*\marg{1in} \glsni{begin}\marg{Large}\glsni{bfseries} A Sample PhD Thesis\glsni{par} \glsni{end}\marg{Large} \glsni{vspace}\marg{1.5in} \glsni{begin}\marg{large}\glsni{bfseries} A. N. Other\glsni{par} \glsni{end}\marg{large} \glsni{vfill} A Thesis submitted for the degree of Doctor of Philosophy \glsni{par} \glsni{vspace}\marg{0.5in} School of Something \glsni{par} University of Somewhere \glsni{par} \glsni{vspace}\marg{0.5in} July 2013 \glsni{par} \glsni{end}\marg{titlepage} \end{alltt} \end{codelisting} The result is shown in \figureref{fig:titlepage}. (If you require \htmlref{double-spacing}{sec:setspace}, you may need to wait until after the title page before switching to double-spacing.) \begin{figure}[htbp] \fboxfigconts {pictures/titlepage} {\caption{Sample Title Page}} {fig:titlepage} \end{figure} \setnode{verbatim} \section{Listings and Other Verbatim Text} \label{sec:verbatim} \faq{Code listings in \LaTeX}{codelist}There may be times when you want to include text exactly as you have typed it into your source code. For example, you may want to include a~short segment of computer code. This can be done using the \gls{env-verbatim} environment. \xminisec{Example:} Note how I~don't need to worry about \htmladdnormallink{special characters}{\baseurl/latex/novices/html/symbols.html}, such as \gls{hashchar}, within the \glsni{env-verbatim} environment: \begin{code} \begin{alltt} \glsni{begin}\marg{verbatim} \#include /* needed for printf */ int main() \marg*{ printf("Hello World\cmdname{n}"); return 1; } \glsni{end}\marg{verbatim} \end{alltt} \end{code} This just produces: \begin{result}[verbatim.html] \begin{verbatim} #include /* needed for printf */ int main() { printf("Hello World\n"); return 1; } \end{verbatim} \end{result} A more sophisticated approach is to use the \isty{listings} package. With this package, you first need to specify the programming language. For example, the above code is in C, so I~need to specify this using: \begin{codeS} \gls{lstset}\marg{language=C} \end{codeS} Now I~can use the \gls{env-lstlisting} environment to typeset my C~code: \begin{code} \begin{alltt} \glsni{begin}\marg{lstlisting} \#include /* needed for printf */ int main() \marg*{ printf("Hello World\cmdname{n}"); return 1; } \glsni{end}\marg{lstlisting} \end{alltt} \end{code} The resulting output looks like: \begin{result}[The keywords 'include' 'int' and 'return' are rendered in bold. The normal serif font is used.] \lstset{language=C} \lstinputlisting{listing-samples/helloworld.c} \end{result} I~can also have inline code snippets using: \begin{definition} \gls{lstinline}\oarg{\meta{options}}\meta{char}\meta{code}\meta{char} \end{definition} This is different syntax to the usual forms of command argument. You can chose any character \meta{char} that isn't the open square bracket \gls{opt.opensq} and that doesn't occur in \meta{code} to delimit the code, but the start and end \meta{char} must match. (The optional argument is discussed below.) So the following are all equivalent: \begin{enumerate} \item \meta{char} is the exclamation mark character: \begin{codeS} \glsni{lstinline}!\#include ! \end{codeS} \item \meta{char} is the vertical bar character: \begin{codeS} \glsni{lstinline}|\#include | \end{codeS} \item \meta{char} is the double-quote character: \begin{codeS} \glsni{lstinline}"\#include " \end{codeS} \item \meta{char} is the plus symbol: \begin{codeS} \glsni{lstinline}+\#include + \end{codeS} \end{enumerate} And so on, but \meta{char} can't be, say, \verb|#| as that occurs in \meta{code}. Example: \begin{code} The stdio header file (required for the \glsni{lstinline}+printf+ function) is loaded using the directive \glsni{lstinline}!\#include ! on the first line. \end{code} Result: \begin{result}[The keyword 'include' is rendered in bold.] \lstset{language=C}% The stdio header file (required for the \lstinline+printf+ function) is loaded using the directive \lstinline!#include ! on the first line. \end{result} Another alternative is to input the code from an external file. For example, suppose my C~code is contained in the file \texttt{helloworld.c}, then I~can input it using: \begin{codeS} \gls{lstinputlisting}\oarg{\meta{options}}\marg{helloworld.c} \end{codeS} (Remember to use a~forward slash \texttt{/} as the directory divider, even if you are using Windows.) All the above (\glsni{lstinline}, \glsni{lstinputlisting} and the \glsni{env-lstlisting} environment) have an optional argument \meta{options} that can be used to override the default settings. These are \meta{key}=\meta{value} options. There are a~lot of options available, but I'm only going to cover a~few. If you want more detail, have a~look at the \sty{listings} documentation~\cite{listings}. \begin{description} \item[\ikeyvalopt{lstinputlisting}{title}=\marg{\meta{text}}] is used to set an unnumbered and unlabelled title. If \meta{text} contains a comma or equal sign, make sure you enclose \meta{text} in curly braces \gls{leftbracechar} and~\gls{rightbracechar}. \item[\ikeyvalopt{lstinputlisting}{caption}=\marg{\oarg{\meta{short}}\meta{text}}] is used to set a~numbered caption. The optional part \meta{short} is an alternative short caption for the list of listings, which can be produced using \begin{definition} \gls{lstlistoflistings} \end{definition} As above, if the caption contains a~comma or equal sign, make sure you enclose it in curly braces \glsni{leftbracechar} and \glsni{rightbracechar}. \item[\ikeyvalopt{lstinputlisting}{label}=\marg{\meta{text}}] is used to assign a~label to this listing so the number can be referenced via \gls{ref}. \item[\ikeyvalopt{lstinputlisting}{numbers}=\marg{\meta{setting}}] The value \meta{setting} may be one of: \texttt{none} (no line numbers), \texttt{left} (line numbers on the left) or \texttt{right} (line numbers on the right). \item[\ikeyvalopt{lstinputlisting}{mathescape}] This is a~boolean key that can either be \texttt{true} (dollar \gls{dollarchar} character acts as the usual math mode shift) or \texttt{false} (deactivates the usual behaviour of \glsni{dollarchar}). \item[\ikeyvalopt{lstinputlisting}{basicstyle}=\marg{\meta{declaration}}] The value (one or more declarations) is used at the start of the listing to set the basic font style. For example, \meta{declaration} could be \gls{ttfamily} (which actually makes more sense for a~listing). \end{description}\reportpagebreak \xminisec{Note:} \warning If you set \ikeyvalopt{lstinputlisting}{basicstyle} to \glsni{ttfamily} and you want bold keywords, make sure you are using a typewriter font that supports bold, as not all of them do. (Recall from \novices[sec:changingfonts]{documentfonts} how to change the font family.) This book uses txtt (see \novices{renewcom}). Other possibilities include \sty{beramono}, \sty{tgcursor}, \sty{courier}, \sty{DejaVuSansMono} (or \sty{dejavu} to load the serif and sans-serif DejaVu fonts as well), \sty{lmodern} and \sty{luximono}. \xminisec{KOMA and \sty{listings}} \warning If you want to use the \sty{listings} package with one of the \koma\ classes, you need to load \sty{scrhack} \emph{before} \sty{listings}, otherwise you will get a~warning that looks like: \begin{verbatim} Class scrbook Warning: Usage of deprecated \float@listhead! (scrbook) You should use the features of package `tocbasic' (scrbook) instead of \float@listhead. (scrbook) Definition of \float@listhead my be removed from (scrbook) `scrbook' soon, so it should not be used on input line 57. \end{verbatim} \xminisec{Example:} \begin{codelisting}{thesis-listings.tex} \null\par \glsni{begin}\marg{lstlisting}\oarg{language=C\comma basicstyle=\glsni{ttfamily}\comma mathescape=true}\newline \#include /* needed for printf */\newline \#include /* needed for sqrt */\newline \mbox{}\newline int main()\newline \marg*{\newline \mbox{}~~~double x = sqrt(2.0); /* \glsni{dollarchar}x = \gls{sqrt}\marg{2}\glsni{dollarchar} */\newline \mbox{}\newline \mbox{}~~~printf("x = \%f\cmdname{n}", x);\newline \mbox{}\newline \mbox{}~~~return 1;\newline }\newline \glsni{end}\marg{lstlisting} \end{codelisting} Result: \begin{result}[The formula in the comment is typeset in math mode and spaces in the string are displayed with the visible space symbol] \lstinputlisting[language=C,basicstyle=\ttfamily,mathescape=true]{listing-samples/sqrt.c} \end{result} If you are using \htmlref{double-spacing}{sec:setspace}, you may need to temporarily switch it off in the listings. You can do this by adding \gls{singlespacing} to the \ikeyvalopt{lstset}{basicstyle} setting. \begin{codeS} \glsni{lstset}\marg{basicstyle=\marg{\glsni{ttfamily}\glsni{singlespacing}}} \end{codeS} (Check with your supervisor to find out if listings should be double- or single-spaced.) \xminisec{Note:} It is not usually appropriate to have reams of listings in your thesis. It can annoy an examiner if you have included every single piece of code you have written during your PhD, as it comes across as padding to make it look as though your thesis is a~lot larger than it really is. (Examiners are not easily fooled, and it's best not to irritate them as it is likely to make them less sympathetic towards you.) If you want to include listings in your thesis, check with your supervisor first to find out whether or not it is appropriate. \warning Be careful when you use verbatim-like environments or commands, such as \gls{env-verbatim}, \gls{env-lstlisting}, \gls{lstinline} and \gls{lstinputlisting}. In general, they can't be used in the argument of another command.\faq{Why doesn't verbatim work within \ldots?}{verbwithin} \setnode{tabbing} \section{Tabbing} \label{sec:tabbing} The \gls{env-tabbing} environment lets you create tab stops so that you can tab to a~particular distance from the left margin. Within the tabbing environment, you can use the command \gls{tabstop} to set a~tab stop, \gls{greaterthan} to jump to the next tab stop, \gls{lessthan} to go back a~tab stop, \gls{plus} to shift the left border by one tab stop to the right, \gls{hyphen-tab} to shift the left border by one tab stop to the left. In addition, \gls{dbbackslashchar} will start a~new line and \gls{kill} will set any tabs stops defined in that line, but will not typeset the line itself. \xminisec{Note:} \warning You may recall two of the above commands from Volume~1: \gls{hyphen-discretionary} was described as a~discretionary hyphen in \xrsectionref{Hyphenation}{hyphenation} and \gls{macron} was described as the macron accent command in \xrsectionref[sec:chars]{Special Characters and Symbols}{symbols}. These two commands take on different meanings when they are used in the \glsni{env-tabbing} environment.\faq{Accents misbehave in \envname{tabbing}}{tabacc} If you want accents in your \glsni{env-tabbing} environment, either use the \isty{inputenc} package (see \novices{inputenc}) or use \gls{a}\meta{accent symbol}\marg{\meta{c}}, for example \verb|\a"{u}| instead of \verb|\"{u}|.\reportpagebreak\screenpagebreak \xminisec{Example:} \begin{code} \begin{alltt} \glsni{begin}\marg{tabbing} Zero \glsni{tabstop}One \glsni{tabstop}Two \glsni{tabstop}Three\glsni{dbbackslashchar} \glsni{greaterthan}First tab stop\glsni{dbbackslashchar} \glsni{greaterthan}A\glsni{greaterthan}\glsni{greaterthan}B\glsni{dbbackslashchar} \glsni{greaterthan}\glsni{greaterthan}Second tab stop \glsni{end}\marg{tabbing} \end{alltt} \end{code} This produces the following output: \begin{result} \begin{tabbing} Zero \=One \=Two \=Three\\ \>First tab stop\\ \>A\>\>B\\ \>\>Second tab stop \end{tabbing} \end{result} \xminisec{Another Example:} This example sets up four tab stops, but ignores the first line: \begin{code} \begin{alltt} \glsni{begin}\marg{tabbing} AAA \glsni{tabstop}BBBB \glsni{tabstop}XX \glsni{tabstop}YYYYYY \glsni{tabstop}Z \glsni{kill} \glsni{greaterthan}\glsni{greaterthan}\glsni{greaterthan}Third tab stop\glsni{dbbackslashchar} \glsni{greaterthan}a \glsni{greaterthan}b \glsni{greaterthan} \glsni{greaterthan}c \glsni{end}\marg{tabbing} \end{alltt} \end{code} This produces the following output: \begin{result}[The first line isn't shown] \begin{tabbing} AAA \=BBBB \=XX \=YYYYYY \=Z \kill \>\>\>Third tab stop\\ \>a \>b \> \>c \end{tabbing} \end{result} \setnode{newtheorem} \section{Theorems} \label{sec:newtheorem} A PhD thesis can often contain theorems, lemmas, definitions etc. The \LaTeX\ kernel comes with the command: \begin{definition} \gls{newtheorem}\marg{\meta{name}}\oarg{\meta{counter}}\marg{\meta{title}}\oarg{\meta{outer counter}} \end{definition} which can be used to create an environment called \meta{name} that has an optional argument. Each instance of the environment starts with \meta{title} followed by the associated counter value. If \meta{counter} is present, the new environment uses that counter instead of having a~new counter defined for it. If \meta{outer counter} is present, the environment counter is reset every time \meta{outer counter} is incremented. The optional arguments are mutually exclusive. In the example below, I've use \glsni{newtheorem} to define a~new environment called \envname{theorem}, which has an associated counter, also called \counter{theorem}, that is dependant on the \counter{chapter} counter.\bookpagebreak \begin{code} \begin{alltt} \glsni{percentchar} in the preamble: \glsni{newtheorem}\marg{theorem}\marg{Theorem}\oarg{chapter} \glsni{percentchar} later in the document: \glsni{begin}\marg{theorem} If proposition \glsni{dollarchar}P\glsni{dollarchar} is a tautology then \glsni{dollarchar}\gls{sim} P\glsni{dollarchar} is a contradiction, and conversely. \glsni{end}\marg{theorem} \end{alltt} \end{code} Resulting output: \begin{result}[theorem.html] \begin{theorem} If proposition $P$ is a tautology then $\sim P$ is a contradiction, and conversely. \end{theorem} \end{result} The optional argument to the new environment can be used to add a caption. Modifying the above example (changes shown \modification{like this}): \begin{code} \begin{alltt} \glsni{percentchar} in the preamble: \glsni{newtheorem}\marg{theorem}\marg{Theorem}\oarg{chapter} \glsni{percentchar} later in the document: \glsni{begin}\marg{theorem}\modification{\oarg{Tautologies and Contradictions}} If proposition \glsni{dollarchar}P\glsni{dollarchar} is a tautology then \glsni{dollarchar}\glsni{sim} P\glsni{dollarchar} is a contradiction, and conversely. \glsni{end}\marg{theorem} \end{alltt} \end{code}\reportpagebreak\noindent Resulting output: \begin{result}[As the previous example except the code in the optional argument is added in paretheses to the running header] \begin{theorem}[Tautologies and Contradictions] If proposition $P$ is a tautology then $\sim P$ is a contradiction, and conversely. \end{theorem} \end{result} Here's an example that uses the first optional argument of \glsni{newtheorem}: \begin{code} \begin{alltt} \glsni{percentchar} in the preamble: \glsni{newtheorem}\marg{exercise}\marg{Exercise} \glsni{newtheorem}\marg{suppexercise}\oarg{exercise}\marg{Supplementary Exercise} \glsni{percentchar} later in the document: \glsni{begin}\marg{exercise} This is an example of how to create a theorem-like environment. \glsni{end}\marg{exercise} \glsni{begin}\marg{suppexercise} This is another example of how to create a theorem-like environment. \glsni{end}\marg{suppexercise} \end{alltt} \end{code} Result: \begin{result}[extheorem.html] \begin{Exercise} This is an example of how to create a theorem-like environment. \end{Exercise} \begin{Suppexercise} This is another example of how to create a theorem-like environment. \end{Suppexercise} \end{result} Unfortunately there isn't a~great deal of flexibility with the environment appearance. \faq{Theorem bodies printed in a~roman font}{theoremfmt}However there are various packages available that provide enhancements to this basic command, allowing you to adjust the appearance to suit your requirements. There seem to be two main contenders: \isty{amsthm} and \isty{ntheorem}. Both have advantages and disadvantages. For example, \sty{ntheorem} is more flexible but \isty{amsthm} is more robust. Therefore I'm going to describe both, and you will have to decide which one you prefer. \xminisec{Note:} \warning If you are using either packages with \isty{amsmath}, you must load \sty{amsmath} first: \begin{code} \begin{alltt} \gls{usepackage}\marg{amsmath} \gls{usepackage}\marg{ntheorem} \end{alltt} \end{code} or \begin{code} \begin{alltt} \gls{usepackage}\marg{amsmath} \gls{usepackage}\marg{amsthm} \end{alltt} \end{code} With both \sty{amsthm} and \sty{ntheorem}, you can still define new theorem-like environments using \gls{newtheorem}, but there is also a~starred version of that command, which can be used to define unnumbered theorem-like environments. \xminisec{Example:} Suppose I~want to have an unnumbered \envname{remark} environment, I~can define the environment like this: \begin{code} \begin{alltt} \glsni{percentchar} in the preamble: \glsni{newtheorem}*\marg{note}\marg{Note} \glsni{percentchar} later in the document: \glsni{begin}\marg{note} This is a note about something. \glsni{end}\marg{note} \end{alltt} \end{code} Result: \begin{result}[Similar to before but the header text is just 'Note' without a number] \begin{plainnote} This is a note about something. \end{plainnote} \end{result} \setnode{amsthm} \subsection{The \sty{amsthm} Package} \label{sec:amsthm} The \isty{amsthm} package provides three predefined theorem styles: \texttt{plain}, \texttt{definition} and \texttt{remark}. When you define a new theorem-like environment with \gls{newtheorem}, it is given the style \emph{currently in effect}. You can change the current style with: \begin{definition} \gls{theoremstyle}\marg{\meta{style name}} \end{definition} where \meta{style name} is the name of the theorem style. \xminisec{Example:} This example defines six theorem-like environments: \envname{theorem}, \envname{lemma}, \envname{defn}, \envname{conj}, \envname{note} and \envname{remark}. The \envname{note} environment is unnumbered as it's defined using the starred version of \glsni{newtheorem}. The definitions have been arranged according to the required theorem style. \begin{code} \begin{alltt} \glsni{theoremstyle}\marg{plain} \glsni{newtheorem}\marg{theorem}\marg{Theorem} \glsni{newtheorem}\marg{lemma}\marg{Lemma} \glsni{theoremstyle}\marg{definition} \glsni{newtheorem}\marg{defn}\marg{Definition} \glsni{newtheorem}\marg{conj}\marg{Conjecture} \glsni{theoremstyle}\marg{remark} \glsni{newtheorem}*\marg{note}\marg{Note} \glsni{newtheorem}\marg{remark}\marg{Remark} \end{alltt} \end{code} \label{amsthm-proof}% The \sty{amsthm} package also provides the \gls{env-proof} environment, which can be used for typesetting proofs. \begin{definition} \glsni{begin}\marg{proof}\oarg{\meta{title}} \end{definition} The optional argument \meta{title} is a replacement for the default title. This environment automatically inserts a~QED symbol at the end of it, but if the default location isn't appropriate (which can happen if the proof ends with an equation) then use \begin{definition} \gls{qedhere} \end{definition} where you want the QED symbol to appear. The symbol is given by \begin{definition} \gls{qedsymbol} \end{definition} This defaults to an unfilled square $\openbox$, but you can redefine \glsni{qedsymbol} to something else if you prefer. (Recall redefining commands from \novices{renewcom}.) \begin{codelisting}{thesis-amsthm.tex} \null\par \glsni{percentchar} in the preamble:\newline \mbox{}\newline \glsni{usepackage}\marg{amsthm}\newline \glsni{theoremstyle}\marg{plain}\newline \glsni{newtheorem}\marg{theorem}\marg{Theorem}\newline \mbox{}\newline \glsni{theoremstyle}\marg{definition}\newline \glsni{newtheorem}\marg{defn}\marg{Definition}\newline \glsni{newtheorem}\marg{xmpl}\marg{Example}\oarg{chapter}\newline \mbox{}\newline \glsni{theoremstyle}\marg{remark}\newline \glsni{newtheorem}\marg{remark}\marg{Remark}\newline \mbox{}\newline \glsni{percentchar} later in the document:\newline \mbox{}\newline \glsni{begin}\marg{defn}\oarg{Tautology}\glsni{label}\marg{def:tautology}\newline A \gls{emph}\marg{tautology} is a proposition that is always true for any value of its variables.\newline \glsni{end}\marg{defn}\newline \mbox{}\newline \glsni{begin}\marg{defn}\oarg{Contradiction}\glsni{label}\marg{def:contradiction}\newline A \glsni{emph}\marg{contradiction} is a proposition that is always false for any value of its variables.\newline \glsni{end}\marg{defn}\newline \mbox{}\newline \glsni{begin}\marg{theorem}\newline If proposition \glsni{dollarchar}P\glsni{dollarchar} is a tautology then \glsni{dollarchar}\gls{sim} P\glsni{dollarchar} is a contradiction, and conversely.\newline \glsni{begin}\marg{proof}\newline If \glsni{dollarchar}P\glsni{dollarchar} is a tautology, then all elements of its truth table are true (by Definition\glsni{tildechar}\glsni{ref}\marg{def:tautology}), so all elements of the truth table for \glsni{dollarchar}\glsni{sim} P\glsni{dollarchar} are false, therefore \glsni{dollarchar}\glsni{sim} P\glsni{dollarchar} is a contradiction (by Definition\glsni{tildechar}\glsni{ref}\marg{def:contradiction}).\newline \glsni{end}\marg{proof}\newline \glsni{end}\marg{theorem}\newline \mbox{}\newline \glsni{begin}\marg{xmpl}\glsni{label}\marg{ex:rain}\newline \gls{quotedblleft}It is raining or it is not raining\gls{quotedblright} is a tautology, but \glsni{quotedblleft}it is not raining and it is raining\glsni{quotedblright} is a contradiction.\newline \glsni{end}\marg{xmpl}\newline \mbox{}\newline \glsni{begin}\marg{remark}\newline Example\gls{tildechar}\glsni{ref}\marg{ex:rain} used De Morgan\gls{quoteright}s Law \glsni{dollarchar}\glsni{sim} (p \gls{vee} q) \gls{equiv} \glsni{sim} p \gls{wedge} \glsni{sim} q\glsni{dollarchar}.\newline \glsni{end}\marg{remark} \end{codelisting} Result: \begin{result}[amsthm.html] \noindent\textbf{Definition \refstepcounter{defn}\thedefn\label{def:tautology}} (Tautology). A \emph{tautology} is a proposition that is always true for any of its variables. \par\vskip\baselineskip\noindent \textbf{Definition \refstepcounter{defn}\thedefn\label{def:contradiction}} (Contradiction). A \emph{contradiction} is a proposition that is always false for any value of its variables. \par\vskip\baselineskip\noindent \textbf{Theorem \refstepcounter{theorem}\thetheorem.} \emph{If proposition $P$ is a tautology then $\sim P$ is a contradiction, and conversely.} \par\vskip\baselineskip \noindent\emph{Proof.} If $P$ is a tautology, then all elements of its true table are true (by Definition~\ref*{def:tautology}), so all elements of the truth table for $\sim P$ are false, therefore $\sim P$ is a contradiction (by Definition~\ref*{def:contradiction}).\hfill$\openbox$\par \vskip\baselineskip\noindent \textbf{Example~\refstepcounter{Example}\theExample\label{ex:rain}.} ``It is raining or it is not raining'' is a tautology, but ``it is not raining and it is raining'' is a contradiction. \par\vskip\baselineskip\noindent \emph{Remark}~\refstepcounter{Remark}\theRemark. Example~\ref*{ex:rain} used De~Morgan's Law $\sim (p \vee q) \equiv \sim p \wedge \sim q$. \end{result} A new theorem style can be created using \begin{definition} \gls{newtheoremstyle}\marg{\meta{name}}\marg{\meta{space above}}\marg{\meta{space below}}\marg{\meta{body font}}\marg{\meta{indent}}\marg{\meta{head font}}\marg{\meta{post head punctuation}}\marg{\meta{post head space}}\marg{\meta{head spec}} \end{definition} This defines a~new theorem style called \meta{name}, which can later be set using \booklinebreak\gls{theoremstyle}. The other arguments are as follows: \begin{fwlist}{\meta{post head punctuation}} \fwitem{\meta{space above}} the amount of space above the theorem-like environment \fwitem{\meta{space below}} the amount of space below the theorem-like environment \fwitem{\meta{body font}} the font to be used in the main theorem body \fwitem{\meta{indent}} the amount of indentation (empty means no indent or use \nxgls{parindent} for normal paragraph indentation) \fwitem{\meta{head font}} the font to be used in the theorem header \fwitem{\meta{post head punctuation}} the punctuation to be inserted after the theorem head \fwitem{\meta{post head space}} the space to put after the theorem head (use \texttt{\marg{\glsni{visiblespace}}} for a~normal interword space or \gls{newline} for a~linebreak) \fwitem{\meta{head spec}} the theorem head spec \end{fwlist} \xminisec{Example:} This example creates a~new style called \texttt{note} that inserts a space of 2ex above the theorem and 2ex below.\footnote{Recall the \Indextt{ex} unit from \novices{length}.} The body font is just the normal font. There is no indent, the theorem header is in small caps, a~full stop is put after the theorem head and a~line break is inserted between the theorem head and body: \begin{code} \begin{alltt} \glsni{newtheoremstyle}\marg{note}\glsni{percentchar} style name \marg{2ex}\glsni{percentchar} above space \marg{2ex}\glsni{percentchar} below space \marg{}\glsni{percentchar} body font \marg{}\glsni{percentchar} indent amount \marg{\gls{scshape}}\glsni{percentchar} head font \marg{.}\glsni{percentchar} post head punctuation \marg{\glsni{newline}}\glsni{percentchar} post head punctuation \marg{}\glsni{percentchar} head spec \end{alltt} \end{code} Once you have defined the style, you can now use it. For example (in the preamble): \begin{code} \begin{alltt} \gls{theoremstyle}\marg{note} \gls{newtheorem}\marg{scnote}\marg{Note} \end{alltt} \end{code} This defines a~theorem-like environment called \envname{scnote}. You can now use it later in the document: \begin{code} \begin{alltt} \glsni{begin}\marg{scnote} This is an example of a theorem-like environment. \glsni{end}\marg{scnote} \end{alltt} \end{code} This produces: \begin{result}[The header is in small-caps followed by a number and a full stop. The body of the environment starts on the next line and is in the normal font.] \vspace{2ex} \noindent \textsc{Note 1.} \vspace{2ex} \noindent This is an example of a theorem-like environment. \end{result} \setnode{ntheorem} \subsection{The \sty{ntheorem} Package} \label{sec:ntheorem} The \isty{ntheorem} package provides nine predefined theorem styles, listed in \tableref{tab:ntheoremstyles}. The default is \texttt{plain}. When you define a new theorem-like environment with \booklinebreak\gls{newtheorem}, it is given the style \emph{currently in effect}. You can change the current style with: \begin{definition} \gls{theoremstyle}\marg{\meta{style name}} \end{definition} where \meta{style name} is the name of the theorem style. \begin{table}[htbp] \caption[Theorem Styles]{Predefined Theorem Styles Provided by \sty{ntheorem}} \label{tab:ntheoremstyles} \centering \begin{tabular}{lp{0.7\linewidth}} \texttt{plain} & Like the original \LaTeX\ style\\ \texttt{break} & Header is followed by a~line break\\ \texttt{change} & Like \texttt{plain} but header and number interchanged\\ \texttt{changebreak} & Combination of \texttt{change} and \texttt{break}\\ \texttt{margin} & Number is set in the margin\\ \texttt{marginbreak} & Like \texttt{margin} but header followed by a~line break\\ \texttt{nonumberplain} & Like \texttt{plain} but without the number\\ \texttt{nonumberbreak} & Like \texttt{break} but without the number\\ \texttt{empty} & No number and no name. Only the optional argument is used in the header. \end{tabular} \end{table} In addition to these styles, you can also use \begin{definition} \gls{theoremheaderfont}\marg{\meta{declarations}} \end{definition} to set the header font to \meta{declarations}, which should consist of font declaration commands such as \gls{normalfont}, \begin{definition} \gls{theorembodyfont}\marg{\meta{declarations}} \end{definition} to set the body font to \meta{declarations}, and \begin{definition} \gls{theoremnumbering}\marg{\meta{style}} \end{definition} to set the appearance of the theorem number, where \meta{style} may be one of: \texttt{arabic}, \texttt{roman}, \texttt{Roman}, \texttt{alph}, \texttt{Alph}, \texttt{greek}, \texttt{Greek} or \texttt{fnsymbol}. Remember that the above commands all need to be used before the new theorem-like environment is defined. For additional commands that affect the style of the theorems, see the \isty{ntheorem} documentation~\cite{ntheorem}. \xminisec{Example:} \begin{code} \begin{alltt} \glsni{percentchar} in the preamble: \glsni{theoremstyle}\marg{marginbreak} \glsni{theorembodyfont}\marg{\glsni{normalfont}} \glsni{newtheorem}\marg{note}\marg{Note}\oarg{chapter} \glsni{percentchar} later in the document: \glsni{begin}\marg{note} This is a sample note. The number is in the margin. \glsni{end}\marg{note} \end{alltt} \end{code} Result\html{ (the vertical line in the image below indicates the boundary of the text area and won't appear in the PDF)}: \begin{result} \begin{note} This is a sample note. The number is in the margin. \end{note} \end{result} If you use the \istyopt{ntheorem}{standard} package option to \sty{ntheorem}, it will automatically define the following environments: \envname{Theorem}, \envname{Lemma}, \envname{Proposition}, \envname{Corollary}, \envname{Satz}, \envname{Korollar}, \envname{Definition}, \envname{Example}, \envname{Beispiel}, \envname{Anmerkung}, \envname{Bemerkung}, \envname{Remark}, \gls*{env-Proof} and \envname{Beweis}. \warning Unlike \isty{amsthm}'s \gls{env-proof} environment, \isty{ntheorem}'s \gls{env-Proof} environment appends its optional argument in parentheses, if present, to the proof title. (Recall from \htmlref{earlier}{amsthm-proof}\doifbook{ \vpageref{amsthm-proof}} that \isty{amsthm}'s \gls{env-proof} environment uses its optional argument as a replacement for the default proof title.) \xminisec{Example:} \begin{codelisting}{thesis-ntheorem.tex} \null\par \glsni{percentchar} in the preamble:\newline \mbox{}\newline \glsni{usepackage}\oarg{standard}\marg{ntheorem}\newline \mbox{}\newline \glsni{percentchar} later in the document:\newline \mbox{}\newline \glsni{begin}\marg{Definition}\oarg{Tautology}\glsni{label}\marg{def:tautology}\newline A \gls{emph}\marg{tautology} is a~proposition that is always true for any value of its variables.\newline \glsni{end}\marg{Definition}\newline \mbox{}\newline \glsni{begin}\marg{Definition}\oarg{Contradiction}\glsni{label}\marg{def:contradiction}\newline A \glsni{emph}\marg{contradiction} is a~proposition that is always false for any value of its variables.\newline \glsni{end}\marg{Definition}\newline \mbox{}\newline \glsni{begin}\marg{Theorem}\newline If proposition \glsni{dollarchar}P\glsni{dollarchar} is a~tautology then \glsni{dollarchar}\gls{sim} P\glsni{dollarchar} is a~contradiction, and conversely.\newline \glsni{begin}\marg{Proof}\newline If \glsni{dollarchar}P\glsni{dollarchar} is a~tautology, then all elements of its truth table are true (by Definition\glsni{tildechar}\glsni{ref}\marg{def:tautology}), so all elements of the truth table for \glsni{dollarchar}\glsni{sim} P\glsni{dollarchar} are false, therefore \glsni{dollarchar}\glsni{sim} P\glsni{dollarchar} is a contradiction (by Definition\glsni{tildechar}\glsni{ref}\marg{def:contradiction}).\newline \glsni{end}\marg{Proof}\newline \glsni{end}\marg{Theorem}\newline \mbox{}\newline \glsni{begin}\marg{Example}\glsni{label}\marg{ex:rain}\newline \glsni{quotedblleft}It is raining or it is not raining\glsni{quotedblright} is a~tautology, but \glsni{quotedblleft}it is not raining and it is raining\glsni{quotedblright} is a~contradiction.\newline \glsni{end}\marg{Example}\newline \mbox{}\newline \glsni{begin}\marg{Remark}\newline Example\gls{tildechar}\glsni{ref}\marg{ex:rain} used De Morgan\glsni{quoteright}s Law \glsni{dollarchar}\glsni{sim} (p \gls{vee} q) \gls{equiv} \glsni{sim} p \gls{wedge} \glsni{sim} q\glsni{dollarchar}.\newline \glsni{end}\marg{Remark}\newline \end{codelisting} Result: \begin{result}[ntheorem.html] \begin{Definition}[Tautology]\label{nthm:def:tautology} A \emph{tautology} is a~proposition that is always true for any value of its variables. \end{Definition} \begin{Definition}[Contradiction]\label{nthm:def:contradiction} A \emph{contradiction} is a~proposition that is always false for any value of its variables. \end{Definition} \begin{Theorem} If proposition $P$ is a~tautology then $\sim P$ is a~contradiction, and conversely. \begin{Proof} If $P$ is a tautology, then all elements of its truth table are true (by Definition~\ref{nthm:def:tautology}), so all elements of the truth table for $\sim P$ are false, therefore $\sim P$ is a contradiction (by Definition~\ref{nthm:def:contradiction}). \end{Proof} \end{Theorem} \begin{Example}\label{ex:rain2} ``It is raining or it is not raining'' is a~tautology, but ``it is not raining and it is raining'' is a~contradiction. \end{Example} \begin{Remark} Example~\ref*{ex:rain2} used De Morgan's Law $\sim (p \vee q) \equiv \sim p \wedge \sim q$. \end{Remark} \end{result} \setnode{algorithms} \section{Algorithms} \label{sec:algorithms} If you want to display an algorithm, such as pseudo-code, you can use a~combination of the \gls{env-tabbing} environment (described in \sectionref{sec:tabbing}) and a~theorem-like environment (described above in \sectionref{sec:newtheorem}). \xminisec{Example:} \begin{code} \begin{alltt} \glsni{percentchar} in the preamble: \gls{theoremstyle}\marg{break} \gls{theorembodyfont}\marg{\glsni{normalfont}} \gls{newtheorem}\marg{algorithm}\marg{Algorithm} \glsni{percentchar} later in the document: \glsni{begin}\marg{algorithm}\oarg{Gauss-Seidel Algorithm} \glsni{begin}\marg{tabbing} 1. \gls{tabstop}For \glsni{dollarchar}k=1\glsni{dollarchar} to maximum number of iterations\gls{dbbackslashchar} \gls{greaterthan}2. For \glsni{tabstop}\glsni{dollarchar}i=1\glsni{dollarchar} to \glsni{dollarchar}n\glsni{dollarchar}\glsni{dbbackslashchar} \glsni{greaterthan}\glsni{greaterthan}Set \glsni{begin}\marg{math} x\gls{underscorechar}i\gls{circumchar}\marg{(k)} = \gls{frac}\marg{b\glsni{underscorechar}i-\gls{sum}\glsni{underscorechar}\marg{j=1}\glsni{circumchar}\marg{i-1}a\glsni{underscorechar}\marg{ij}x\glsni{underscorechar}j\glsni{circumchar}\marg{(k)} -\glsni{sum}\glsni{underscorechar}\marg{j=i+1}\glsni{circumchar}\marg{n}a\glsni{underscorechar}\marg{ij}x\glsni{underscorechar}j\glsni{circumchar}\marg{(k-1)}}\glsni{percentchar} \marg{a\glsni{underscorechar}\marg{ii}} \glsni{end}\marg{math} \glsni{dbbackslashchar} \glsni{greaterthan}3. If \glsni{dollarchar}\gls{lvert}\gls{vec}\marg{x}\glsni{circumchar}\marg{(k)}-\glsni{vec}\marg{x}\glsni{circumchar}\marg{(k-1)}\gls{rvert} < \gls{epsilon}\glsni{dollarchar}, where \glsni{dollarchar}\glsni{epsilon}\glsni{dollarchar} is a specified stopping criteria, stop. \glsni{end}\marg{tabbing} \glsni{end}\marg{algorithm} \end{alltt} \end{code} Result: \begin{result} \begin{algorithm}[Gauss-Seidel Algorithm] \begin{tabbing} 1. \=For $k=1$ to maximum number of iterations\\ \>2. For \=$i=1$ to $n$\\ \>\>Set \begin{math} x_i^{(k)} = \frac{b_i-\sum_{j=1}^{i-1}a_{ij}x_j^{(k)} -\sum_{j=i+1}^{n}a_{ij}x_j^{(k-1)}}% {a_{ii}} \end{math} \\ \>3. If $\lvert\vec{x}^{(k)}-\vec{x}^{(k-1)}\rvert < \epsilon$, where $\epsilon$ is a specified stopping criteria, stop. \end{tabbing} \end{algorithm} \end{result} (See \novices[sec:vec]{vectors} to find out how to redefine \glsni{vec} to display its argument in bold.) If you want a~more sophisticated approach, there are some packages available on \gls{ctan}, such as \isty{alg}, \isty{algorithm2e}, \isty{algorithms} and \isty{algorithmicx}. I'm going to briefly introduce the \sty{algorithm2e} package here. This provides the \gls{env-algorithm} floating environment. Like the \gls{env-figure} and \gls{env-table} environments described in \novices[ch:floats]{floats}, the \glsni{env-algorithm} environment has an optional argument that specifies the placement. \begin{definition} \glsni{begin}\marg{algorithm}\oarg{\meta{placement}} \end{definition} If you are using a~class or package that already defines an \envname{algorithm} environment, you can use the \istyopt{algorithm2e}{algo2e} package option: \begin{codeS} \glsni{usepackage}\oarg{algo2e}\marg{algorithm2e} \end{codeS} This will define an environment called \gls{env-algorithm2e} instead of \glsni{env-algorithm} to avoid conflict. Within the body of the environment, you must mark the end of each line with \gls{algo2e.semicolon} regardless of whether you want a semi-colon to appear. To suppress the default end-of-line semi-colon, use \begin{definition} \gls{DontPrintSemicolon} \end{definition} To switch it back on again, use \begin{definition} \gls{PrintSemicolon} \end{definition} There are a~variety of commands that may be used within the \glsni{env-algorithm} environment. Some of the commands are described below, but for a~complete list you should consult the \sty{algorithm2e} documentation~\cite{algorithm2e}.\bookpagebreak First there are the commands for the algorithm input, output and data: \begin{definition} \gls{KwIn}\marg{\meta{input}}\newline \gls{KwOut}\marg{\meta{output}}\newline \gls{KwData}\marg{\meta{input}}\newline \gls{KwResult}\marg{\meta{output}} \end{definition} Next there are commands for basic keywords: \begin{definition} \gls{KwTo}\newline \gls{KwRet}\marg{\meta{value}}\newline \gls{Return}\marg{\meta{value}} \end{definition} There are a~lot of conditionals, but here's a~selection: \begin{definition} \gls{If}\marg{\meta{condition}}\marg{\meta{then block}}\newline \gls{uIf}\marg{\meta{condition}}\marg{\meta{then block without end}}\newline \gls{ElseIf}\marg{\meta{else-if block}}\newline \gls{uElseIf}\marg{\meta{else-if block without end}}\newline \gls{Else}\marg{\meta{else block}} \end{definition} Similarly there are a~lot of loops, but here's a~selection: \begin{definition} \gls{For}\marg{\meta{condition}}\marg{\meta{body}}\newline \gls{While}\marg{\meta{condition}}\marg{\meta{body}} \end{definition} \xminisec{Example:} The above algorithm can be written using the \gls{env-algorithm2e} environment as follows (this document has used the \istyopt{algorithm2e}{algo2e} package option): \begin{codelisting}{thesis-algorithms.tex} \begin{alltt} \glsni{begin}\marg{algorithm2e} \gls{caption}\marg{Gauss-Seidel Algorithm}\gls{label}\marg{alg:gauss-seidel} \glsni{KwIn} \marg{\glsni{percentchar} scalar \glsni{dollarchar}\glsni{epsilon}\glsni{dollarchar}, matrix \glsni{dollarchar}\gls{mathbf}\marg{A} = (a\glsni{underscorechar}\marg{ij})\glsni{dollarchar}, vector \glsni{dollarchar}\gls{vec}\marg{b}\glsni{dollarchar} and initial vector \glsni{dollarchar}\glsni{vec}\marg{x}\gls{circumchar}\marg{(0)}\glsni{dollarchar} } \glsni{For}\marg{\glsni{dollarchar}k\gls{leftarrow} 1\glsni{dollarchar} \glsni{KwTo} maximum iterations} \marg{ \glsni{For}\marg{\glsni{dollarchar}i\glsni{leftarrow} 1\glsni{dollarchar} \glsni{KwTo} \glsni{dollarchar}n\glsni{dollarchar}} \marg{ \glsni{dollarchar} x\glsni{underscorechar}i\glsni{circumchar}\marg{(k)} = \glsni{frac} \marg{ b\glsni{underscorechar}i-\glsni{sum}\glsni{underscorechar}\marg{j=1}\glsni{circumchar}\marg{i-1}a\glsni{underscorechar}\marg{ij}x\glsni{underscorechar}j\glsni{circumchar}\marg{(k)} -\glsni{sum}\glsni{underscorechar}\marg{j=i+1}\glsni{circumchar}\marg{n}a\glsni{underscorechar}\marg{ij}x\glsni{underscorechar}j\glsni{circumchar}\marg{(k-1)} }\glsni{percentchar} \marg{a\glsni{underscorechar}\marg{ii}} \glsni{dollarchar}\glsni{algo2e.semicolon} } \glsni{If}\marg{\glsni{dollarchar}\glsni{lvert}\glsni{vec}\marg{x}\glsni{circumchar}\marg{(k)}-\glsni{vec}\marg{x}\glsni{circumchar}\marg{(k-1)}\glsni{rvert} < \glsni{epsilon}\glsni{dollarchar}} \marg{Stop} } \glsni{end}\marg{algorithm2e} \end{alltt} \end{codelisting} The result is shown in Algorithm~\ref{alg:gauss-seidel}. \begin{algorithm2e} \caption{Gauss-Seidel Algorithm}\label{alg:gauss-seidel} \KwIn {% Scalar $\epsilon$, matrix $\mathbf{A}=(a_{ij})$, vector $\vec{b}$ and initial vector $\vec{x}^{(0)}$ } \For{$k\leftarrow 1$ \KwTo maximum iterations} { \For{$i\leftarrow 1$ \KwTo $n$} { $ x_i^{(k)} = \frac{b_i-\sum_{j=1}^{i-1}a_{ij}x_j^{(k)} -\sum_{j=i+1}^{n}a_{ij}x_j^{(k-1)}}% {a_{ii}} $\; } \If{$\lvert\vec{x}^{(k)}-\vec{x}^{(k-1)}\rvert < \epsilon$}{Stop} } \end{algorithm2e} The \gls{env-algorithm} environment (as defined by \sty{algorithm2e} without the \istyopt{algorithm2e}{algo2e} option) or \gls{env-algorithm2e} environment (as defined with the \optfmt{algo2e} option) uses the \icounter{algocf} counter. So in this document, to ensure that the \envname{algorithm} environment defined with \gls{newtheorem} used the same counter as \gls{env-algorithm2e}, I~had the following in my preamble: \begin{code} \begin{alltt} \glsni{usepackage}\marg{ntheorem} \glsni{usepackage}\oarg{algo2e}\marg{algorithm2e} \glsni{theoremstyle}\marg{break} \glsni{theorembodyfont}\marg{\glsni{normalfont}} \glsni{newtheorem}\marg{algorithm}\oarg{algocf}\marg{Algorithm} \end{alltt} \end{code} \setnode{siunitx} \section{Formatting SI Units} \label{sec:siunitx} If you need to typeset numbers and units then I~strongly recommend that you use the \isty{siunitx} package. This section just provides a~brief introduction to that package. You will need to read the \sty{siunitx} package documentation~\cite{siunitx} if you want further details. \begin{definition} \gls{num}\marg{\meta{number}} \end{definition} This command typesets \meta{number}, adding appropriate spacing between number groups where necessary. It also adds a~leading zero if omitted before the decimal point and identifies exponents. Note that the command recognises both \dq{\texttt{.}} and \dq{\texttt{,}} as the decimal marker. If you want one of these characters between number groups (instead of the default space) you can change the settings, but it's best to stick to the default settings unless instructed to do otherwise. \xminisec{Example:} \begin{code} Out of \glsni{num}\marg{12890} experiments, \glsni{num}\marg{1289} of them had a mean squared error of \glsni{num}\marg{.346} and \glsni{num}\marg{128} of them had a mean squared error of \glsni{num}\marg{1.23e-6}. \end{code}\reportpagebreak\noindent Result: \begin{result}[sinum.html] Out of \num{12890} experiments, \num{1289} of them had a mean squared error of \num{.346} and \num{128} of them had a mean squared error of \num{1.23e-6}. \end{result} \begin{definition} \gls{ang}\marg{\meta{angle}} \end{definition} This command typesets an angle. The argument \meta{angle} may be a single number or three (some possibly empty) values separated by semi-colons. \xminisec{Example:} \begin{codeS} The result formed an arc from \glsni{ang}\marg{45} to \glsni{ang}\marg{60;2;3}. \end{codeS} Result: \begin{resultS}[siang.html] The result formed an arc from \ang{45} to \ang{60;2;3}. \end{resultS} \begin{definition} \gls{si}\marg{\meta{unit}} \end{definition} This command typesets a~unit. The \meta{unit} can be formed from commands like \gls{metre}, \gls{gram}, \gls{second} or \gls{kilo}. (See the \sty{siunitx} documentation~\cite{siunitx} for the full list.)\bookpagebreak \xminisec{Example:} \begin{code} The distance was measured in \glsni{si}\marg{\glsni{kilo}\glsni{metre}} and the area in \glsni{si}\marg{\glsni{kilo}\glsni{metre}\gls{squared}}. The acceleration was given in \glsni{si}\marg{\glsni{metre}\gls{per}\gls{square}\glsni{second}}. \end{code}\screenpagebreak Result: \begin{result}[siunit.html] The distance was measured in \si{\kilo\metre} and the area in \si{\kilo\metre\squared}. The acceleration was given in \si{\metre\per\square\second}. \end{result} \begin{definition} \gls{SI}\marg{\meta{number}}\marg{\meta{unit}} \end{definition} This combines the functionality of \glsni{num} and \glsni{si} so that you can typeset both a~number and a~unit. \xminisec{Example:} \begin{code} The acceleration was approximately \glsni{SI}\marg{9.78}\marg{\glsni{metre}\glsni{per}\glsni{square}\glsni{second}}. \end{code} Result: \begin{resultS}[sinumunit.html] The acceleration was approximately \SI{9.78}{\metre\per\square\second}. \end{resultS} % Generating a~Bibliography \setnode{citations} \chapter{Generating a~Bibliography} \label{ch:citations} \novices[sec:bib]{biblio} introduced the \gls{env-thebibliography} environment. While it is possible to write this environment yourself, as was done in Volume~1, it's not practical with a~large number of citations. Instead, the preferred method is to create an external database of bibliographic data and use an application that fetches the relevant information from that database and writes a~file containing the \glsni{env-thebibliography} environment, which can then be input into your document. This means that: \begin{enumerate} \item Only the references that you cite are included in the bibliography. (Examiners tend to fault uncited references\footnote{% When your examiners read through your thesis, they can check off each citation they encounter against your bibliography. When they reached the end of the thesis, they can then look through the bibliography for unchecked entries. One or two may appear the result of carelessness, whereas a~large quantity will look like padding and may lead the examiners to suspect a~certain amount of duplicity on your part.}.) \item References are displayed in a~consistent manner. \item Entries can be sorted in order of citation or alphabetically. \end{enumerate} Traditionally the \iappnamelink{bibtex}{http://www.bibtex.org/} application is used to generate the \glsni{env-thebibliography} environment. It comes with \TeX\ distributions and most books on \LaTeX\ cover \appname{bibtex}. Unfortunately \appname{bibtex} has some drawbacks, most notably the complexity of creating your own custom style. UTF-8 has also been a~problem, although newer versions of \appname{bibtex} apparently fix this. In 2006, Philipp Lehman brought out the \isty{biblatex} package to provide a~more flexible way of typesetting bibliographies. This originally used \appname{bibtex} to just sort the entries and used \LaTeX\ macros to deal with the actual formatting, but it is now moving over to using \iappnamelink{biber}{http://biblatex-biber.sourceforge.net/} instead of \appname{bibtex}. Since some journals, conferences or other types of scientific publishers require you to use \appname{bibtex}, \sectionref{sec:bibtex} provides a~brief introduction to \appname{bibtex} and then \sectionref{sec:biblatex} discusses \sty{biblatex} and \appname{biber}. But first \sectionref{sec:creatingbibfile} covers creating the actual database, which is required for both methods. \setnode{creatingbibfile} \section{Creating a~Bibliography Database} \label{sec:creatingbibfile} This section covers creating a~\texttt{.bib} file that contains the bibliographic information you want to cite in your documents. You can use an ordinary text editor to create a~bibliographic database (as described in \sectionref{sec:bibformat}) but it can be difficult to remember the names of the required fields and it's easy to make syntactic mistakes. It can also be hard to keep track of entries in a~large database. To make life easier, there are a~number of bibliography reference managers available that provide a convenient graphical interface. One such application is \appname{JabRef} and is described \htmlref{next}{sec:jabref}. \setnode{jabref} \subsection{JabRef} \label{sec:jabref} I've chosen to describe \iappname{JabRef} here because it's an open source Java application that can run on any operating system that has the \htmladdnormallinkfoot{Java Runtime Environment}{http://www.java.com/getjava/} installed (at least version 1.5). You can download \appname{JabRef} from \url{http://jabref.sourceforge.net/download.php}. Linux users may also be able to install it via their \dq{Add/Remove Software} tool. (If you have successfully been using \htmlref{\appname{arara}}{sec:arara}, you already have Java installed.) Once you have installed it, run \appname{JabRef} and select \menu{File}\menuto\menu{New database} to create a~new database (see \figureref{fig:jabref1}). When you save your data, it's saved as a BibTeX (\texttt{.bib}) file. Note that if you use the \isty{inputenc} package in your thesis (see \novices{inputenc}) you'll have to make sure \appname{JabRef} is using the same encoding as your document. You can do this by selecting \menu{Options}\menuto\menu{Preferences} to open the Preferences dialog box and set the default encoding as appropriate. For example, I~use UTF-8 so I've set that as the default encoding (see \figureref{fig:jabref-pref}). I~also need to change the database encoding in the \dq{Database properties} dialog, \figureref{fig:jabref-dataprop}, which can be opened using \menu{File}\menuto\menu{Database properties}. \begin{figure}[htbp] \figconts {pictures/jabref1} {\caption{JabRef}} {fig:jabref1} \end{figure} \begin{figure}[htbp] \figconts {pictures/jabref-pref} {\caption{JabRef Preferences}} {fig:jabref-pref} \end{figure} \begin{figure}[htbp] \figconts {pictures/jabref-dataprop} {\caption{JabRef Database Properties}} {fig:jabref-dataprop} \end{figure} To create a~new entry you can select \menu{BibTeX}\menuto\menu{New entry}, which will open the dialog box shown in \figureref{fig:jabref2}. Now you need to click on the button appropriate to the entry. For example, click on \dq{Article} for an article in a~journal or click on \dq{Inproceedings} for a~paper in a~conference proceedings. \xminisec{Example (Book):} Suppose I~want to enter information about a book. I~need to select \menu{BibTeX}\menuto\menu{New entry} and then click on the button labelled \dq{Book}. This now displays fields in which I~can enter the relevant information (see \figureref{fig:jabref3}). \begin{figure}[htbp] \figconts {pictures/jabref2} {\caption{JabRef (Select Entry Type)}} {fig:jabref2} \end{figure} \begin{figure}[htbp] \figconts {pictures/jabref3} {\caption{JabRef (New Entry)}} {fig:jabref3} \end{figure} Next I~need to enter information in the \dq{Required fields} tab. This will usually include the title and the author. I~also need to specify a key that uniquely identifies this entry. If you have read \novices[sec:bib]{biblio} this key corresponds to the mandatory argument of \gls{bibitem} and is also used in \glsi{cite}. \figureref{fig:jabref4} shows the details for my new entry. I've set the key to the author's surname followed by the year to make it easy to remember. This key won't appear anywhere in the document, it's just used to identify the entry, just like the \gls{label}/\gls{ref} mechanism. Alternatively, I~can click on the \dq{Generate BibTeX Key} button \incGraphics{pictures/generatekey} to automatically insert a~unique key. \begin{figure}[htbp] \figconts {pictures/jabref4} {\caption{JabRef (Entering the Required Fields)}} {fig:jabref4} \end{figure} There are also optional fields you can specify as well. In \figureref{fig:jabref5}, I've added the book's edition. \begin{figure}[htbp] \figconts {pictures/jabref5} {\caption{JabRef (Entering Optional Fields)}} {fig:jabref5} \end{figure} \xminisec{Example (Journal Article):} Now I~want to enter an article in a~journal. So I~need to go back to \menu{BibTeX}\menuto\menu{New entry} and click on \dq{Article}. This time I've used the \dq{Generate BibTeX Key} button to generate the key to save me typing. (See \figureref{fig:jabref6}.) I've also used the \dq{General} tab to enter the DOI for this article. The entry now has an icon \incGraphics{pictures/doibutton} next to it. I~can click on this button to direct my web browser to the article's entry on the Internet. \begin{figure}[htbp] \figconts {pictures/jabref6} {\caption{JabRef (Adding an Article)}} {fig:jabref6} \end{figure} BibTeX uses the European assumption\faq{\BiBTeX\ sorting and name prefixes}{bibprefixsort} that names are composed of forenames, an optional \dq{von} part which starts with a~lower case letter, a~surname and an optional \dq{jr} part. In order to enable BibTeX to correctly identify these components, names in the author or editor fields must be entered in one of the formats listed in \tableref{tab:bibnameformats}. \begin{table}[htbp] \caption{Name Formats for Bibliographic Data} \label{tab:bibnameformats} \centering \begin{tabular}{l} \meta{forenames} \meta{von} \meta{surname}\\ \meta{von} \meta{surname}, \meta{forenames}\\ \meta{von} \meta{surname}, \meta{jr}, \meta{forenames} \end{tabular} \end{table} \bookpagebreak \xminisec{Examples:} \begin{center} \begin{tabular}{ll} \toprule \bfseries Entry & \bfseries Abbreviated as\\ \midrule \ttfamily Alex Thomas von Neumann & A.T. von Neumann\\ \ttfamily John Chris \{Smith Jones\} & J.C. Smith Jones\\ \ttfamily van de Klee, Mary-Jane & M.-J. van de Klee\\ \ttfamily Smith, Jr, Fred John & F.J. Smith, Jr\\ \ttfamily \verb|Maria {\MakeUppercase{d}e La} Cruz| & M. De La Cruz\\ \bottomrule \end{tabular} \end{center} Compare the last example with: \texttt{Maria De La Cruz} which would be abbreviated to: M.~D.~L.~Cruz, which is incorrect. Let's analyse this last example in more detail: BibTeX always expects the \dq{von} part to start with a~lower case letter, but \dq{De} and \dq{La} both start with an upper case letter, so BibTeX will assume that these form part of the forenames. However, BibTeX will ignore any \LaTeX\ commands such as \faq{Case-changing oddities}{casechange}\gls{MakeUppercase} in \verb|\MakeUppercase{d}e| since it assumes that the command is an accent command\faq{Accents in bibliographies}{bibaccent}. So when it parses \verb|\MakeUppercase{d}e| it will skip \glsni{MakeUppercase} and look at the following letter. In this case it is \dq{d} which is lower case, so from BibTeX's point of view the word \verb|\MakeUppercase{d}e| starts with a~lower case letter (\dq{d}), so it is therefore the \dq{von} part. You can either do the same with the \dq{La} part, or, as in the above example, you can place it in the same group as \verb|\MakeUppercase{d}e|. Multiple authors should be separated by the keyword \dq{and}. \textbf{Don't use a~comma to separate the authors.} Here is an example with three authors: \begin{codeS} Gavin C. Cawley and Nicola L. C. Talbot and Mark Girolami \end{codeS} If the author is an institution or company that happens to have the word \dq{and} in its name, such as \dq{Smith and Jones Inc}, then you need to group the \dq{and} to indicate that you mean the word \dq{and} rather than the keyword: \begin{codeS} Smith \marg{and} Jones Inc \end{codeS} \figureref{fig:jabref7} shows the entry for a~paper in a~conference proceedings, so for that one I~used \menu{BibTeX}\menuto\menu{New entry} and clicked on the \dq{Inproceedings} button. \begin{figure}[htbp] \figconts {pictures/jabref7} {\caption{JabRef (Adding a~Conference Paper)}} {fig:jabref7} \end{figure} Notice the way I've written the title for this entry: \begin{code} Sparse multinomial logistic regression via \marg{Bayesian} \marg{L1} regularisation \end{code} BibTeX automatically converts the title to lower case (apart from the initial letter) but here both \dq{Bayesian} and \dq{L1} should begin with a~capital. I~therefore need to enclose those words in braces to instruct BibTeX not to convert their case. Multiple editors must also be separated by the \dq{and} keyword, as shown in \figureref{fig:jabref8}. For that entry, the editors are listed as: \begin{codeS} Bernhard Sch\"{o}lkopf and John Platt and Thomas Hofmann \end{codeS} Note that if I~don't use the \isty{inputenc} package, I~need to change this to: \begin{codeS} Bernhard Sch\gls{doublequote}\marg{o}lkopf and John Platt and Thomas Hofmann \end{codeS} \begin{figure}[htbp] \figconts {pictures/jabref8} {\caption{JabRef (Adding Editor List)}} {fig:jabref8} \end{figure} It's also possible to import entries from other formats, such as Copac or ISI, using \menu{File}\menuto\menu{Import into new database} or \menu{file}\menuto\menu{Import into current database}. Alternatively, you can copy and paste a~plain text reference using \menu{BibTeX}\menuto\menu{New entry from plain text}. This again opens the dialog box where you need to click on the entry type, but then it opens the \dq{Plain text import} window. \xminisec{Example:} Suppose I~want to add an entry for an article whose DOI is 10.1007/s10994-008-5055-9. First, I~direct my browser to \htmladdnormallink{\texttt{http://dx.doi.org/10.1007/\reportlinebreak\screenlinebreak s10994-008-5055-9}}{http://dx.doi.org/10.1007/s10994-008-5055-9}, which takes me to the article's web page. In this case, it's in a~journal published by Springer, so my browser is redirected to the SpringerLink cite. There I~can use the export as text only option, then copy and paste the reference into \appname{JabRef}'s import window, as shown in \figureref{fig:jabref-textimport1}. \begin{figure}[htbp] \figconts {pictures/jabref-textimport1} {\caption{Importing a~Plain Text Reference}} {fig:jabref-textimport1} \end{figure} Next, I~need to select text, for example an author's name, and select the appropriate field in the \dq{Work options} list. Then click on the \dq{Insert} button. For example, in \figureref{fig:jabref-textimport2} I~have selected an author's name then I~selected the \dq{author} field in the \dq{Work options} list. \begin{figure}[htbp] \figconts {pictures/jabref-textimport2} {\caption{Importing a~Plain Text Reference (Selecting a~Field)}} {fig:jabref-textimport2} \end{figure} Next I~clicked on the \dq{Insert} button. Now the author's name is highlighted in red and the author field has a~tick next to it (see \figureref{fig:jabref-textimport3}). I~can repeat this process for the next author. (Just make sure the \dq{Append} rather than \dq{Override} radio button is selected.) \begin{figure}[htbp] \figconts {pictures/jabref-textimport3} {\caption{Importing a~Plain Text Reference (Field Selected)}} {fig:jabref-textimport3} \end{figure} I~can repeat this for all the different fields. Each time, I~select the text in the raw source panel, then select the appropriate field from the \dq{Work options} list and then click \dq{Insert}. Once I~have finished, I~then need to click \dq{Accept}. \setnode{bibformat} \subsection{Writing the .bib File Manually} \label{sec:bibformat} It may be that you don't want to or can't use a~bibliography management application, such as \htmlref{\appname{JabRef}}{sec:jabref}. In which case, you can create the \texttt{.bib} file in an ordinary text editor, such as the one you use to write your \LaTeX\ documents. When you save the file, make sure you give it the extension \texttt{.bib}. Entries in this file should have the following form\faq{Creating a~BibTeX bibliography}{buildbib}: \vbox{\begin{ttfamily} \begin{tabbing} xxx\=\kill @\meta{entry type}\{\meta{keyword},\\ \>\meta{field name} \== "\meta{text}",\\ \>\> \vellipsis\\ \>\meta{field name} = "\meta{text}"\\ \} \end{tabbing} \end{ttfamily}} \noindent where \meta{entry type} indicates the type of entry (e.g.\ book or article). Standard entry types are listed in \tableref{tab:entrytype}. \begin{table}[htbp] \caption{Standard BiBTeX entry types} \label{tab:entrytype} \centering \begin{tabular}{ll} \toprule \bfseries Entry Name & \bfseries Description\\ \midrule \ibibentry{article} & Article from a~journal\\ \ibibentry{book} & Published book\\ \ibibentry{booklet} & Printed work without a~publisher\\ \ibibentry{conference} & Identical to \ibibentry{inproceedings}\\ \ibibentry{inbook} & Part, chapter, section etc of a~book\\ \ibibentry{incollection} & A chapter of a~book with its own author and title\\ \ibibentry{inproceedings} & An article in a~conference proceedings\\ \ibibentry{manual} & Technical documentation\\ \ibibentry{mastersthesis} & A master's thesis\\ \ibibentry{misc} & Non-standard work\\ \ibibentry{phdthesis} & PhD thesis\\ \ibibentry{proceedings} & Conference proceedings\\ \ibibentry{techreport} & Report published by an institution\\ \ibibentry{unpublished} & Unpublished work with an author and title\\ \bottomrule \end{tabular} \end{table} Within an entry, \meta{keyword} is a~short label that is used to cite this work with the \glsi{cite} command. If you have written bibliographies with the \gls{env-thebibliography} environment, it's the same as the argument to \gls{bibitem}. There then follows a~comma-separated list of fields of the form \meta{field name} \texttt{=} \meta{value}. The \meta{field name} indicates what kind of field it is, e.g.\ \ibibfield{title} or \ibibfield{author}. \tableref{tab:fields} lists the standard fields. Note that some bibliography styles may define additional non-standard fields, such as \texttt{email} or \texttt{url}.\faq{URLS in BibTeX bibliographies}{citeURL} See the Bib\TeX\ documentation~\cite{bibtex} for information about other fields not listed in \tableref{tab:fields}. \begin{table}[hbtp] \caption{Standard BiBTeX fields} \label{tab:fields} \centering\latex{\ifscreen\small\fi} \begin{tabular}{ll} \ibibfield{address} & Publisher/Institution's address\\ \ibibfield{author} & Author names\\ \ibibfield{booktitle} & Title of book where only a~part of the book is being cited\\ \ibibfield{chapter} & Chapter or section number\\ \ibibfield{edition} & The edition of the book\\ \ibibfield{howpublished} & How a~non-standard work was published\\ \ibibfield{institution} & The institute sponsoring the work\\ \ibibfield{journal} & The name of the journal\\ \ibibfield{month} & The month the work was published\\ \ibibfield{note} & Any additional information\\ \ibibfield{number} & The number of the journal, technical report etc\\ \ibibfield{organization} & Organization sponsoring conference or manual\\ \ibibfield{pages} & Page number or page range\\ \ibibfield{publisher} & Publisher's name\\ \ibibfield{school} & Academic institution where thesis was written\\ \ibibfield{series} & Name of a~series\\ \ibibfield{title} & The title of the work\\ \ibibfield{type} & The type of technical report\\ \ibibfield{volume} & The volume number. \end{tabular} \end{table} The required and optional fields for the standard entry types are listed in \tableref{tab:reqopt}. If an entry has a~field that is neither required nor optional, BibTeX will ignore it. This means that you can have a~field called, say, \ibibfield{abstract}, which will be ignored by the standard bibliography styles, but will be included if you use a~bibliography style that has an \ibibfield{abstract} field. So you can store additional information in the database that won't appear in the bibliography. \begin{fieldtab} \ibibentry{article} & \ibibfield{author}, \ibibfield{title}, \ibibfield{journal}, \ibibfield{year} & \ibibfield{volume}, \ibibfield{month}, \ibibfield{note}, \ibibfield{number}, \ibibfield{pages} \tabularnewline \ibibentry{book} & \ibibfield{author} or \ibibfield{editor}, \ibibfield{title}, \ibibfield{publisher}, \ibibfield{year} & \ibibfield{address}, \ibibfield{edition}, \ibibfield{volume} or \ibibfield{number}, \ibibfield{month}, \ibibfield{note}, \ibibfield{pages}, \ibibfield{series} \tabularnewline \ibibentry{booklet} & \ibibfield{title} & \ibibfield{author}, \ibibfield{address}, \ibibfield{howpublished}, \ibibfield{month}, \ibibfield{note}, \ibibfield{year} \tabularnewline \ibibentry{inbook} & \ibibfield{author} or \ibibfield{editor}, \ibibfield{chapter} or \ibibfield{pages}, \ibibfield{title}, \ibibfield{publisher}, \ibibfield{year} & \ibibfield{address}, \ibibfield{edition}, \ibibfield{volume} or \ibibfield{number}, \ibibfield{month}, \ibibfield{note}, \ibibfield{series}, \ibibfield{type} \tabularnewline \ibibentry{incollection} & \ibibfield{author}, \ibibfield{title}, \ibibfield{booktitle}, \ibibfield{publisher}, \ibibfield{year} & \ibibfield{address}, \ibibfield{chapter}, \ibibfield{editor}, \ibibfield{edition}, \ibibfield{volume} or \ibibfield{number}, \ibibfield{month}, \ibibfield{note}, \ibibfield{pages}, \ibibfield{series}, \ibibfield{type} \tabularnewline \ibibentry{inproceedings} & \ibibfield{author}, \ibibfield{title}, \ibibfield{booktitle}, \ibibfield{year} & \ibibfield{address}, \ibibfield{editor}, \ibibfield{volume} or \ibibfield{number}, \ibibfield{month}, \ibibfield{note}, \ibibfield{organization}, \ibibfield{pages}, \ibibfield{publisher}, \ibibfield{series}, \ibibfield{type} \tabularnewline \ibibentry{manual} & \ibibfield{title} & \ibibfield{author}, \ibibfield{address}, \ibibfield{edition}, \ibibfield{month}, \ibibfield{note}, \ibibfield{organization}, \ibibfield{year}\tabularnewline \ibibentry{mastersthesis} & \ibibfield{author}, \ibibfield{title}, \ibibfield{school}, \ibibfield{year} & \ibibfield{address}, \ibibfield{month}, \ibibfield{note}, \ibibfield{type}\tabularnewline \ibibentry{misc} & --- & \ibibfield{author}, \ibibfield{howpublished}, \ibibfield{month}, \ibibfield{note}, \ibibfield{title}, \ibibfield{year}\tabularnewline \ibibentry{phdthesis} & \ibibfield{author}, \ibibfield{title}, \ibibfield{school}, \ibibfield{year} & \ibibfield{address}, \ibibfield{month}, \ibibfield{note}, \ibibfield{type}\tabularnewline \ibibentry{proceedings} & \ibibfield{title}, \ibibfield{year} & \ibibfield{editor}, \ibibfield{organization}, \ibibfield{address}, \ibibfield{volume} or \ibibfield{number}, \ibibfield{series}, \ibibfield{month}, \ibibfield{publisher}, \ibibfield{note}\tabularnewline \ibibentry{techreport} & \ibibfield{author}, \ibibfield{title}, \ibibfield{institution}, \ibibfield{year} & \ibibfield{type}, \ibibfield{number}, \ibibfield{address}, \ibibfield{month}, \ibibfield{note} \tabularnewline \ibibentry{unpublished} & \ibibfield{author}, \ibibfield{title}, \ibibfield{note} & \ibibfield{month}, \ibibfield{year} \end{fieldtab}\screenpagebreak The \ibibfield{author} and \ibibfield{editor} fields have the same format as described in \sectionref{sec:jabref}. That is, each name should be in one of the forms listed in \tableref{tab:bibnameformats}, and multiple authors or editors must be separated with the keyword \dq{and}. \xminisec{Example (Multiple Authors):} This example uses the \ibibentry{book} entry: \begin{code} \begin{verbatim} @book{goossens97, author = "Goossens, Michel and Rahtz, Sebastian and Mittelbach, Frank", title = "The \LaTeX\ graphics companion: illustrating documents with \TeX\ and {PostScript}", publisher = "Addison Wesley Longman, Inc", year = 1997 } \end{verbatim} \end{code} In this example, the \meta{keyword} is \verb|goossens97|. That is the identifying key used in \glsi{cite}, described \htmlref{below}{sec:bibtex}. The standard bibliography styles usually convert titles to lower case, so the name PostScript is enclosed in curly braces to prevent this from happening. Note that curly braces \verb|{}| can be used instead of double quotes. The above example can just as easily be written: \begin{code} \begin{verbatim} @book{goossens97, author = {Goossens, Michel and Rahtz, Sebastian and Mittelbach, Frank}, title = {The \LaTeX\ graphics companion: illustrating documents with \TeX\ and {PostScript}}, publisher = {Addison Wesley Longman, Inc}, year = 1997 } \end{verbatim} \end{code} Numbers (such as the year 1997) don't need to be delimited with quotes or braces. So you can have \begin{codeS} pages = 10 \end{codeS} but a~page range would need to be delimited: \begin{codeS} \begin{alltt} pages = "10\gls{endash}45" \end{alltt} \end{codeS} Bibliography styles always have three-letter abbreviations for months: \texttt{jan}, \texttt{feb}, \texttt{mar}, etc. These should be used instead of typing them in explicitly, as their format depends on the bibliography style. These abbreviations should be entered without quotes. For example: \begin{code} \begin{verbatim} @inproceedings{talbot97, author = "Talbot, Nicola and Cawley, Gavin", title = "A fast index assignment algorithm for robust vector quantisation of image data", booktitle = "Proceedings of the I.E.E.E. International Conference on Image Processing", address = "Santa Barbara, California, USA", month = oct, year = 1997 } \end{verbatim} \end{code} \setnode{bibtex} \section{BibTeX} \label{sec:bibtex} Now that we've created a~\texttt{.bib} file (as described \htmlref{above}{sec:creatingbibfile}) we next need to look at how to incorporate the information in the database into a~\LaTeX\ document. As mentioned in \novices[sec:bib]{biblio}, entries are cited in the document using: \begin{definition} \gls{cite}\oarg{\meta{text}}\marg{\meta{key list}} \end{definition} where \meta{key list} is a~comma-separated list of keys. Each key uniquely identifies an entry in the database. If you used \appname{JabRef} (\sectionref{sec:jabref}), this is the key you entered in the \dq{Bibtexkey} field. If you wrote the \texttt{.bib} file in a~text editor (\sectionref{sec:bibformat}) it's the \meta{keyword} bit at the start of the list of fields for the entry. Next you need to specify what type of bibliography style you want to use. There are many available, but the basic ones are:\faq{Choosing a~bibliography style}{whatbst} \begin{description} \item[\ibst{abbrv}] Entries sorted alphabetically with abbreviated first names, months and journal names. \item[\ibst{alpha}] Entries sorted alphabetically with the citation represented by abbreviated author surname and year instead of a~number. \item[\ibst{plain}] Entries sorted alphabetically, with the citation represented by a~number. \item[\ibst{unsrt}] Entries sorted according to citation with the citation represented by a~number. \end{description} The style is specified in your \LaTeX\ document with the command: \begin{definition} \gls{bibliographystyle}\marg{\meta{style}} \end{definition} where \meta{style} is the name of the style. Some people put this command in the document's preamble and some people put it near their bibliography, but wherever you choose to put it, this command should only be used once. The actual bibliography itself is input into the document using \begin{definition} \gls{bibliography}\marg{\meta{database}} \end{definition} where \meta{database} is the name of the database \emph{without the \texttt{.bib} extension}. In fact, this argument can be a comma-separated list of databases if your entries are stored across multiple files. Recall the example thesis in \listingref{ex:thesis1} ended with: \begin{code} \begin{alltt} \glsni{percentchar} The bibliography will go here \glsni{end}\marg{document} \end{alltt} \end{code}\bookpagebreak\noindent If my references are stored in the file \examplecode{thesis-ref.bib}, then I~can replace the above comment as follows: \begin{codelisting}{thesis-biblio.tex} \begin{alltt} \glsni{bibliographystyle}\marg{plain} \glsni{bibliography}\marg{thesis-ref} \glsni{end}\marg{document} \end{alltt} \end{codelisting} Elsewhere in my document I~need some citations. For example: \begin{code} See Turabian\glsni{tildechar}\glsni{cite}\marg{turabian96} for a~comprehensive guide on preparing a~thesis. \end{code} If you are using \iappname{arara} (see \sectionref{sec:arara}) you need the following lines in your source code: \begin{code} \begin{alltt} \gls{percentchar.arara} pdflatex: \marg{ synctex: on } \glsni{percentchar.arara} \iappname{bibtex} \glsni{percentchar.arara} pdflatex: \marg{ synctex: on } \glsni{percentchar.arara} pdflatex: \marg{ synctex: on } \end{alltt} \end{code} If you are using \iappname{latexmk} (see \sectionref{sec:latexmk}) make sure you are using the \texttt{-bibtex} argument (\figureref{fig:texworks-latexmkbibtex}). If you are not using either \appname{latexmk} or \appname{arara}, you will need to run \PDFLaTeX, then run Bib\TeX, then run \PDFLaTeX\ twice more (see \sectionref{sec:build}). If your citations appear as two question marks ?? in your PDF, then the citation key you used hasn't been recognised. This could be that you've forgotten the Bib\TeX\ and subsequent \emph{two} \PDFLaTeX\ calls, or it could be that the key hasn't been defined, or you have misspelt it. Recall from \novices[sec:bib]{biblio} that the bibliography doesn't usually get added to the table of contents for most class files, but the \koma\ classes provide the options \scrclsopt[totocnumbered]{bibliography} and \scrclsopt[totoc]{bibliography}, that add a~numbered or unnumbered bibliography to the table of contents. You can add backlinks from your bibliography back to the section or page where the entries were cited using the \istyopt{hyperref}{backref} option of the \isty{hyperref} package. (The \sty{hyperref} package should usually be loaded last.) For example, to have backreferences to the pages on which the citation occurs: \begin{codeS} \gls{usepackage}\oarg{backref}\marg{hyperref} \end{codeS} The \sty{hyperref} package is covered in more detail in \latexpdfdoc. \setnode{natbib} \subsection{Author--Year Citations} \label{sec:natbib} The default behaviour of citations with bibliography styles such as \texttt{plain} is to produce a~numerical reference in square brackets. If you're using \iappname{bibtex} (rather than \isty{biblatex}, described \htmlref{below}{sec:biblatex}) you can override this using a~number of packages. One such package is \isty{natbib}. This comes with some drop-in replacements for the standard bibliography styles: \ibst{plainnat}, \ibst{unsrtnat} and \ibst{abbrvnat}. The \sty{natbib} package comes with a~variety of package options, but I'm just going to mention a~few of them: \istyopt{natbib}{authoryear} for author--year citations (default), \istyopt{natbib}{numbers} for numerical citations, \istyopt{natbib}{super} for superscripted numerical citations, \istyopt{natbib}{round} for round parentheses, \istyopt{natbib}{square} for square parentheses and \istyopt{natbib}{sort\&compress} which sorts multiple citations and compresses consecutive numbers into a~range. For example, [4,2,8,3] will become [2--4,8]. So for citations that give the author and year rather than a~number, you need to load \sty{natbib} in the preamble: \begin{codeS} \glsni{usepackage}\oarg{round}\marg{natbib} \end{codeS} and specify one of the \sty{natbib} bibliography styles: \begin{codeS} \gls{bibliographystyle}\marg{plainnat} \end{codeS} There are two main replacements for \gls{cite}: \begin{definition} \gls{citet}\oarg{\meta{pre}}\oarg{\meta{post}}\marg{\meta{key}} \end{definition} for textual citations and \begin{definition} \gls{citep}\oarg{\meta{pre}}\oarg{\meta{post}}\marg{\meta{key}} \end{definition} for parenthetical citations. Unlike \glsni{cite}, these commands have two optional arguments. The second \meta{post} is a~suffix, the same as \glsni{cite}'s only optional argument. The first optional argument \meta{pre} is a~prefix. If only one optional argument is present, it is assumed to be \meta{post}, so if you only want a~prefix and no suffix, you have to specify an empty argument for \meta{post}. \xminisec{Example:} (Using the same \examplecode{thesis-ref.bib} database as earlier.) \begin{codelisting}{thesis-nat.tex} \null\par A textual citation \glsni{citet}\marg{turabian96} and a~parenthetical citation \glsni{citep}\oarg{see}\oarg{Chapter 9}\marg{goossens97}. \end{codelisting} Result: \begin{result}[natbib.html] A textual citation Turabian (1996) and a~parenthetical citation (see Goossens et al., 1997, Chapter 9). \end{result} \setnode{bibtextroubleshooting} \subsection{Troubleshooting} \label{sec:bibtextroubleshooting} \begin{itemize} \item \BiBTeX\ writes the \gls{env-thebibliography} environment to a~\texttt{.bbl} file, which is then input into the document by \gls{bibliography}. If you have made a~\LaTeX\ error in the \texttt{.bib} file, this error will be copied to the \texttt{.bbl} file. If you have corrected the error in the \texttt{.bib} file, but you are still getting an error when you \LaTeX\ your document, try deleting the \texttt{.bbl} file. (In TeXworks, you can use the menu item \menu{File}\menuto\menu{Remove Aux Files}.) \item Remember to use double quotes or braces to delimit the field names in your \texttt{.bib} file. \item Remember to put a~comma at the end of each field entry (except the last). \item It is better to only use alphanumerical characters in the keywords. Some punctuation characters such as \texttt{.} (full stop) should be fine (unless you're using a~package such as \sty{babel} that makes them active), but spaces are not recommended, and commas should definitely be avoided. \item If you have entered a~field in the \texttt{.bib} file, but it doesn't appear in the bibliography, check to make sure that the field is required or optional for that type of entry, and check the spelling. (You can avoid this problem by using a~bibliography management system such as \htmlref{JabRef}{sec:jabref}.) \item Check the \BiBTeX\ log file (\texttt{.blg}) for messages. \item If you get an error that looks something like: \begin{verbatim} ERROR - Cannot find control file 'thesis-ref.bcf'! - did you pass the "backend=biber" option to BibLaTeX? \end{verbatim} then you have inadvertently used \iappname{biber} (see \htmlref{below}{sec:biblatex}) instead of \iappname{bibtex}. \item If you get an error that looks something like: \begin{verbatim} I found no \citation commands---while reading file thesis1.aux I found no \bibdata command---while reading file thesis1.aux I found no \bibstyle command---while reading file thesis1.aux \end{verbatim} then you probably forgot to use the \gls{bibliography} and \gls{bibliographystyle} commands in your document. \end{itemize} \setnode{biblatex} \section{Biblatex} \label{sec:biblatex} The \istystart{biblatex} package is a~reimplementation of \LaTeX's bibliographic facilities. The formatting of the bibliography is governed by \LaTeX\ commands instead of selecting a~BibTeX style (as was done with \gls{bibliographystyle} described \htmlref{above}{sec:bibtex}). This package uses \iappnamelink{biber}{http://biblatex-biber.sourceforge.net/} instead of BibTeX to process the bibliographic database and sort the entries. Legacy BibTeX is also supported, but with a~reduced feature set. The \sty{biblatex} package also supports multiple bibliographies, for example a~bibliography for each chapter in the document. The \sty{biblatex} package requires e-\TeX, so make sure you have a~recent \TeX\ distribution. Biber comes with the latest version of \TeX~Live. If you are using \iappname{JabRef} (described in \sectionref{sec:jabref}) there is a~BibLaTeX mode option in the Advanced tab of the \appname{JabRef} preferences dialog, illustrated in \figureref{fig:jabref9}. (Use \menu{Options}\menuto\menu{Preferences} to open the dialog.) You will have to quit and restart \appname{JabRef} after enabling this option. When you restart, you should find extra fields when you edit an entry or create a~new entry, as illustrated in \figureref{fig:jabref10}. You should also find that there are more entry types available (see \figureref{fig:jabref11}). \begin{figure}[htbp] \figconts {pictures/jabref9} {\caption{JabRef Advanced Preferences}} {fig:jabref9} \end{figure} \begin{figure}[htbp] \figconts {pictures/jabref10} {\caption{JabRef in BibLaTeX Mode}} {fig:jabref10} \end{figure} \begin{figure}[htbp] \figconts {pictures/jabref11} {\caption{JabRef in BibLaTeX Mode (Select Entry Type)}} {fig:jabref11} \end{figure} With BibTeX, there was a~\ibibfield{month} and \ibibfield{year} field. BibLaTeX provides a~replacement \ibibfield{date} field, although if this field is missing it will fall back on the \bibfield{month} and \bibfield{year} fields. In \figureref{fig:jabref12}, I've edited my earlier example to use the new \bibfield{date} field. Note that the date should be specified as \meta{year}\texttt{-}\meta{month}\texttt{-}\meta{day} where \texttt{-}\meta{day} or \texttt{-}\meta{month}\texttt{-}\meta{day} maybe omitted. A slash \texttt{/} should be used to indicate a range, for example \texttt{2002-01/2002-02}. \begin{figure}[htbp] \figconts {pictures/jabref12} {\caption{JabRef in BibLaTeX Mode (Setting the Publication Date)}} {fig:jabref12} \end{figure} Recall from \figureref{fig:jabref-pref} and \figureref{fig:jabref-dataprop} that I~set the default encoding to UTF-8. With BibLaTeX and \appname{biber}, my UTF-8 bibliography can be correctly sorted, but I~need to make sure that I~load the \isty{inputenc} package before \sty{biblatex} in my document: \begin{code} \glsni{usepackage}\oarg{utf8}\marg{inputenc}\newline \glsni{usepackage}\marg{biblatex} \end{code} \sectionref{sec:natbib} described the \isty{natbib} package. BibLaTeX has a~compatibility module: \begin{codeS} \glsni{usepackage}\oarg{natbib}\marg{biblatex} \end{codeS} This provides the same commands (such as \gls{citet} and \gls{citep}) that \isty{natbib} provides. The default sorting order is name, title and year. This can be changed using the \istyopt{biblatex}{sorting} package option. For example, to sort by name, year and title: \begin{codeS} \glsni{usepackage}\oarg{sorting=nyt}\marg{biblatex} \end{codeS} Or you can suppress the sorting, so that all entries are in citation order: \begin{codeS} \glsni{usepackage}\oarg{sorting=none}\marg{biblatex} \end{codeS} For other possible values, see the \sty{biblatex} documentation~\cite{biblatex}. If you want a~list of back-references in the bibliography, referring to the pages on which the entries were cited, you can use the \istyopt{biblatex}{backref} option: \begin{codeS} \glsni{usepackage}\oarg{backref}\marg{biblatex} \end{codeS} The default database backend is \iappname{biber}, which is recommended, but if for some reason you want to stick to using \iappname{bibtex} you can use the \istyopt{biblatex}{backend} option to switch to \optfmt{bibtex}: \begin{codeS} \glsni{usepackage}\oarg{backend=bibtex}\marg{biblatex} \end{codeS} There are also options that govern whether certain fields are printed in the bibliography, such as \istyopt{biblatex}{isbn}, \istyopt{biblatex}{url} or \istyopt{biblatex}{doi}. For example: \begin{codeS} \glsni{usepackage}\oarg{isbn,url,doi}\marg{biblatex} \end{codeS} The style can be set using the \istyopt{biblatex}{style} option. The default is \optfmt{numeric}, which produces a~numeric citation, such as [1]. There is also \optfmt{numeric-comp}, which is like \isty{natbib}'s \istyopt{natbib}{sort\&compress} option, described in \sectionref{sec:natbib}, or \optfmt{authoryear} which displays \meta{author} \meta{year} citations. There are many other citation styles. For these and for other package options, see the \sty{biblatex} documentation~\cite{biblatex}. With BibLaTeX, you don't use the \gls{bibliography} command, described in \sectionref{sec:bibtex}. Instead, you add the bib file as a~resource \emph{in the preamble} using: \begin{definition} \gls{addbibresource}\oarg{\meta{options}}\marg{\meta{resource}} \end{definition} where \meta{resource} is the name of the bib file \emph{including the file extension}. However, the resource doesn't have to be a~bib file. You can only add one resource at a~time: \begin{code} \glsni{addbibresource}\marg{bibfile1.bib}\newline \glsni{addbibresource}\marg{bibfile2.bib} \end{code} The resource can be a~remote one, in which case you need to use the \ikeyvalopt{addbibresource}{location} option with the value \texttt{remote} and specify the URL: \begin{code} \glsni{addbibresource}\oarg{location=remote}\marg{http://www.somewhere.com/bibfile2.bib} \end{code} This is only available if you use \iappname{biber} as the backend. Another option is \ikeyvalopt{addbibresource}{datatype} which specifies the format of the resource. The default is \texttt{bibtex}, but it can also be \Indextt{ris}, \Indextt{zoterordfxm} or \Indextt{endnotexml}. See the \sty{biblatex} and \sty{biber} documentation~\cite{biber} for further details. The bibliography itself is displayed using \begin{definition} \gls{printbibliography}\oarg{\meta{options}} \end{definition} This should go in the document where you want the bibliography to be displayed. Like the \isty{natbib} commands described in \sectionref{sec:natbib}, the \sty{biblatex} commands generally have two optional arguments, indicating the prenote and postnote, and a mandatory argument specifying the key or a~comma-separated list of keys. If you want a~prenote but not a~postnote, you need to give an empty second optional argument. The basic commands are: \begin{definition} \gls{cite}\oarg{\meta{prenote}}\oarg{\meta{postnote}}\marg{\meta{key}}\newline \gls{Cite}\oarg{\meta{prenote}}\oarg{\meta{postnote}}\marg{\meta{key}} \end{definition} These are bare citation commands. The latter is provided if the citation occurs at the start of a~sentence. \begin{definition} \gls{parencite}\oarg{\meta{prenote}}\oarg{\meta{postnote}}\marg{\meta{key}}\newline \gls{Parencite}\oarg{\meta{prenote}}\oarg{\meta{postnote}}\marg{\meta{key}} \end{definition} These commands are like \glsni{cite} and \glsni{Cite} but enclose the citation in parentheses (square if the numeric style is used). \begin{definition} \gls{textcite}\oarg{\meta{prenote}}\oarg{\meta{postnote}}\marg{\meta{key}}\newline \gls{Textcite}\oarg{\meta{prenote}}\oarg{\meta{postnote}}\marg{\meta{key}} \end{definition} These commands are used for citations in the flow of text. The latter is provided if the citation occurs at the start of a sentence. For other citation commands, see the \sty{biblatex} documentation~\cite{biblatex}. So, the example document from \listingref{ex:thesis1}, can now be edited so that the preamble looks like: \begin{codelisting}{thesis-biblatex.tex}\label{ex:thesis-biblatex} \begin{alltt} \glsni{percentchar.arara} pdflatex: \marg{ synctex: on } \glsni{percentchar.arara} \iappname{biber} \glsni{percentchar.arara} pdflatex: \marg{ synctex: on } \glsni{percentchar.arara} pdflatex: \marg{ synctex: on } \glsni{documentclass}\oarg{oneside}\marg{scrbook} \glsni{usepackage}\oarg{backend=biber}\marg{biblatex} \glsni{addbibresource}\marg{thesis-refs.bib} \end{alltt} \end{codelisting} (where \examplecode{thesis-refs.bib} is the name of my bibliography database, see \sectionref{sec:creatingbibfile}) and the end of the document looks like: \begin{code} \begin{alltt} \glsni{printbibliography} \glsni{end}\marg{document} \end{alltt} \end{code} Elsewhere in the document, I~need to cite some of the entries in my bibliography database: \begin{code} \null First of all, let's cite a book\glsni{tildechar}\glsni{parencite}\marg{wainwright93} now let's cite a~journal paper and a~conference proceedings\glsni{tildechar}\glsni{parencite}\marg{cawley96,talbot97}. Finally, let's cite a~chapter in a book\glsni{tildechar}\glsni{parencite}\oarg{Chapter 9}\marg{goossens97}. \null \end{code} If you want to build the document using \iappname{arara} (\sectionref{sec:arara}) remember to include the \gls{percentchar.arara} comments (as shown above). If you are using \iappname{latexmk} (\sectionref{sec:latexmk}) remember to use the \texttt{-bibtex} option as illustrated in \figureref{fig:texworks-latexmkbibtex}. If you're not using an automated method, such as \appname{arara} or \appname{latexmk}, you need a~\PDFLaTeX\ run, a \appname{biber} run (or \appname{bibtex} if you've chosen that as your backend) followed by two more \PDFLaTeX\ runs. \setnode{biblatextroubleshooting} \subsection{Troubleshooting} \label{sec:biblatextroubleshooting} Most of the comments from the \htmlref{\BiBTeX\ troubleshooting section}{sec:bibtextroubleshooting}\doifbook{ (see page~\pageref{sec:bibtextroubleshooting})} also apply here. If you get an error that looks like: \begin{verbatim} I found no \citation commands---while reading file thesis-biblatex.aux I found no \bibdata command---while reading file thesis-biblatex.aux I found no \bibstyle command---while reading file thesis-biblatex.aux \end{verbatim} then you have inadvertently used \iappname{bibtex} instead of \iappname{biber}. If you actually want to use \appname{bibtex} with the \istyend{biblatex} package remember that you have to specify \appname{bibtex} using: \begin{codeS} \glsni{usepackage}\oarg{backend=bibtex}\marg{biblatex} \end{codeS} % Indexes and Glossaries \setnode{indgloss} \chapter{Generating Indexes and Glossaries} \label{ch:indgloss} Most theses will need a~glossary of terms or a~list of acronyms or notation. It's less likely that you'll need an index in your thesis, but since the same mechanism is used to generate glossaries and indexes, both topics are covered in this chapter. There are two basic methods of generating a~glossary or index: \begin{enumerate} \item The glossary or indexing information is written to a~temporary file by \LaTeX\ while the document is being built. An external application is then used to collate and sort the entries defined in that temporary file and \LaTeX\ code to display the result is written to another file. You then need to run (PDF)\LaTeX\ on your document to ensure the sorted and collated glossary or index is displayed. (You may then need an additional \LaTeX\ run to ensure the table of contents is up-to-date.) This is similar to the way you had to use \appname{bibtex} or \appname{biber} between LaTeX runs in the \htmlref{previous chapter}{ch:citations}. \item The glossary or indexing information is collated and sorted by \LaTeX\ during the document build. (At least two runs are required, but no external indexing application is needed.) \end{enumerate} The first approach (see \sectionref{sec:makeindexglos}) is more efficient, but a~lot of users, especially beginners, have difficulty with the intermediate step where the external indexing application is run. The second approach (see \sectionref{sec:datagidx}) is slower, but you don't need to worry about running an indexing application. If you're not writing in English (in particular if you are not using the Latin alphabet) you're better off using the first approach with \iappname{xindy}. In this chapter I'll describe both approaches and you can choose which you prefer. \setnode{makeindexglos} \section{Using an External Indexing Application} \label{sec:makeindexglos} This section describes how to create indexes (\sectionref{sec:makeidx}) or glossaries (\sectionref{sec:makeglossaries}) using an external indexing application. There are two popular indexing applications: \iappname{makeindex} and \iappname{xindy}. All \TeX\ distributions should come with \appname{makeindex}. The \TeX~Live distribution also comes with \appname{xindy}, but if you have a different \TeX\ distribution (such as MikTeX) you may need to fetch \iappname{xindy} from \url{http://www.xindy.org/}. \xminisec{Note:} You must have Perl installed in order to use \iappname{xindy} as it's a~Perl script. (See \xrsectionref{Perl}{perl} from Volume~1.) If you have successfully been using \htmlref{\appname{latexmk}}{sec:latexmk}, you already have Perl installed. \setnode{makeidx} \subsection{Creating an Index (\texorpdfstring{\sty{makeidx}}{makeidx} package)} \label{sec:makeidx} \novices[ch:newcom]{newcom} introduced the command: \begin{definition} \gls{index}\marg{\meta{text}} \end{definition} to index the word given in \meta{text}. For example, if \verb|\index{circuit}| occurs on page~42, then \dq{42} will be added to the \keyword{location list} for the term \dq{circuit}. \xminisec{Note:} \warning \glsni{index} doesn't display any text. It just adds a~line to the index file with the information required by \appname{makeindex} or \appname{xindy} to sort and collate the information. The default action of \glsni{index} simply ignores its argument. To ensure the indexing mechanism works, you must activate it by placing \begin{definition} \gls{makeindex} \end{definition} in the document preamble. Finally, you need to use \begin{definition} \gls{printindex} \end{definition} (defined in the \isty{makeidx} package) to display the index. \xminisec{Note:} \warning \glsni{printindex} won't produce any text until you have run the external indexing application. Here's an example document: \begin{codelisting}{sample-index.tex}\label{ex:sample-index} \begin{alltt} \glsni{percentchar.arara} pdflatex: \marg{ synctex: on } \glsni{percentchar.arara} \iappname{makeindex} \glsni{percentchar.arara} pdflatex: \marg{ synctex: on } \glsni{documentclass}\oarg{12pt,oneside}\marg{scrbook} \glsni{usepackage}\marg{makeidx} \glsni{makeindex} \glsni{title}\marg{Sample Document} \glsni{author}\marg{Me} \glsni{begin}\marg{document} \glsni{maketitle} \glsni{chapter}\marg{Sample} Stuff about eigenvectors\glsni{index}\marg{eigenvector} and eigenvalues\glsni{index}\marg{eigenvalue}. \glsni{chapter}\marg{Another Sample} Some more stuff about eigenvectors\glsni{index}\marg{eigenvector} and eigenvalues\glsni{index}\marg{eigenvalue}. Something about eigen-decomposition\glsni{index}\marg{eigen-decomposition}. \glsni{backmatter} \glsni{printindex} \glsni{end}\marg{document} \end{alltt} \end{codelisting} If you are using \iappname{arara} to build your document (see \sectionref{sec:arara}), remember to include the \glsni{percentchar.arara} comments, as shown in the above listing. If you are using \iappname{latexmk} to build your document, remember to include the \texttt{.idx} custom dependency to your RC file, as described in \sectionref{sec:latexmk}. If you aren't using an automated method to build your document, you will need to run \PDFLaTeX, then run \iappname{makeindex}, and then run \PDFLaTeX\ again (see \sectionref{sec:build}). If you prefer to use \iappname{xindy} instead of \iappname{makeindex}, you need to run \iappname{texindy} (a \appname{xindy} wrapper customised for \LaTeX\ documents). If you are using \iappname{arara}, change the line: \begin{alltt} \gls{percentchar.arara} makeindex \end{alltt} to (change the language as required): \begin{alltt} \gls{percentchar.arara} texindy: \marg*{ language: english } \end{alltt} (Make sure you have added the \appname{texindy} rule as described in \sectionref{sec:arara}.) If you are using \iappname{latexmk} to build your document, you will need to change the custom dependency for \texttt{.idx} files, as described in \sectionref{sec:latexmk}. \setnode{makeindexindexsort} \subsubsection{Overriding the Default Sort} \label{sec:makeindexsort} By default the index entry will be sorted according to the word being indexed. However, you can override this by writing the argument of \glsni{index} in the form: \begin{definition} \meta{sort}\gls{atchar}\meta{word} \end{definition} where \meta{sort} is how to sort the term and \meta{word} is how the term should appear in the index. The \iappname{makeindex} application doesn't understand \LaTeX\ commands. It simply sorts the term as is. So, for example, if you do \begin{codeS} \glsni{index}\marg{\gls{AE} olian} \end{codeS} then \iappname{makeindex} will sort it according to the characters \gls{backslashchar}, A, E, \glsni{visiblespace} (space), o, l, i, a, n. Since \appname{makeindex} sorts symbols (such as \glsni{backslashchar}) before letters, it will put \texttt{\glsni{AE}\glsni{visiblespace}olian} before, say, \texttt{adze}, since \glsni{backslashchar} comes before \dq{a}. To get around this, you need to specify the sort key: \begin{codeS} \glsni{index}\marg{AEolian\glsni{atchar}\glsni{AE} olian} \end{codeS} Now \appname{makeindex} will put \dq{\AE olian} after \dq{adze}. Here's another example that indexes a~function or method: \begin{codeS} \glsni{index}\marg{sqrt()\glsni{atchar}\gls{texttt}\marg{sqrt()}} \end{codeS} You will also need to do something similar if you are entering the character directly via the \sty{inputenc} package: \begin{codeS} \glsni{index}\marg{elite\glsni{atchar}\'elite} \end{codeS} Note, however, that you don't need to do this if you are using \iappname{xindy}. You just need to make sure you match the input encoding. For example: \begin{code} \begin{alltt} \glsni{percentchar.arara} pdflatex: \marg{ synctex: on } \glsni{percentchar.arara} \iappname{texindy}: \marg{ language: english, codepage: latin1} \glsni{percentchar.arara} pdflatex: \marg{ synctex: on } \glsni{documentclass}\oarg{12pt,oneside}\marg{scrbook} \glsni{usepackage}\oarg{latin1}\marg{inputenc} \glsni{usepackage}\marg{makeidx} \end{alltt} \end{code} Later in the document: \begin{codeS} \glsni{index}\marg{\'elite} \end{codeS} \setnode{makeindexformat} \subsubsection{Setting the Location Format} \label{sec:makeindexformat} Each index entry has an associated \keyword{location list} that directs the reader to the pages in the document associated with that entry. For example, if you look up \gls*{index} in this book's \latexhtml{\htmlref{index}{ch:index}}{\htmladdnormallink{index}{bookindex.html}}, the entry's location list will include this page. If the location list is long, it's helpful to highlight a~particular location to direct the reader to the principle definition or discussion related to that term. This is usually done by formatting the relevant location in a~different font, for example bold or italic. \bookpagebreak You can specify the format for the location by writing the argument of \glsni{index} in the form: \begin{definition} \meta{word}\gls{index.barchar}\meta{format} \end{definition} where \meta{format} is the name of a~text-block command \emph{without} the leading backslash. For example: \begin{codeS} \glsni{index}\marg{eigenvector\glsni{index.barchar}textbf} \end{codeS} You can combine \gls{atchar} and \glsni{index.barchar}. For example: \begin{codeS} \glsni{index}\marg{sqrt()\glsni{atchar}\gls{texttt}\marg{sqrt()}\glsni{index.barchar}textbf} \end{codeS} \xminisec{Note:} \warning Make sure the format you use is the name of a~command that takes an argument. While it won't cause an error to use, say, \texttt{bfseries} instead of \texttt{textbf}, it will cause the unexpected side-effect of rendering the rest of your index in that font, instead of just that particular location. You can also use \meta{format} to cross-reference another entry. If you have an entry that's just a~synonym for another entry, you can use: \begin{definition} \meta{word}\glsni{index.barchar}see\marg{\meta{name}} \end{definition} where \meta{name} is the other entry. If you want to direct the reader to a~similar topic, you can use: \begin{definition} \meta{word}\glsni{index.barchar}seealso\marg{\meta{topic}} \end{definition} where \meta{topic} is the other entry. For example: \begin{codeS} \glsni{index}\marg{eigenvector\glsni{index.barchar}seealso\marg{eigenvalue}} \end{codeS} \setnode{makeindexsublevels} \subsubsection{Sub Levels} \label{sec:makeindexsublevels} An entry in the index may have sub-items. With \iappname{makeindex} you can have a~maximum of three levels. With \iappname{xindy} you can have an arbitrary number of levels. However, it's a~good idea to consider the advice in the Oxford Style Manual~\cite{oxford}: \dq{In all but the most complex indexes, subentries within subentries (\emph{sub-subentries}) should be avoided.} In other words, just because it's possible to do something doesn't mean you should do it. To indicate a~subentry, the argument of \gls{index} should be in the form: \begin{definition} \meta{main entry}\gls{makeindex.exclamchar}\meta{subentry} \end{definition} For example: \begin{codeS} \glsni{index}\marg{reptile\glsni{makeindex.exclamchar}caiman} \end{codeS}\bookpagebreak If you really must have a~sub-subentry: \begin{codeS} \glsni{index}\marg{reptile\glsni{makeindex.exclamchar}crocodylian\glsni{makeindex.exclamchar}caiman} \end{codeS} You can combine \gls{atchar}, \gls{index.barchar} and \glsni{makeindex.exclamchar}. For example: \begin{codeS} \glsni{index}\marg{methods\glsni{makeindex.exclamchar}sqrt()\glsni{atchar}\glsni{texttt}\marg{sqrt()}\glsni{index.barchar}textbf} \end{codeS} \listingref{ex:thesis-biblatex} can now be modified as follows (download the document for the complete code):\reportpagebreak \begin{codelisting}{thesis-index.tex}\label{ex:thesis-index} \begin{alltt} \glsni{percentchar} In the preamble: \glsni{percentchar.arara} pdflatex: \marg{ synctex: on } \glsni{percentchar.arara} biber \glsni{percentchar.arara} \iappname{makeindex} \glsni{percentchar.arara} pdflatex: \marg{ synctex: on } \glsni{percentchar.arara} pdflatex: \marg{ synctex: on } \glsni{documentclass}\oarg{oneside,12pt}\marg{scrbook} \glsni{usepackage}\marg{makeidx} \gls{makeindex} \glsni{percentchar} Later in the document: Some sample code is shown in Listing\gls{tildechar}\gls{ref}\marg{lst:sample}. This uses the function \gls{lstinline}"sqrt()"\glsni{percentchar} \gls{index}\marg{sqrt()\gls{atchar}\glsni{texttt}\marg{sqrt()}}\glsni{percentchar} \glsni{index}\marg{functions\gls{makeindex.exclamchar}sqrt()\glsni{atchar}\glsni{texttt}\marg{sqrt()}}\glsni{percentchar} \glsni{index}\marg{square root\gls{index.barchar}see\marg{\glsni{texttt}\marg{sqrt()}}}. \glsni{begin}\marg{Definition}\oarg{Tautology} A \glsni{emph}\marg{tautology}\glsni{index}\marg{tautology\glsni{index.barchar}textbf} is a proposition that is always true for any value of its variables. \glsni{end}\marg{Definition} \glsni{begin}\marg{Definition}\oarg{Contradiction} A \glsni{emph}\marg{contradiction}\glsni{index}\marg{contradiction\glsni{index.barchar}textbf} is a proposition that is always false for any value of its variables. \glsni{end}\marg{Definition} \glsni{percentchar} At the end of the document: \glsni{printbibliography} \glsni{printindex} \glsni{end}\marg{document} \end{alltt} \end{codelisting} The index for the above document looks like: \begin{result}[sampleindex.html] \begin{simtheindex} \item contradiction, \textbf{2} \indexspace \item functions \subitem \texttt{sqrt()}, 2 \indexspace \item \texttt{sqrt()}, 2 \item square root, \see{\texttt{sqrt()}}{2} \indexspace \item tautology, \textbf{2} \end{simtheindex} \end{result} \setnode{makeindextroubleshooting} \subsubsection{Troubleshooting} \label{sec:makeindextroubleshooting} \begin{itemize} \item My index hasn't appeared. \begin{enumerate} \item Make sure you have the command \gls{printindex} at the place where you want the index to appear (this command is defined in the \isty{makeidx} package). \item Make sure you have the command \gls{makeindex} in the preamble. \item If you are building the document using \iappname{arara} make sure you included all the \gls{percentchar.arara} directives as shown in \listingref{ex:thesis-index}. If you are using \iappname{latexmk}, make sure you have included the \texttt{.idx} dependency, as described in \sectionref{sec:latexmk}. If you're not using an automated tool, make sure you run (PDF)\LaTeX, then \iappname{makeindex} and then (PDF)\LaTeX\ again (see \sectionref{sec:build}). \item Check \appname{makeindex}'s log file (which has the extension \texttt{.ilg} by default) for error messages. \end{enumerate} \item I~want to index the character \gls{makeindex.doublequote}, \gls{atchar}, \gls{makeindex.exclamchar} or \gls{index.barchar} but it's not working. If you want any of these symbols in your index, you will need to prefix the character with the double-quote symbol \glsni{makeindex.doublequote}. For example to index the \glsni{atchar} symbol: \begin{codeS} \gls{index}\marg{\glsni{makeindex.doublequote}\glsni{atchar}} \end{codeS} \item I~have multiple entries of the same item. For example: \null~~~identity matrix, 10, 22--30\newline \null~~~identity matrix, 4\newline\null Check to make sure the sort argument to each of the corresponding \glsni{index} commands is the same. Pay particular attention to spaces as \iappname{makeindex} will treat the following entries differently: \begin{code} \begin{alltt} \glsni{index}\marg{identity\glsni{visiblespace}matrix} \glsni{index}\marg{identity\glsni{visiblespace}\glsni{visiblespace}matrix} \end{alltt} \end{code} \LaTeX\ however treats multiple spaces the same as a~single space, so the text will appear the same in the index. \item \LaTeX\ says that the command \gls{printindex} is undefined. You have forgotten to load the \sty{makeidx} package. \end{itemize} % Generating glossaries \setnode{makeglossaries} \subsection{Creating Glossaries, Lists of Symbols or Acronyms (\texorpdfstring{\sty{glossaries}}{glossaries} package)} \label{sec:makeglossaries} There are a~number of packages available to assist producing a~list of acronyms (such as the \isty{acronym} package) or a~glossary (such as the \isty{nomencl} package). You can see a~list of available packages in the \htmladdnormallinkfoot{OnLine \TeX\ Catalogue's Topic Index}{http://mirror.ctan.org/help/Catalogue/bytopic.html\#index}~\cite{texcattopic}. Here, I've chosen to describe the \isty{glossaries} package. Firstly, it encompasses the functionality of both \sty{acronym} and \sty{nomencl} as \isty{glossaries} allows you to define multiple lists of acronyms, lists of symbols or glossaries. Secondly, I~wrote the \sty{glossaries} package, so it's the one with which I~am most familiar. The \sty{glossaries} package is very flexible, but the downside to that is that it has too many features to cover briefly. I'm therefore only going to introduce the basics here. If you want more detail you'll have to read the user manual~\cite{glossaries}. I~will use the term \dq{glossary} to mean a~list of terms or a~list of notation or a~list of symbols or a~list of acronyms. \xminisec{Note:} \warning If you want to use both \isty{glossaries} and \isty{hyperref}, you must load \sty{hyperref} \emph{before} \sty{glossaries}. This is an exception to the usual advice of loading \sty{hyperref} last. \setnode{newglossaryentry} \subsubsection{Defining Glossary Entries} \label{sec:newglossaryentry} Firstly, in order to make the glossary (or glossaries, if you have more than one) appear, you must use the command \begin{definition} \gls{makeglossaries} \end{definition} in the preamble. This is analogous to the \glsni{makeindex} command described in \sectionref{sec:makeidx}. Next you need to define the terms you want to appear in the glossary. This is done using the command: \begin{definition} \gls{newglossaryentry}\marg{\meta{label}}\marg{\meta{key-val list}} \end{definition} The first argument \meta{label} is a~unique label so that you can refer to this entry in your document text. The entry will only appear in the glossary if you have referenced it in the document using one of the commands listed later. The second argument is a comma-separated list of \meta{key}=\meta{value} options. Common keys are: \begin{itemize} \item \ikeyvalopt{newglossaryentry}{name} The name of the entry (as it will appear in the glossary). \item \ikeyvalopt{newglossaryentry}{description} A brief description of this entry (to appear in the glossary). \item \ikeyvalopt{newglossaryentry}{text} How this entry will appear in the document text where the singular form is required. If this key is omitted the value of \keyvalopt{name} will be used. \item \ikeyvalopt{newglossaryentry}{first} How this entry will appear in the document text the first time it is used, where the first use requires the singular form. If this key is omitted the value of \keyvalopt{text} is used. \item \ikeyvalopt{newglossaryentry}{plural} How this entry will appear in the document text where the plural form is required. If this key is omitted, the value is obtained by appending the letter \dq{s} to the value of the \keyvalopt{text} key. \item \ikeyvalopt{newglossaryentry}{firstplural} How this entry will appear in the document text the first time it is used, where the first use requires the plural form. If this field is omitted, the value is obtained by appending the letter \dq{s} to the value of the \keyvalopt{first} key. \item \ikeyvalopt{newglossaryentry}{symbol} This key is provided to allow the user to specify an associated symbol, but most glossary styles ignore this value. \item \ikeyvalopt{newglossaryentry}{sort} This value indicates how to sort this entry (analogous to using the \gls{atchar} character in the argument of \gls{index}, as described in \sectionref{sec:makeidx}). If this key is omitted the value of \keyvalopt{name} is used. \item \ikeyvalopt{newglossaryentry}{type} This is the glossary type to which this entry belongs (see \sectionref{sec:newglossary}). If omitted the main (default) glossary is assumed. \end{itemize} \xminisec{Examples:} The following defines the term \dq{set} and assigns a brief description. The term is given the label \texttt{set}. This is the minimum amount of information you must give: \begin{code} \begin{alltt} \gls{newglossaryentry}\marg{set}\glsni{percentchar} the label \marg{\glsni{percentchar} name=\marg{set},\glsni{percentchar} the term description=\marg{a collection of objects}\glsni{percentchar} a brief description } \end{alltt} \end{code} The following entry also has an associated symbol: \begin{code} \begin{alltt} \glsni{newglossaryentry}\marg{U}\glsni{percentchar} the label \marg{\glsni{percentchar} name=\marg{universal set},\glsni{percentchar} the term description=\marg{the set of all things},\glsni{percentchar} a brief description symbol=\marg{\gls{ensuremath}\marg{\gls{mathcal}\marg{U}}}\glsni{percentchar} the associate symbol } \end{alltt} \end{code} The plural of the word \dq{matrix} is \dq{matrices} not \dq{matrixs}, so the term needs the plural form set explicitly: \begin{code} \begin{alltt} \glsni{newglossaryentry}\marg{matrix}\glsni{percentchar} the label \marg{name=\marg{matrix},\glsni{percentchar} the term description=\marg{a rectangular table of elements},\glsni{percentchar} brief description plural=\marg{matrices}\glsni{percentchar} the plural } \end{alltt} \end{code} The \sty{glossaries} package also provides the shortcut command: \begin{definition} \gls{newacronym}\oarg{\meta{key-val list}}\marg{\meta{label}}\marg{\meta{abbrv}}\marg{\meta{long}} \end{definition} The default behaviour of this command is equivalent to: \begin{code} \glsni{newglossaryentry}\marg{\meta{label}}\marg{name=\marg{\meta{abbrv}}\comma description=\marg{\meta{long}}\comma text=\marg{\meta{abbrv}}\comma first=\marg{\meta{long} (\meta{abbrv})}\comma plural=\marg{\meta{abbrv}s}\comma firstplural=\marg{\meta{long}s (\meta{abbrv}s)}\comma \meta{key-val list}} \end{code} \xminisec{Example:} \begin{codeS} \glsni{newacronym}\marg{svm}\marg{SVM}\marg{support vector machine} \end{codeS} is equivalent to \begin{code} \begin{alltt} \glsni{newglossaryentry}\marg{svm}\glsni{percentchar} the label \marg{\glsni{percentchar} name=\marg{SVM},\glsni{percentchar} description=\marg{support vector machine},\glsni{percentchar} first=\marg{support vector machine (SVM)},\glsni{percentchar} firstplural=\marg{support vector machines (SVMs)},\glsni{percentchar} text=\marg{SVM},\glsni{percentchar} plural=\marg{SVMs}\glsni{percentchar} } \end{alltt} \end{code} There are some package options that modify the behaviour of \glsni{newacronym}. For example, the package option \istyopt{glossaries}{description} changes \glsni{newacronym} so that you need to explicitly set the description in the optional argument. For example: \begin{code} \glsni{usepackage}\oarg{description}\marg{glossaries}\newline \mbox{}\newline \glsni{newacronym}\oarg{description=\marg{a~statistical pattern recognition technique}}\marg{svm}\marg{SVM}\marg{support vector machine} \end{code} Another package option is \istyopt{glossaries}{footnote} which will modify the behaviour of \glsni{newacronym} so that the long form is displayed as a~footnote on first use. For a~full list of available options, see the \sty{glossaries} documentation~\cite{glossaries}. \setnode{glslink} \subsubsection{Displaying Terms in the Document} \label{sec:glslink} Any glossary term that has been defined using \gls{newglossaryentry} or \reportlinebreak\gls{newacronym}, as described above, can be displayed in the document using one of the commands described in this section. (There are other less commonly used commands available as well, see the \sty{glossaries} documentation~\cite{glossaries} for details of them.) Each term has an associated \keyword{first use flag}. This is a boolean (true\slash false) switch that determines whether or not the entry has been used. This is how the \sty{glossaries} package determines whether to display the value of the \ikeyvalopt{newglossaryentry}{first} key or to display the value of the \ikeyvalopt{newglossaryentry}{text} key. You can reset this flag using:\bookpagebreak \begin{definition} \gls{glsreset}\marg{\meta{label}} \end{definition} Conversely, you can unset it using: \begin{definition} \gls{glsunset}\marg{\meta{label}} \end{definition} To display a~term that has previously been defined using either \reportlinebreak\gls{newglossaryentry} or \gls{newacronym} you can use one of the following commands: \begin{definition} \gls{gls}\oarg{\meta{options}}\marg{\meta{label}}\oarg{\meta{insert}} \end{definition} \begin{definition} \gls{glspl}\oarg{\meta{options}}\marg{\meta{label}}\oarg{\meta{insert}} \end{definition} \begin{definition} \gls{Gls}\oarg{\meta{options}}\marg{\meta{label}}\oarg{\meta{insert}} \end{definition} \begin{definition} \gls{Glspl}\oarg{\meta{options}}\marg{\meta{label}}\oarg{\meta{insert}} \end{definition} These commands all have the same syntax: \meta{label} is the label that uniquely identifies the term (as supplied in \gls{newglossaryentry} or \gls{newacronym}), \meta{insert} is additional text to insert after the term (but inside the hyperlink, if used with the \isty{hyperref} package), and \meta{options} is a \meta{key}=\meta{value} list of options. Available options are: \begin{itemize} \item \ikeyvalopt{gls}{format} This specifies how to format the associated location for this entry. It is analogous to the \gls{index.barchar} special character used in \gls{index} (see \sectionref{sec:makeindexformat}). As with \glsni{index}, the format must not include the initial backslash. For example, \booklinebreak\texttt{format=textbf} indicates that the location should be displayed in bold. (If you are using the \isty{hyperref} package, you should use the \texttt{hyper}\meta{xx} formats instead, such as \texttt{hyperbf}, see the \sty{glossaries} documentation~\cite{glossaries} for further detail.) \item \ikeyvalopt{gls}{counter} This specifies which counter to use for the associated location in the glossary. This is usually the page number, but can be changed to, say, the section in which the term is used. \item \ikeyvalopt{gls}{hyper} This is a~boolean key which can be used to enable/disable the hyperlink to the relevant entry in the glossary. Note that setting \texttt{hyper=true} will only have an effect if hyperlinks are supported (through loading the \isty{hyperref} package before loading the \isty{glossaries} package). The above commands all have starred versions that are a~shortcut for \texttt{hyper=false}. For example \texttt{\glsni{gls}*\marg{svm}} is equivalent to \texttt{\glsni{gls}\oarg{hyper=false}\marg{svm}}. \end{itemize} The above commands \glsni{gls} and \glsni{Gls} will display the value of the \ikeyvalopt{newglossaryentry}{first} or \ikeyvalopt{newglossaryentry}{text} key, depending on whether or not the entry has already been used. Similarly, \glsni{glspl} and \glsni{Glspl} will display the value of the \ikeyvalopt{newglossaryentry}{firstplural} or \ikeyvalopt{newglossaryentry}{plural} key, depending on whether or not the entry has already been used. The upper case forms, \glsni{Gls} and \glsni{Glspl}, will capitalise the first letter. \xminisec{Example:} Suppose I~have defined the following entry: \begin{code} \begin{alltt} \glsni{newglossaryentry}\marg{matrix}\glsni{percentchar} the label \marg{name=\marg{matrix},\glsni{percentchar} the term description=\marg{a rectangular table of elements},\glsni{percentchar} brief description plural=\marg{matrices}\glsni{percentchar} the plural } \end{alltt} \end{code} Then (later in the document) \begin{code} \gls{Glspl}\marg{matrix} are usually denoted by a~bold capital letter, such as \glsni{dollarchar}\glsni{mathbf}\marg{A}\glsni{dollarchar}. The \gls{gls}\marg{matrix}\oarg{'s} \glsni{dollarchar}(i,j)\glsni{dollarchar}th element is usually denoted \glsni{dollarchar}a\gls{underscorechar}\marg{ij}\glsni{dollarchar}. \gls{Gls}\marg{matrix} \glsni{dollarchar}\glsni{mathbf}\marg{I}\glsni{dollarchar} is the identity \gls{gls}\marg{matrix}. \end{code} will display: \begin{result}[matrix.html] Matrices are usually denoted by a~bold capital letter, such as $\mathbf{A}$. The matrix's $(i,j)$th element is usually denoted $a_{ij}$. Matrix $\mathbf{I}$ is the identity matrix. \end{result} If you have used the \ikeyvalopt{newglossaryentry}{symbol} key when you defined a~term, you can access its value with: \begin{definition} \gls{glssymbol}\oarg{\meta{options}}\marg{\meta{label}}\oarg{\meta{insert}} \end{definition} This has the same syntax as commands like \gls{gls} but it doesn't affect or query the first use flag. Terms that have been defined using \gls{newacronym} can also be referenced using the commands: \begin{definition} \gls{acrshort}\oarg{\meta{options}}\marg{\meta{label}}\oarg{\meta{insert}}\newline \gls{Acrshort}\oarg{\meta{options}}\marg{\meta{label}}\oarg{\meta{insert}} \end{definition} \begin{definition} \gls{acrlong}\oarg{\meta{options}}\marg{\meta{label}}\oarg{\meta{insert}}\newline \gls{Acrlong}\oarg{\meta{options}}\marg{\meta{label}}\oarg{\meta{insert}} \end{definition} \begin{definition} \gls{acrfull}\oarg{\meta{options}}\marg{\meta{label}}\oarg{\meta{insert}}\newline \gls{Acrfull}\oarg{\meta{options}}\marg{\meta{label}}\oarg{\meta{insert}} \end{definition} \emph{These commands don't affect the first use flag.} The first two (\glsni{acrshort} and \glsni{Acrshort}) will display the abbreviation only, the middle two (\glsni{acrlong} and \glsni{Acrlong}) will display the long form only, and the last two (\glsni{acrfull} and \glsni{Acrfull}) display both the long and short form. These commands have the same syntax as \glsni{gls} and \gls{Gls}. If you find these commands a~little long-winded to type, you can use the package option \istyopt{glossaries}{shortcuts}, which will provide shorter synonyms, such as \gls{acs}, \gls{acl} and \gls{acf}. This option also defines \gls{ac} which is equivalent to \gls{gls}. See the glossaries user guide~\cite{glossaries} for further details. \xminisec{Another Example:} Suppose I~have defined an acronym as follows: \begin{codeS} \glsni{newacronym}\marg{svm}\marg{SVM}\marg{support vector machine} \end{codeS} Then (later in the document): \begin{code} First use: \gls{gls}\marg{svm}\gls{at}. Next use: \gls{gls}\marg{svm}\glsni{at}. Short: \glsni{acrshort}\marg{svm}\glsni{at}. Long: \glsni{acrlong}\marg{svm}. Full: \glsni{acrfull}\marg{svm}\glsni{at}. \end{code} produces: \begin{result}[svm.html] First use: support vector machine (SVM)\@. Next use: SVM\@. Short: SVM\@. Long: support vector machine. Full: support vector machine (SVM)\@. \end{result} (Recall \glsni{at} from \novices{intersentencespacing}.) \xminisec{Note:} \warning Avoid using commands like \gls{gls} in section headings or captions. Instead, use commands like: \begin{definition} \gls{glsentrytext}\marg{\meta{label}} \end{definition} (displays the value of the \ikeyvalopt{newglossaryentry}{text} key without a~hyperlink) or \begin{definition} \gls{glsentryfirst}\marg{\meta{label}} \end{definition} (displays the value of the \ikeyvalopt{newglossaryentry}{first} key without a~hyperlink). These commands don't affect the first use flag. For related commands, see the glossaries user guide~\cite{glossaries}. \warning Take care if you want to use the uppercase variants, such as \gls{Gls} or \gls{Acrlong}. If the first letter is an accent (either entered using accents commands such as \texttt{\gls{acute}\marg{e}} or entered directly such as \texttt{\'{e}} with the \isty{inputenc} package) then you must group that letter when you define the term.\bookpagebreak \xminisec{Example:} \begin{code} \begin{alltt} \gls{newglossaryentry}\marg{elite}\glsni{percentchar} label \marg{\glsni{percentchar} name=\marg{\marg{\'{e}}lite},\glsni{percentchar} description=\marg{select group or class}\glsni{percentchar} } \end{alltt} \end{code} \setnode{newglossary} \subsubsection{Defining New Glossaries} \label{sec:newglossary} If you want the list of acronyms to be separate from the main glossary, you need to use the package option \istyopt{glossaries}{acronym}. This will change the effect of \gls{newacronym} so that it adds the term to the list of acronyms instead of to the main glossary. You can also define your own custom glossaries using \begin{definition} \gls{newglossary}\oarg{\meta{log-ext}}\marg{\meta{name}}\marg{\meta{in-ext}}\marg{\meta{out-ext}}\marg{\meta{title}}\oarg{\meta{counter}} \end{definition} where \meta{name} is a~label that uniquely defines this new glossary and \meta{title} is the title to be used when the glossary is displayed in the document via \glsi{printglossary} or \glsi{printglossaries}, see \sectionref{sec:printglossaries}. The other mandatory arguments, \meta{in-ext} and \meta{out-ext}, specify the file extensions to give to the input and output files for this new glossary. The first optional argument \meta{log-ext} is the extension for the log file. This information is provided for the benefit of the \iappname{makeglossaries} application. The final optional argument \meta{counter} is the name of the counter used by default in the location lists for this new glossary. If omitted, the \icounter{page} counter is used (unless overridden by the \istyopt{glossaries}{counter} package option). \xminisec{Note:} \warning All glossaries must be defined before \gls{makeglossaries} to ensure that the relevant output files are opened. \xminisec{Example:} The following defines a~new glossary called \dq{notation}: \begin{codeS} \glsni{newglossary}\oarg{nlg}\marg{notation}\marg{not}\marg{ntn}\marg{Notation} \end{codeS} When it gets displayed (using \glsni{printglossary} or \glsni{printglossaries}) the title will default to \dq{Notation}. I~now need to use the \ikeyvalopt{newglossaryentry}{type} key if I~want to define an entry to go in this new glossary: \begin{code} \begin{alltt} \gls{newglossaryentry}\marg{not:set}\glsni{percentchar} label \marg{\glsni{percentchar} type=notation,\glsni{percentchar} glossary type name=\marg{\glsni{dollarchar}\gls{mathcal}\marg{S}\glsni{dollarchar}},\glsni{percentchar} description=\marg{A set},\glsni{percentchar} sort=\marg{S}\glsni{percentchar} } \end{alltt} \end{code} Later in the document I~can use this entry: \begin{codeS} A \gls{gls}\marg{not:set} is a~collection of objects. \end{codeS} \setnode{printglossaries} \subsubsection{Displaying Glossaries} \label{sec:printglossaries} Now that you know how to define entries and how to use them in the document text, let's now look at the more complicated task of displaying the glossaries. To display all the defined glossaries use: \begin{definition} \gls{printglossaries} \end{definition} To only display a~particular glossary use: \begin{definition} \gls{printglossary}\oarg{\meta{options}} \end{definition} where \meta{options} is a~comma-separated list of \meta{key}=\meta{value} options. Available keys: \begin{itemize} \item \ikeyvalopt{printglossary}{type} The glossary to print. If omitted, the main (default) glossary is assumed. \item \ikeyvalopt{printglossary}{style} The glossary style to use. There are a~lot of predefined styles to choose from, such as \texttt{list}, \texttt{long} or \texttt{tree}. See the \sty{glossaries} user manual~\cite{glossaries} for further details. \item \ikeyvalopt{printglossary}{title} Overrides the default title for this glossary. \item \ikeyvalopt{printglossary}{toctitle} Overrides the default title for the table of contents. \item \ikeyvalopt{printglossary}{numberedsection} Put this glossary in a~numbered section (instead of an unnumbered section). \item \ikeyvalopt{printglossary}{nonumberlist} Suppress the location lists for this glossary. \end{itemize} \xminisec{Note:} \warning By default, the glossaries aren't added to the table of contents. If you want them added to the table of contents use the package option \istyopt{glossaries}{toc}.\bookpagebreak \begin{codeS} \glsni{usepackage}\oarg{toc}\marg{glossaries} \end{codeS} Only those entries that have been used in the document (via commands like \gls{gls}) are displayed in the glossary. If you want to add an entry without displaying it in the document, use \begin{definition} \gls{glsadd}\oarg{\meta{options}}\marg{\meta{label}} \end{definition} where \meta{label} is the unique label identifying the entry. The optional argument \meta{options} is the same as for commands like \glsni{gls} except there is no \ikeyvalopt{newglossaryentry}{hyper} key. Alternatively, you can add all defined entries using: \begin{definition} \gls{glsaddall}\oarg{\meta{options}} \end{definition} where \meta{options} is the same as for \glsni{glsadd} except that there is also a~\ikeyvalopt{glsaddall}{types} key where the value should be a~comma-separated list of all the glossaries to iterate over. For example, to add all entries defined in the \dq{acronym} glossary and the \dq{notation} glossary, but not the \dq{main} glossary: \begin{codeS} \glsni{glsaddall}\oarg{types=\marg{acronym,notation}} \end{codeS} \xminisec{Note:} \warning As with \gls{printindex} the glossaries won't be displayed until the relevant files have been created either by \iappname{makeindex} or by \iappname{xindy}. Unlike in \sectionref{sec:makeidx}, if you want to use \appname{xindy} to create your glossary files, you can't use the \iappname{texindy} wrapper but must either use \appname{xindy} directly or use the \iappname{makeglossaries} wrapper, described below. If you want to use \appname{xindy} with the \isty{glossaries} package, you must use the \istyopt{glossaries}{xindy} package option: \begin{codeS} \glsni{usepackage}\oarg{xindy}\marg{glossaries} \end{codeS} If omitted, \iappname{makeindex} will be assumed. If you have Perl installed, you can use the \iappname{makeglossaries} application that comes with the \isty{glossaries} package. If you have been using \iappname{latexmk} or \appname{xindy}, then you already have Perl installed. If you don't want to install Perl for some reason, there's a~Java alternative to \appname{makeglossaries} called \iappname{makeglossariesgui} that's available from \gls{ctan}. However, if you don't install Perl, you are restricting your options as you won't be able to use \appname{xindy}\footnote{or a~lot of other useful Perl scripts, such as \appname{epstopdf}}. If you are using \iappname{arara} (see \sectionref{sec:arara}), then all you need to do is add another \gls{percentchar.arara} directive in your source code: \begin{codeS} \glsni{percentchar.arara} \iappname{makeglossaries} \end{codeS} If you are using \iappname{latexmk}, then make sure you have added the custom dependencies for \texttt{.gls} as described in \sectionref{sec:latexmk}. If you are not using any automated tool to build your document, you will have to invoke \iappname{makeglossaries} between (PDF)\LaTeX\ runs (see \sectionref{sec:build}). Adding to \listingref{ex:thesis-index}: \begin{codelisting}{thesis-glossaries.tex}\label{ex:thesis-glossaries} \begin{alltt} \glsni{percentchar} In the preamble: \glsni{percentchar.arara} pdflatex: \marg{ synctex: on } \glsni{percentchar.arara} biber \glsni{percentchar.arara} \iappname{makeglossaries} \glsni{percentchar.arara} makeindex \glsni{percentchar.arara} pdflatex: \marg{ synctex: on } \glsni{percentchar.arara} pdflatex: \marg{ synctex: on } \glsni{documentclass}\oarg{oneside,12pt}\marg{scrbook} \glsni{usepackage}\oarg{toc,acronym}\marg{glossaries} \glsni{newglossary}\oarg{nlg}\marg{notation}\marg{not}\marg{ntn}\marg{Notation} \gls{makeglossaries} \glsni{newglossaryentry}\marg{matrix}\glsni{percentchar} the label \marg{name=\marg{matrix},\glsni{percentchar} the term description=\marg{a rectangular table of elements},\glsni{percentchar} brief description plural=\marg{matrices}\glsni{percentchar} the plural } \glsni{newacronym}\marg{svm}\marg{SVM}\marg{support vector machine} \gls{newglossaryentry}\marg{not:set}\glsni{percentchar} label \marg{\glsni{percentchar} type=notation,\glsni{percentchar} glossary type name=\marg{\glsni{dollarchar}\gls{mathcal}\marg{S}\glsni{dollarchar}},\glsni{percentchar} description=\marg{A set},\glsni{percentchar} sort=\marg{S}\glsni{percentchar} } \glsni{percentchar} Later in the document: \end{alltt} \gls{Glspl}\marg{matrix} are usually denoted by a bold capital letter, such as \glsni{dollarchar}\glsni{mathbf}\marg{A}\glsni{dollarchar}. The \gls{gls}\marg{matrix}\oarg{'s} \glsni{dollarchar}(i,j)\glsni{dollarchar}th element is usually denoted \glsni{dollarchar}a\gls{underscorechar}\marg{ij}\glsni{dollarchar}. \gls{Gls}\marg{matrix} \glsni{dollarchar}\glsni{mathbf}\marg{I}\glsni{dollarchar} is the identity \gls{gls}\marg{matrix}.\newline \mbox{}\newline First use: \gls{gls}\marg{svm}\gls{at}. Next use: \gls{gls}\marg{svm}\glsni{at}. Short: \gls{acrshort}\marg{svm}\glsni{at}. Long: \gls{acrlong}\marg{svm}. Full: \gls{acrfull}\marg{svm}\glsni{at}.\newline \mbox{}\newline A \gls{gls}\marg{not:set} is a collection of objects. \begin{alltt} \glsni{percentchar} At the end of the document: \glsni{backmatter} \glsni{printglossaries} \end{alltt} \end{codelisting} \setnode{glossariestroubleshooting} \subsubsection{Troubleshooting} \label{sec:glossariestroubleshooting} If you run into difficulties with the \isty{glossaries} package, first consult the \htmladdnormallinkfoot{glossaries FAQ}{http://www.dickimaw-books.com/faqs/glossariesfaq.html}\@. You can also check my \htmladdnormallinkfoot{bug tracker}{http://www.dickimaw-books.com/cgi-bin/bugtracker.cgi?category=glossaries} if you think you've stumbled on a~bug. If you are using TeXnicCenter instead of TeXworks, there are instructions on how to get TeXnicCenter to run \iappname{makeglossaries} in an article I~wrote on the \LaTeX\ Community's Know How section~\cite{latexcommunitygloss}. If you're completely confused about how to generate the glossary files, you might want to consider using \sty{datagidx} instead, described \htmlref{next}{sec:datagidx}. % Using datagidx \setnode{datagidx} \section{Using \LaTeX\ to Sort and Collate Indexes or Glossaries (\sty{datagidx} package)} \label{sec:datagidx} \sectionref{sec:makeindexglos} described how to create an index or glossaries using an external indexing application. Some users stumble when it comes to invoking the indexing application. There is an alternative where \TeX\ does the sorting and collating. This by-passes the need to use \iappname{makeindex}, \iappname{xindy} or \iappname{makeglossaries}, but it's less efficient and takes longer to build your document. This section describes how to do this using the \isty{datagidx} package. This package comes with my \isty{datatool} bundle (at least version 2.13). The documentation for \sty{datagidx} is included in the \sty{datatool} user manual~\cite{datatool}. The \sty{datatool} package allows you to define databases that you can access in your document. The \sty{datagidx} package has a~special interface to this facility that allows you to define databases for the purposes of indexing. These databases and their definitions must be defined in the preamble. In this section, the term \dq{indexing} will be used to refer to either indexes or glossaries, as the same mechanism is used for both tasks. A new indexing database is defined using: \begin{definition} \gls{newgidx}\marg{\meta{label}}\marg{\meta{title}} \end{definition} where \meta{label} is a~label that uniquely identifies this database and \meta{title} is the title to be used when the index (or glossary) is displayed. For example: \begin{codeS} \glsni{newgidx}\marg{index}\marg{Index} \end{codeS} creates a~new database labelled \texttt{index}. When the index is displayed, it will have the section heading \dq{Index}. As in \sectionref{sec:makeindexglos}, each term in the index (or glossary) database has an associated location list. This list is initially null. The locations are added to terms used in the document on \emph{the second} \LaTeX\ run. When you display the index, only those entries with a~non-null location list or a cross-reference will be shown. The default location is the page number on which the entry was referenced. The \sty{datagidx} package knows about the following page numbering styles: \texttt{arabic}, \texttt{roman}, \texttt{Roman}, \texttt{alph} and \texttt{Alph}. If your document has another type of numbering style, or if you want to use a~different counter for the location, consult the \sty{datagidx} section of the \sty{datatool} manual~\cite{datatool}. Once you have defined the indexing database, you can now define terms associated with that database using \begin{definition} \gls{newterm}\oarg{\meta{options}}\marg{\meta{name}} \end{definition} where \meta{name} is the term and \meta{options} is a~list of \meta{key}=\meta{value} options. The following keys are available: \begin{itemize} \item \ikeyvalopt{newterm}{database} Identifies the database in which to store this term. For example: \begin{definition} \glsni{newterm}\oarg{database=index}\marg{eigenvalue} \end{definition} It can be somewhat cumbersome having to type the database for each new term. Instead you can define the default database using: \begin{definition} \gls{DTLgidxSetDefaultDB}\marg{\meta{label}} \end{definition} For example: \begin{code} \begin{alltt} \glsni{newgidx}\marg{index}\marg{Index} \glsni{DTLgidxSetDefaultDB}\marg{index} \glsni{newterm}\marg{eigenvalue} \glsni{newterm}\marg{eigenvector} \end{alltt} \end{code} \item \ikeyvalopt{newterm}{label} A label uniquely identifying this term. If omitted the label is extracted from \meta{name}. \item \ikeyvalopt{newterm}{sort} The sort key. If omitted this is extracted from \meta{name}. \item \ikeyvalopt{newterm}{parent} The parent entry, if this is a~sub-term. (The value should be the label identifying the parent, which must already be defined.) \item \ikeyvalopt{newterm}{text} How the entry should appear in the document text. If omitted, \meta{name} is used. If present, \meta{name} indicates how the term should appear in the index\slash glossary. \item \ikeyvalopt{newterm}{description} An optional associated description. \item \ikeyvalopt{newterm}{plural} The plural form of this term. If omitted this value is obtained by appending \dq{s} to \meta{name} (or the value of \keyvalopt{text} if supplied). \item \ikeyvalopt{newterm}{symbol} An optional associated symbol. \item \ikeyvalopt{newterm}{short} An associated short form, if required. (Defaults to \meta{name} if omitted.) \item \ikeyvalopt{newterm}{long} An associated long form, if required. (Defaults to \meta{name} if omitted.) \item \ikeyvalopt{newterm}{shortplural} The plural of the associated short form. If omitted, the value is obtained by appending \dq{s} to the short form. \item \ikeyvalopt{newterm}{longplural} The plural of the associated long form. If omitted, the value is obtained by appending \dq{s} to the long form. \item \ikeyvalopt{newterm}{see} A cross-reference to a~synonym. The value should be the label of another entry. This entry will not have a~location list, just the reference to the other term. \item \ikeyvalopt{newterm}{seealso} A cross-reference to a~closely related term. Both this term and the cross-referenced term should have a~location list. \end{itemize} It's also possible to add your own custom keys. See the \sty{datagidx} section of the \sty{datatool} user guide~\cite{datatool} for further details. As with \gls{newglossaryentry}, discussed in \sectionref{sec:newglossaryentry}, if the term starts with an accented letter (or a~ligature) the letter must be grouped. \xminisec{Example:}\label{ex:elite} \begin{code} \begin{alltt} \glsni{newterm}\oarg{label=elite,sort=elite}\marg{\marg{\'{e}}lite} \glsni{newterm} \oarg{\glsni{percentchar} plural=\marg{\marg{\oe}sophagi}, label=\marg{oesophagus}, sort=\marg{oesophagus}, description=\marg{tube connecting throat and stomach} } \marg{\marg{\oe}sophagus} \end{alltt} \end{code} There is a~shortcut command for defining acronyms: \begin{definition} \gls{newacro}\oarg{\meta{options}}\marg{\meta{short}}\marg{\meta{long}} \end{definition} where \meta{short} is the abbreviation and \meta{long} is the long form. The optional argument \meta{options} is the same as for \glsni{newterm}. This is equivalent to: \begin{alltt} \glsni{newterm} \oarg{\glsni{percentchar} description=\marg{\glsni{capitalisewords}\marg{\meta{long}}},\glsni{percentchar} short=\marg{\glsni{acronymfont}\marg{\meta{short}}},\glsni{percentchar} long=\marg{\meta{long}},\glsni{percentchar} text=\marg{\glsni{DTLgidxAcrStyle}\marg{\meta{long}}\marg{\glsni{acronymfont}\marg{\meta{short}}}},\glsni{percentchar} plural=\marg{\glsni{DTLgidxAcrStyle}\marg{\meta{long}s}\marg{\glsni{acronymfont}\marg{\meta{short}s}}},\glsni{percentchar} sort=\marg{\meta{short}},\glsni{percentchar} \meta{options}\glsni{percentchar} }\glsni{percentchar} \marg{\glsni{MakeTextUppercase}\marg{\meta{short}}} \end{alltt} where \begin{definition} \gls{DTLgidxAcrStyle}\marg{\meta{long}}\marg{\meta{short}} \end{definition} formats the full version of the acronym. This defaults to: \meta{long} (\meta{short}), and \begin{definition} \gls{acronymfont}\marg{\meta{text}} \end{definition} is the font used to format acronyms. By default this just displays its argument, but can be redefined if you want the acronyms formatted in a~particular style or font (such as small-caps). The other commands used above are: \begin{definition} \gls{MakeTextUppercase}\marg{\meta{text}} \end{definition} This is defined by the \isty{textcase} package and converts \meta{text} to uppercase. \begin{definition} \gls{capitalisewords}\marg{\meta{text}} \end{definition} This is defined by the \isty{mfirstuc} package and capitalises the first letter of each word in \meta{text}. \xminisec{Example:}\label{ex:svm} \begin{codeS} \glsni{newacro}\marg{svm}\marg{support vector machine} \end{codeS} Once you have defined the terms in the preamble, you can later use them in the document: \begin{definition} \gls{datagidx.gls}\marg{\oarg{\meta{format}}\meta{label}} \end{definition} \begin{definition} \gls{datagidx.glspl}\marg{\oarg{\meta{format}}\meta{label}} \end{definition} \begin{definition} \gls{datagidx.Gls}\marg{\oarg{\meta{format}}\meta{label}} \end{definition} \begin{definition} \gls{datagidx.Glspl}\marg{\oarg{\meta{format}}\meta{label}} \end{definition} These are similar to those described in \sectionref{sec:glslink}, but they have a~different syntax. Here \meta{format} is the name of a~text-block commands (such as \gls{textbf}) \emph{without} the initial backslash that should be used to format the location for this reference. This is analogous to the \gls{index.barchar} special character described in \sectionref{sec:makeindexformat}. There are also commands associated with acronyms: \begin{definition} \gls{datagidx.acr}\marg{\oarg{\meta{format}}\meta{label}} \end{definition} \begin{definition} \gls{datagidx.acrpl}\marg{\oarg{\meta{format}}\meta{label}} \end{definition} \begin{definition} \gls{datagidx.Acr}\marg{\oarg{\meta{format}}\meta{label}} \end{definition} \begin{definition} \gls{datagidx.Acrpl}\marg{\oarg{\meta{format}}\meta{label}} \end{definition} \warning Unlike the \isty{glossaries} package, described in \sectionref{sec:makeglossaries}, there is a~difference between \isty{datagidx}'s \glsni{datagidx.gls} and \glsni{datagidx.acr}. Here \glsni{datagidx.gls} will always display the value of the \ikeyvalopt{newterm}{text} field, whereas \glsni{datagidx.acr} will display the full form on first use (the \ikeyvalopt{newterm}{text} field) and the abbreviation on subsequent use (the \ikeyvalopt{newterm}{short} field). You can also add terms to the index without creating any link text: \begin{definition} \gls{datagidx.glsadd}\marg{\meta{label}} \end{definition} This adds the term uniquely identified by \meta{label}. \begin{definition} \gls{datagidx.glsaddall}\marg{\meta{database name}} \end{definition} This adds all the terms defined in the database uniquely identified by \meta{database name}. \xminisec{Note:} \warning Unlike most commands, the optional part of the above commands occurs \emph{inside} the mandatory argument. \xminisec{Examples:} Given the \texttt{elite} and \texttt{oesophagus} examples defined \xpageref{earlier}{ex:elite}, I~can reference those entries in the text as follows: \begin{codeS} \gls{datagidx.Gls}\marg{elite} and \gls{datagidx.glspl}\marg{oesophagus}. \end{codeS} This produces: \begin{resultS}[Elite and oesophagi] \'{E}lite and \oe sophagi. \end{resultS} Elsewhere, I~might have the main topic about \oe sophagi: \begin{code} The \gls{datagidx.gls}\marg{\oarg{textbf}oesophagus} connects the throat and the stomach. \end{code} This produces: \begin{resultS}[The oesophagus connects the throat and the stomach.] The \oe sophagus connects the throat and the stomach. \end{resultS} and the associated location will be typeset in bold. Here's an example using the \texttt{svm} example defined \xpageref{earlier}{ex:svm}: \begin{code} First use: \gls{datagidx.acr}\marg{svm}\gls{at}. Subsequent use: \gls{datagidx.acr}\marg{svm}\glsni{at}. Full form: \gls{datagidx.gls}\marg{svm}. \end{code} This produces: \begin{result}[datagidx-svm.html] First use: support vector machine (SVM)\@. Subsequent use: SVM\@. Full form: support vector machine (SVM). \end{result} You can unset and reset acronyms using \begin{definition} \gls{glsunset}\marg{\meta{label}} \end{definition} and \begin{definition} \gls{glsreset}\marg{\meta{label}} \end{definition} To display the index or glossary or list of acronyms use: \begin{definition} \gls{printterms}\oarg{\meta{options}} \end{definition} where \meta{options} is a~comma-separated \meta{key}=\meta{value} list. Common options are: \begin{itemize} \item \ikeyvalopt{printterms}{database} The label uniquely identifying the database containing the relevant terms. \item \ikeyvalopt{printterms}{postdesc} This may have the value \optfmt{dot} (put a~full stop after the description, if there is a~description) or \optfmt{none} (don't put a~full stop after the description). \item \ikeyvalopt{printterms}{columns} This value must be an integer greater than or equal to 1, indicating the number of columns for the page layout. \item \ikeyvalopt{printterms}{style} The style to use. There are a~number of predefined styles, such as \texttt{index} or \texttt{gloss}. See the user guide~\cite{datatool} for further details. \item \ikeyvalopt{printterms}{namecase} Indicates whether any case change should be applied to the entry's name. Available values are: \optfmt{nochange} (no change), \optfmt{uc} (convert to uppercase), \optfmt{lc} (convert to lower case), \optfmt{firstuc} (convert the first letter to uppercase) and \optfmt{capitalise} (capitalise each initial letter using \gls{capitalisewords}). \end{itemize} For a~full list of options see the \sty{datagidx} section of the \sty{datatool} user guide~\cite{datatool}. \listingref{ex:thesis-glossaries} can now be rewritten as follows: \begin{codelisting}{thesis-datagidx.tex} \begin{alltt} \glsni{percentchar.arara} pdflatex: \marg{ synctex: on } \glsni{percentchar.arara} biber \glsni{percentchar.arara} pdflatex: \marg{ synctex: on } \glsni{percentchar.arara} pdflatex: \marg{ synctex: on } \glsni{documentclass}\oarg{oneside,12pt}\marg{scrbook} \glsni{usepackage}\marg{datagidx} \glsni{newgidx}\marg{index}\marg{Index} \glsni{newgidx}\marg{glossary}\marg{Glossary} \glsni{newgidx}\marg{acronym}\marg{Acronyms} \glsni{newgidx}\marg{notation}\marg{Notation} \glsni{DTLgidxSetDefaultDB}\marg{glossary} \glsni{newterm} \oarg{\glsni{percentchar} description=\marg{a rectangular table of elements},\glsni{percentchar} brief description plural=\marg{matrices}\glsni{percentchar} the plural }% \marg{matrix}% the name \glsni{DTLgidxSetDefaultDB}\marg{acronym} \glsni{newacro}\marg{svm}\marg{support vector machine} \glsni{DTLgidxSetDefaultDB}\marg{notation} \glsni{newterm} \oarg{\glsni{percentchar} label=\marg{not:set},\glsni{percentchar} label description=\marg{A set},\glsni{percentchar} sort=\marg{S}\glsni{percentchar} }\glsni{percentchar} \marg{\gls{ensuremath}\marg{\gls{mathcal}\marg{S}}} \glsni{DTLgidxSetDefaultDB}\marg{index} \glsni{newterm} \oarg{\glsni{percentchar} label=\marg{function},\glsni{percentchar} text=\marg{function}\glsni{percentchar} }\glsni{percentchar} \marg{functions} \glsni{newterm} \oarg{\glsni{percentchar} see=\marg{sqrt},\glsni{percentchar} }\glsni{percentchar} \marg{square root} \glsni{newterm} \oarg{\glsni{percentchar} label=\marg{fn.sqrt}, parent=\marg{function} }\glsni{percentchar} \marg{\glsni{texttt}\marg{sqrt()}} \glsni{newterm} \oarg{\glsni{percentchar} label=\marg{sqrt}, }\glsni{percentchar} \marg{\texttt{sqrt()}} \glsni{newterm}\marg{tautology} \glsni{newterm}\marg{contradiction} \glsni{percentchar} later in the document: \end{alltt} \glsni{datagidx.Glspl}\marg{matrix} are usually denoted by a~bold capital letter, such as \glsni{dollarchar}\glsni{mathbf}\marg{A}\glsni{dollarchar}. The \glsni{datagidx.gls}\marg{matrix}'s \glsni{dollarchar}(i,j)\glsni{dollarchar}th element is usually denoted \glsni{dollarchar}a\glsni{underscorechar}\marg{ij}\glsni{dollarchar}. \glsni{datagidx.Gls}\marg{matrix} \glsni{dollarchar}\glsni{mathbf}\marg{I}\glsni{dollarchar} is the identity \glsni{datagidx.gls}\marg{matrix}.\newline \mbox{}\newline First use: \gls{datagidx.acr}\marg{svm}\glsni{at}. Next use: \gls{datagidx.acr}\marg{svm}\glsni{at}. Full: \gls{datagidx.gls}\marg{svm}\glsni{at}.\newline \mbox{}\newline A \gls{datagidx.gls}\marg{not:set} is a~collection of objects.\newline \mbox{}\newline \ldots\newline \mbox{}\newline Some sample code is shown in Listing\glsni{tildechar}\glsni{ref}\marg{lst:sample}. This uses the function \gls{datagidx.gls}\marg{fn.sqrt}.\gls{datagidx.glsadd}\marg{sqrt}\newline \mbox{}\newline \ldots\newline \mbox{}\newline \glsni{begin}\marg{Definition}\oarg{Tautology}\newline A \glsni{emph}\marg{\glsni{datagidx.gls}\marg{\oarg{textbf}tautology}} is a~proposition that is always true for any value of its variables.\newline \glsni{end}\marg{Definition}\newline \mbox{}\newline \glsni{begin}\marg{Definition}\oarg{Contradiction}\newline A \glsni{emph}\marg{\gls{datagidx.gls}\marg{\oarg{textbf}contradiction}} is a proposition that is always false for any value of its variables.\newline \glsni{end}\marg{Definition} \begin{alltt} \glsni{percentchar} At the end of the document: \glsni{backmatter} \glsni{printterms}\oarg{database=glossary} \glsni{printterms}\oarg{database=acronym} \glsni{printterms}\oarg{database=notation} \glsni{printbibliography} \glsni{printterms}\oarg{database=index} \end{alltt} \end{codelisting} Note that there is now no need to call either \iappname{makeindex} or \iappname{makeglossaries}. The only external application being called is \iappname{biber} for the bibliography. \appendix \setnode{generaladvice} \chapter{General Advice} \label{ch:generaladvice} If you encounter any \LaTeX\ problems, check \latexhtml{Appendix~\ref{nov-ch:errors} (Common Errors) and Appendix~\ref{nov-ch:help} (Need More Help?) in \latexnovices.}{% \xrsectionref{Common Errors}{commonerrors} and \xrsectionref{Need More Help?}{help}} \setnode{toomanyunprocessedfloats} \section{Too Many Unprocessed Floats} \label{sec:toomanyunprocessedfloats} A common problem PhD student's encounter when writing a~thesis is the \dq{too many unprocessed floats} error.\faq{Too many unprocessed floats}{tmupfl} This is usually caused by having too many figures and tables in the results chapter and not enough surrounding text. If this happens, there are a~number of things you can try doing: \begin{enumerate} \item Make sure you haven't been too restrictive in where you want your floats to go. If you use a~placement specifier, give LaTeX as many options as possible. For example: \begin{codeS} \glsni{begin}\marg{figure}\oarg{htbp} \end{codeS} which indicates that the figure can be placed \dq{here} (\texttt{h}), at the top of a~page (\texttt{t}), at the bottom of the page (\texttt{b}) or on a~page solely consisting of floats (\texttt{p}). If you just use the \texttt{h} placement specifier then you are stating: \dq{I~want it \emph{here} and \emph{nowhere else}!} If \TeX\ can't put it \emph{exactly here}, then you have given no alternative place to put it, and it won't get placed anywhere, unless a~\gls{clearpage} command is issued, at which point all remaining unprocessed floats will be dumped at that point. If you are determined that an image must be placed \emph{exactly here} then it should not be placed in a~floating environment. \item Try increasing the amount of text in the chapter. Remember that you should never simply print all the figures and tables in a results chapter without discussing them to some extent. \item If all else fails, try using the \glsni{clearpage} command. This forces all unprocessed floats to be processed immediately, and start a~new page. This may result in the page ending prematurely, if you wish to avoid this, you can use the \isty{afterpage} package, and use the command: \begin{codeS} \gls{afterpage}\marg{\gls{clearpage}} \end{codeS} \end{enumerate} For other problems, check the FAQ~\cite{ukfaq}. \setnode{thesiswritingadvice} \section{General Thesis Writing Advice} \label{sec:thesiswritingadvice} This section is not specific to \LaTeX. Some of the points have already been mentioned in asides or footnotes. Remember that each college or university or even school within a~university may have different requirements, and requirements will also vary according to country, so some of this advice may not apply to you. I~am writing from the point of view of an English scientist, and am basing it on my own experience and on the comments of English science-based PhD examiners and supervisors. I~cannot guarantee that your own department or university will agree with them. \emph{If in doubt, check with your supervisor.} \begin{enumerate} \item Find out the thesis style requirements from your supervisor or your department's website. Many universities still require \htmlref{double-spaced}{sec:setspace}, single-sided documents with \htmlref{wide margins}{geometry}. Double-spacing is by and large looked down on in the world of typesetting, but this requirement for a~PhD thesis has nothing to do with \ae sthetics or readability. In England the purpose of the PhD viva is to defend your work\footnote{I~gather this is not the case in some other countries, where the viva is more informal, and the decision to pass or fail you has already been made before your viva.}. Before your viva, paper copies of your thesis are sent to your examiners. The double spacing and wide margins provide the examiners room to write the comments and criticisms they wish to raise during the viva, as well as any typographical corrections. Whilst they could write these comments on a~separate piece of paper, cross-referencing the page in the thesis, it is more efficient for the comments to actually be on the relevant page of the thesis. That way, as they go through the manuscript during your viva, they can easily see the comments, questions or criticisms they wish to raise alongside the corresponding text. If you present them with a~single-spaced document with narrow margins, you are effectively telling them that you don't want them to criticise your work! \item Don't try to pad your thesis with irrelevant information. This includes adding items in your bibliography that are not referenced in the text, adding figures or tables that are not explained in the text, and supplying all the source code you have written. The outcome of your viva will not depend on the physical size of your thesis, but on the clarity of your writing and on the quality of your work. \item Clearly delineate your thesis through the use of chapters and sections, outlining your original aims and objectives, an overview of the subject matter including references to other people's work in the area, the methods you employed to extend or innovate the field, your results and conclusions. \item Make sure your references include some recent journal or conference papers to illustrate that you are aware of new developments in your field. Remember that due to the nature of publishing, most books are dated by the time they reach the book shelves. Journal and conference papers are likely to be more up-to-date\footnote{Having said that, I~know someone who submitted an article to a~journal, and it took three and a~half years before the reviewers came back with comments. In the end, the author withdrew the manuscript because by that time the topic was out of date.}. \item Always explain acronyms, technical terms and symbols. It is a good idea to include a~glossary of terms, list of notation or list of acronyms to avoid confusion (see \chapterref{ch:indgloss}). \item If you have equations, make sure you explain the variables used, and how you go from one equation to the next. Depending on your field, you might also consider clarifying the mathematics by providing graphical representations of the equations\footnote{When I was a~PhD student, I~was once rendered speechless when asked to provide a~graphical illustration of an equation involving a quadruple summation that had no graphical meaning from my point of view. Perhaps this was a~drawback of being a~mathematician doing a PhD in an electronics department.}. \item If you include any graphs, bar charts, pie charts or any other form of data plot, make sure it is clearly labelled and no distortion is introduced (such as using three-dimensional bar charts or pie charts\footnote{The sole purpose of 3D pie charts or bar charts appears to be to look pretty and impress people who have no understanding of mathematics.}.) \item If you have used a~computer application to generate numerical results, make sure you have some understanding of the underlying process and what the results mean. This doesn't necessarily mean that you need to understand complex computer code, or complex algorithms, but what you shouldn't do is say something along the lines of, \dq{well, I~clicked on this button, and it said $m=0.678$.} What is the purpose of the button? What does $m$ represent? What does the result $ m=0.678$ signify? What value were you expecting or hoping to get? Numbers on their own are meaningless. If I~ran into a room shouting \dq{I've got 42!} What does that mean? Forty-two what? Forty-two brilliant reviews? (Great!) Forty-two percent in an exam? (Not good.) Forty-two spots on my face? (Very bad!) \item\label{itm:gcc1} Don't waste time worrying about the best way to word your thesis in your first draft. Write first, then edit it later or you will never get started. \item\label{itm:gcc2} If your supervisor offers to critique chapters of your thesis, don't say no! Such offers are not made out of politeness, but a~desire to ensure that you pass. Don't be embarrassed and worry that it's not good enough, that's the whole point in your supervisor helping you improve it\footnote{but don't expect your supervisor to actually write your thesis!}. \item Write in a~clear concise manner. A thesis is a~technical document, not a~novel, so don't be tempted to write something along the lines of: \dq{I~awaited with bated breath, my whole body quivering with excitement at the eager anticipation that my algorithm would prove superior to all others, and, oh joy, my experiments proved me right.} \item Don't decorate your thesis with irrelevant clip art. It is unprofessional and highly inappropriate in the sciences. \item Make regular backups of your work. Be prepared for any of the following: accidentally deleting your thesis, accidentally overwriting your thesis with another file, software failure, hardware failure, viruses, fire and theft. Consider using at least a~two-tier system where you keep one backup in a safe place where you live and ask a close relative or friend to take care of another backup. \end{enumerate} Items~\ref{itm:gcc1} and~\ref{itm:gcc2} above were supplied by Dr~Gavin Cawley\footnote{School of Computing Sciences, University of East Anglia} who has been both a~PhD supervisor and examiner. \backmatter \clearpage\phantomsection \setnode{bibliography} \bibliographystyle{plain} \bibliography{thesis} \setnode{acronyms} \printglossary[type=acronym,nonumberlist,style=acronyms] \setnode{summary} \setdoublecolumnglo \printglossary[style=summary] \setnode{bookindex} \printindex \setnode{licence} \chapter{GNU Free Documentation License} \input{../fdl} \setnode{history} \chapter{History}\markright{History} \historyitem{\protect\bookdocdate\ (Version 1.3)} \begin{itemize} \item Added recap on building the document. \item Added sections on \appname{latexmk} and \appname{arara}. \item Changed examples to use KOMA. \item Added sections on \appname{jabref}, \sty{natbib} and \sty{biblatex}. \item Added information about the \sty{listings}, \sty{siunitx}, \sty{amsthm}, \sty{ntheorem} and \sty{algorithm2e} packages. \item Added section on \sty{datagidx} to the chapter on indexes and glossaries. \item Added summary section. \item Some sections have been reordered. \item Removed section on modifying textual tags such as \cmdname{contentsname} (now in Volume~1). \end{itemize} \backcoverheading (See \url{http://www.gnu.org/licenses/fdl-howto-opt.html#SEC2}.) \backcovertext \end{document}