\ProvidesFile{dccpaper-base.doc}[2015/06/23 automatically generated with makedoc.sty] \begin{lstlisting}[firstnumber=1] %% %% This is file `dccpaper-base.tex', %% generated with the docstrip utility. %% %% The original source files were: %% %% dccpaper.dtx (with options: `base') %% %% ---------------------------------------------------------------- %% The dccpaper bundle: Classes for submissions to IJDC and IDCC %% Author: Alex Ball %% E-mail: a.ball@ukoln.ac.uk %% License: Released under the LaTeX Project Public License v1.3c or later %% See: http://www.latex-project.org/lppl.txt %% ---------------------------------------------------------------- %% \def\Version{2015/06/22 v1.4.1} \ProvidesFile{dccpaper-base.tex} [\Version\space Common class code for IJDC and IDCC papers.] \end{lstlisting} The \textsf{dccpaper} classes are deliberately very similar. This file contains the common code. All the classes are based on the \textsf{article} class, and use A4 paper. \begin{lstlisting}[firstnumber=last] \LoadClass[a4paper,12pt,twoside]{article} \end{lstlisting} We use British English orthography. \begin{lstlisting}[firstnumber=last] \RequirePackage[british]{babel} \end{lstlisting} The macro patching commands from \textsf{etoolbox} come in useful for handling author and date information, and also for compatibility with \textsf{apacite}. \begin{lstlisting}[firstnumber=last] \RequirePackage{etoolbox} \end{lstlisting} The \textsf{dccpaper} classes use Times as the main text font. We prefer \textsf{newtx} as it provides support for mathematics, but the standard \textsf{times} package will do. In case they are needed, we also provide Helvetica for the sans serif font and Computer Modern Teletype for the monospaced. \begin{lstlisting}[firstnumber=last] \RequirePackage[T1]{fontenc} \IfFileExists{newtxtext.sty}% {\RequirePackage{newtxtext,newtxmath}}% {\RequirePackage{times}} \RequirePackage[scaled=0.92]{helvet} \renewcommand{\ttdefault}{cmtt} \end{lstlisting} We will need support for included graphics and colour. The structural elements are a medium turquoise, while the links are slightly darker to make them easier to read on screen. \begin{lstlisting}[firstnumber=last] \RequirePackage{graphicx} \IfFileExists{xcolor.sty}% {\RequirePackage{xcolor}}% {\RequirePackage{color}} \definecolor[named]{struct}{rgb}{0,0.5,0.5} \definecolor[named]{links}{rgb}{0,0.4,0.4} \end{lstlisting} We will calculate some lengths later. \begin{lstlisting}[firstnumber=last] \RequirePackage{calc} \end{lstlisting} Ragged right text is easier to read on screen, but fully justified text looks better. The |\raggedyright| layout from Peter Wilson's \textsf{memoir} class (2005/09/25 v1.618) is an excellent compromise. The code below replicates it. First we save the original definitions of |\\| and |\parindent| as |\OrigLineBreak| and |\RaggedParindent| respectively. \begin{lstlisting}[firstnumber=last] \let\OrigLineBreak\\ \newdimen\RaggedParindent \setlength{\RaggedParindent}{\parindent} \end{lstlisting} The |\raggedyright| layout more or less lays text out as with full justification, but then lets the shorter lines relax a bit from the right margin. It is the default for DCC papers. \begin{lstlisting}[firstnumber=last] \newcommand{\raggedyright}[1][2em]{% \let\\\@centercr\@rightskip \z@ \@plus #1\relax \rightskip\@rightskip \leftskip\z@skip \parindent\RaggedParindent} \AtBeginDocument{\raggedyright} \end{lstlisting} The |\flushleftright| layout restores full justification, in case it is needed. \begin{lstlisting}[firstnumber=last] \newcommand*{\flushleftright}{% \let\\\OrigLineBreak \leftskip\z@skip \rightskip\leftskip \parfillskip\@flushglue \everypar{}} \end{lstlisting} Pdf\TeX\ 1.40.15 introduced the facility to include dummy interword spaces to improve text extraction and reflow. We test for this facility and use it if available. \begin{lstlisting}[firstnumber=last] \RequirePackage{ifpdf} \ifpdf \ifdef{\pdfinterwordspaceon}{\pdfinterwordspaceon}{} \fi \end{lstlisting} The classes have some special metadata requirements. We start with the author information. The macro |\thecorrespondence| is used in the title page footer for the name, postal address and email address of the corresponding author. \begin{lstlisting}[firstnumber=last] \def\thecorrespondence{} \newcommand*{\correspondence}[1]{\def\thecorrespondence{#1}} \end{lstlisting} The handling of authors here is inspired by Patrick W Daly's \textsf{authblk}, (2001/02/27 1.3), and defines the familiar user commands. Authors are presented in blocks, one affiliation but perhaps several authors per block. We make the presentation of the author information configurable (just in case) with some hooks and lengths: \begin{itemize} \item |\Authfont| is the font used for author names; \item |\Affilfont| is the font used for affiliations; \item |\affilsep| is the line spacing between author names and affiliations; \item |\authblksep| is the line spacing between author name/affiliation blocks. \end{itemize} \begin{lstlisting}[firstnumber=last] \newcommand*{\Authfont}{\normalfont} \newcommand*{\Affilfont}{\normalfont\small} \newlength{\affilsep}\setlength{\affilsep}{0pt} \newlength{\authblksep}\setlength{\authblksep}{1.5\baselineskip} \end{lstlisting} The |\AuthorBlock| command will be used to collect all the author information. \begin{lstlisting}[firstnumber=last] \newcommand{\AuthorBlock}{} \end{lstlisting} The |\MainAuthor| command collects abbreviated author information for use in the headers. \begin{lstlisting}[firstnumber=last] \newcommand{\MainAuthor}{} \end{lstlisting} The |\dccp@author| command, meanwhile, collects a full list of authors for the PDF metatdata. \begin{lstlisting}[firstnumber=last] \newcommand{\dccp@author}{} \end{lstlisting} We define counters for \begin{itemize} \item the total number of authors defined; \item the number of authors in the current block; \item the number of blocks defined. \end{itemize} \begin{lstlisting}[firstnumber=last] \newcounter{authors} \newcounter{authorsinblock} \newcounter{block} \end{lstlisting} The |block| counter will increase once in the preamble, and again when the information is typeset, so we need to reset it at the beginning of the document. \begin{lstlisting}[firstnumber=last] \AtBeginDocument{\setcounter{block}{0}} \end{lstlisting} The new definition of the author command starts here. \begin{lstlisting}[firstnumber=last] \renewcommand{\author}[1]{% \end{lstlisting} If this is the first or second |\author| command, we add the name to our abbreviated list of authors. Otherwise, we replace the name of the second and subsequent authors with `et al.' in that list. At the same time, we use a simpler technique to populate |\dccp@author|. \begin{lstlisting}[firstnumber=last] \ifnum\theauthors=0 \def\dccp@author{#1}% \def\MainAuthor{#1}% \else \appto\dccp@author{, #1}% \ifnum\theauthors=1% \def\OtherMainAuthors{ and #1}% \appto\MainAuthor{\OtherMainAuthors}% \else \ifnum\theauthors=2% \def\OtherMainAuthors{ et al.}% \fi \fi \fi \stepcounter{authors}% \end{lstlisting} Each block has its author names collected in a macro like |\blocki@auth|, and its affiliation collected in a macro like |\blocki@affil| (the `i' is a serial number). If this is the first author in a block, we need to create the block and add it to |\AuthorBlock|; second and subsequent blocks are preceded by a |\quad| of space. \begin{lstlisting}[firstnumber=last] \ifnum\theauthorsinblock=0% \stepcounter{block}% \expandafter\def\csname block\roman{block}@auth\endcsname{\Authfont#1}% \ifnum\theblock>1\appto\AuthorBlock{\quad}\fi \appto\AuthorBlock{% \stepcounter{block}% \begin{minipage}[t]{0.45\textwidth}\centering \csname block\roman{block}@auth\endcsname \ifx\undefined\csname block\roman{block}@affil\endcsname \else \\[\affilsep]\csname block\roman{block}@affil\endcsname \fi \end{minipage}% \rule{0pt}{2\baselineskip}% } \else \end{lstlisting} Otherwise we just add the name to the right |\blocki@auth|-style macro. \begin{lstlisting}[firstnumber=last] \csappto{block\roman{block}@auth}{, #1}% \fi \stepcounter{authorsinblock}% } \end{lstlisting} The |\affil| command adds an affiliation to the current block and closes it by resetting the |authorsinblock| counter. \begin{lstlisting}[firstnumber=last] \newcommand{\affil}[1]{% \expandafter\def\csname block\roman{block}@affil\endcsname{\Affilfont#1}% \setcounter{authorsinblock}{0}% } \end{lstlisting} The |\HeadTitle| collects the abbreviated title for use in the headers. \begin{lstlisting}[firstnumber=last] \newcommand*{\HeadTitle}{} \end{lstlisting} We wrap the normal |\title| command with code to populate |\HeadTitle| with the optional argument if provided, or the mandatory one otherwise. We also provide a persistent |\thetitle| macro, stripped of any |\thanks|. \begin{lstlisting}[firstnumber=last] \let\ProperTitle=\title \renewcommand{\title}[2][\empty]{ \ifx\empty #1% \renewcommand*{\HeadTitle}{#2}% \else \renewcommand*{\HeadTitle}{#1}% \fi% \begingroup\let\footnote\@gobble \ProperTitle{#2}% \begingroup \renewcommand{\thanks}[1]{} \protected@xdef\thetitle{#2} \endgroup\endgroup } \end{lstlisting} To prevent \LaTeX\ throwing an error if no title is supplied, we supply a fallback value and issue a warning instead. \begin{lstlisting}[firstnumber=last] \def\thetitle{% Please supply a title \ClassWarning{dccpaper}{% Please supply a title in the preamble with \noexpand\title{}% } } \end{lstlisting} We make |\thedate| persistent, borrowing the technique used in Peter Wilson's \textsf{memoir} class (2005/09/25 v1.618). \begin{lstlisting}[firstnumber=last] \pretocmd{\date}{\begingroup\let\footnote\@gobble}{}{}% \apptocmd{\date}{% \begingroup \renewcommand{\thanks}[1]{} \protected@xdef\thedate{#1} \endgroup\endgroup% }{}{} \end{lstlisting} We ensure |\thedate| is defined using a dummy date. \begin{lstlisting}[firstnumber=last] \date{20xx} \end{lstlisting} IJDC articles have extra bibliographic information: \begin{itemize} \item |\volume| sets the volume number, |\thevolume|; \item |\issue| sets the issue number, |\theissue|; \item |\subno| sets the submission number, |\thesubno|. \end{itemize} These numbers are used to build the DOI, |\thedoi|. \begin{lstlisting}[firstnumber=last] \newcommand*{\thevolume}{0} \newcommand*{\volume}[1]{\renewcommand*{\thevolume}{#1}} \newcommand*{\theissue}{0} \newcommand*{\issue}[1]{\renewcommand*{\theissue}{#1}} \newcommand*{\thesubno}{0} \newcommand*{\subno}[1]{\renewcommand*{\thesubno}{#1}} \newcommand*{\thedoi}{10.2218/ijdc.v\thevolume i\theissue .\thesubno} \end{lstlisting} They also display the page range. The following code was borrowed from Peter Wilson's \textsf{memoir} class (2005/09/25 v1.618). It defines a counter |lastpage| which, on the second run, will contain the number of the last page. \begin{lstlisting}[firstnumber=last] \newcounter{lastpage} \setcounter{lastpage}{0} \newcommand{\dol@stpage}{% \if@filesw \addtocounter{page}{-1}% \immediate\write\@auxout% {\string\setcounter{lastpage}{\the\c@page}}% \stepcounter{page}% \fi } \AtBeginDocument{\AtEndDocument{\clearpage\dol@stpage}} \end{lstlisting} DCC papers display some important dates. We collect these in |\dccp@dates|, initially setting the value to something sensible for papers in draft. \begin{lstlisting}[firstnumber=last] \def\dccp@dates{\emph{Draft from} \today} \end{lstlisting} Several types of date can be added: \begin{itemize} \item |\submitted| for when the authors submitted the paper (intended for IDCC papers). \item |\received| for when the paper was received by the editorial board (intended for IJDC papers). \item |\revised| for when the most recent version was received by the editorial board. \item |\accepted| for when the paper was accepted by the editorial board. \end{itemize} \begin{lstlisting}[firstnumber=last] \newcommand*{\submitted}[1]{% \def\dccp@dates{\emph{Submitted} #1}} \newcommand*{\received}[1]{% \def\dccp@dates{\emph{Received} #1}} \newcommand*{\revised}[1]{% \appto\dccp@dates{% \space\space\space\textbar\space\space\space \emph{Revision received} #1}% } \newcommand*{\accepted}[1]{% \appto\dccp@dates{% \space\space\space\textbar\space\space\space \emph{Accepted} #1}% } \end{lstlisting} IJDC papers need to say if they had a previous life as a conference paper. This statement goes in |\dccp@conf|, which is initially empty. The user command for setting this text is |\conference|. \begin{lstlisting}[firstnumber=last] \let\dccp@conf=\empty \newcommand*{\conference}[1]{% \renewcommand*{\dccp@conf}{An earlier version of this paper was presented at #1.}% } \end{lstlisting} The macro |\FixTextHeight| will be useful when switching from the first page geometry to the regular geometry for the rest of the paper. It is based on code from Hideo Umeki's \textsf{geometry} package (2002/07/08 v3.2). \begin{lstlisting}[firstnumber=last] \newcommand{\FixTextHeight}{% \setlength\@tempdima{\textheight}% \addtolength\@tempdima{-\topskip}% \@tempcnta\@tempdima \@tempcntb\baselineskip \divide\@tempcnta\@tempcntb \setlength\@tempdimb{\baselineskip}% \multiply\@tempdimb\@tempcnta \advance\@tempdima-\@tempdimb \global\advance\footskip\@tempdima \multiply\@tempdima\tw@ \ifdim\@tempdima>\baselineskip \addtolength\@tempdimb{\baselineskip}% \global\advance\footskip-\baselineskip \fi \addtolength\@tempdimb{\topskip}% \global\textheight\@tempdimb } \end{lstlisting} The width of the textblock (on all pages) is 150mm, which on A4 paper implies margins of 30mm each. (Making both horizontal margins the same in a two-sided context makes the paper more pleasant to read on screen). \begin{lstlisting}[firstnumber=last] \setlength{\textwidth}{150mm} \setlength{\oddsidemargin}{30mm - \hoffset - 1in} \setlength{\evensidemargin}{30mm - \hoffset - 1in} \end{lstlisting} It is rare to have marginal notes, but in case we ever do, we centre them in the margin. \begin{lstlisting}[firstnumber=last] \setlength{\marginparwidth}{30mm - 2\marginparsep} \end{lstlisting} We also want a distance of 15mm from the top of the page to the top of the header, and two blank lines between the bottom of the header and the top of the textblock. \begin{lstlisting}[firstnumber=last] \setlength{\topmargin}{15mm - \voffset - 1in} \setlength{\headsep}{2\baselineskip} \end{lstlisting} IJDC editorials have slightly different headers and footers. This requires testing for |\dccp@editorial| if it exists. In case it doesn't, we provide it. \begin{lstlisting}[firstnumber=last] \providecommand{\dccp@editorial}{Editorial} \end{lstlisting} The height of the footer can vary a lot. To keep it a fixed distance from the bottom of the page rather than the top, we need to vary the |\textheight| accordingly. This means we need to measure the height of the footer. (The header is more predictable but we may as well measure it while we are at it). Here we define the header and footer of the title page (i.e.\ the \textsf{title} page style), making sure we save them to auxiliary macros |\TitleHead| and |\TitleFoot| so we can measure them. \begin{lstlisting}[firstnumber=last] \def\ps@title{% \def\@oddhead{% \begin{minipage}{\textwidth}% \centering \LARGE\bfseries\color{struct}% \ifx\dccp@type\dccp@editorial \dccp@publ@long \else \dccp@publ@short\space\space\textbar\space\space\emph{\dccp@type}% \fi \par \end{minipage}% }% \let\@evenhead=\@oddhead \let\TitleHead=\@oddhead \def\@oddfoot{% \begin{minipage}[b]{\textwidth}% \fontsize{9pt}{11pt}\selectfont \ifx\dccp@type\dccp@editorial \else {\centering\dccp@dates\par} \bigskip Correspondence should be addressed to \thecorrespondence\par \bigskip \fi \ifx\empty\dccp@conf \else \dccp@conf\par \bigskip \fi \dccp@publ@msg\par \bigskip \begin{minipage}[b]{\linewidth - 25mm} Copyright rests with the authors. This work is released under a Creative Commons Attribution 4.0 International Licence. For details please see \url{http://creativecommons.org/licenses/by/4.0/}% \end{minipage}\hfill \begin{minipage}[b]{19mm} \href{http://creativecommons.org/licenses/by/4.0/}% {\includegraphics[width=\hsize]{dccpaper-by}}% \end{minipage} \par \bigskip \makebox[0pt][l]{\parbox{0.4\hsize}{% \ifx\undefined\dccp@titlefoot@bib\else\dccp@titlefoot@bib\fi }}\hfill \makebox[0pt][c]{\normalsize\thepage}\hfill \makebox[0pt][r]{\parbox{0.4\hsize}{% \raggedleft\ifx\undefined\dccp@titlefoot@doi\else\dccp@titlefoot@doi\fi }}% \end{minipage}% }% \let\@evenfoot=\@oddfoot \let\TitleFoot=\@oddfoot } \end{lstlisting} We set the normal page style to \textsf{title} here so that |\TitleHead| and |\TitleFoot| are defined, but we will override it with the \textsf{dccpaper} page style later. \begin{lstlisting}[firstnumber=last] \pagestyle{title} \end{lstlisting} The first page should use the \textsf{title} page style, however. \begin{lstlisting}[firstnumber=last] \AtBeginDocument{\thispagestyle{title}} \end{lstlisting} Here are the normal headers and footers (i.e.\ the \textsf{dccpaper} page style). We save them to |\NormalHead| and |\NormalFoot|, again so we can measure them. \begin{lstlisting}[firstnumber=last] \def\ps@dccpaper{% \def\@oddhead{% \begin{minipage}{\textwidth}\frenchspacing {% \fontsize{9pt}{11pt}\selectfont \ifx\undefined\dccp@normhead@doi\else\dccp@normhead@doi\fi }\hfill {\MainAuthor}\space\space\space \textcolor{struct}{\textbar}\space\space\space \thepage\par \vskip6pt\color{struct}{\hrule height 1bp}\par \end{minipage} }% \def\@evenhead{% \begin{minipage}{\textwidth} \thepage\space\space\space \textcolor{struct}{\textbar}\space\space\space {\HeadTitle}\hfill {% \fontsize{9pt}{11pt}\selectfont \ifx\undefined\dccp@normhead@doi\else\dccp@normhead@doi\fi }\par \vskip6pt\color{struct}{\hrule height 1bp}\par \end{minipage} }% \let\NormalHead=\@oddhead \def\@oddfoot{\begin{minipage}[b]{\textwidth} \centering\bfseries\normalsize\color{struct} \ifx\dccp@type\dccp@editorial \dccp@publ@long \else \dccp@publ@short\space\space\textbar\space\space\emph{\dccp@type}% \fi \par \end{minipage}}% \let\@evenfoot=\@oddfoot \let\NormalFoot=\@oddfoot } \pagestyle{dccpaper} \end{lstlisting} We need to wait until the author has supplied the necessary information before we can do our measuring and set the remainder of the geometry, so we do it at the end of the preamble. First we put our saved macros into boxes we can measure (i.e.\ |\dccp@firstpagehead|, |\dccp@firstpagefoot|, |\dccp@restpagehead|, |\dccp@restpagefoot|). \begin{lstlisting}[firstnumber=last] \AtEndPreamble{ \newsavebox{\dccp@firstpagehead} \sbox\dccp@firstpagehead{\normalfont\TitleHead} \newsavebox{\dccp@firstpagefoot} \sbox\dccp@firstpagefoot{\normalfont \def\email#1{#1}\def\url#1{#1}\def\href#1#2{#2}\TitleFoot} \newsavebox{\dccp@restpagehead} \sbox\dccp@restpagehead{\normalfont\NormalHead} \newsavebox{\dccp@restpagefoot} \sbox\dccp@restpagefoot{\normalfont\NormalFoot} \end{lstlisting} We can now set the geometry of the title page\dots \begin{lstlisting}[firstnumber=last] \setlength{\headheight}{\ht\dccp@firstpagehead + \dp\dccp@firstpagehead} \setlength{\footskip}{% 2\baselineskip + \ht\dccp@firstpagefoot + \dp\dccp@firstpagefoot } \setlength{\textheight}{% \paperheight - 30mm % 15mm top and bottom - \headheight - \headsep - \footskip } \end{lstlisting} {\dots}and provide a macro that will reset the geometry for the remaining pages. \begin{lstlisting}[firstnumber=last] \def\dccp@resetgeometry{% \setlength{\headheight}{\ht\dccp@restpagehead + \dp\dccp@restpagehead} \global\headheight=\headheight \setlength{\footskip}{% 2\baselineskip + \ht\dccp@restpagefoot } \global\footskip=\footskip \setlength{\textheight}{% \paperheight - 30mm % 15mm top and bottom - \headheight - \headsep - \footskip } \FixTextHeight \global\textheight=\textheight } } \end{lstlisting} The |\maketitle| command is redefined to the correct formatting. At the end it sets a hook that will reset the geometry when the first page is shipped out, i.e.\ with effect from the second page. It is here rather than at the end of the abstract in case the abstract itself spills over to the second page. \begin{lstlisting}[firstnumber=last] \RequirePackage{atbegshi} \renewcommand{\maketitle}{% \null\nobreak\vspace*{-0.528\baselineskip}% \begingroup \centering {\Large\thetitle\par} \vspace{0.7\baselineskip} \AuthorBlock\par \vspace{1.7\baselineskip} \endgroup \AtBeginShipoutNext{\dccp@resetgeometry}% } \end{lstlisting} The \textsf{abstract} environment is redefined in terms of an environment \textsf{widequote}, which mimics the \textsf{quote} environment, but is a bit wider. We also provide a hook, |\afterabstract|, so that if some annotation needs to be appended to the title page after the abstract, we can do that. \begin{lstlisting}[firstnumber=last] \newenvironment{widequote}{% \list{}{% \setlength{\rightmargin}{2\parindent}% \setlength{\leftmargin}{2\parindent}% }% \flushleftright\item[]% }{% \endlist } \def\afterabstract{} \renewenvironment{abstract}{% \vskip1em% \begin{center}% {\bfseries\abstractname\vspace{-.5em}\vspace{\z@}}% \end{center}% \widequote\footnotesize }{% \endwidequote\afterabstract\newpage } \end{lstlisting} We use the \textsf{titlesec} package to give headings the correct formatting. The settings below try to space out headings so they occupy an integer number of normal lines (an attempt at grid typesetting). They are a little complicated because we want it to work even if the heading appears at the top of the page. \begin{lstlisting}[firstnumber=last] \RequirePackage{titlesec} \titlespacing*{\section}{0pt}{0pt}{\baselineskip} \titlespacing*{\subsection}{0pt}{0pt}{0.6\baselineskip} \titlespacing{\subsubsection}{\parindent}{\baselineskip}{0pt} \titlespacing{\paragraph}{\parindent}{\baselineskip}{0pt} \titlespacing{\subparagraph}{\parindent}{\baselineskip}{0pt} \end{lstlisting} An unfortunate side effect of spacing headings like this is that if a |\subsection| immediately follows a |\section| it forms an unsightly gap. To remedy this, we count how many paragraphs there have been since the last |\section|. Note that as we do not normally number the sections, an automatic reset of the |sectionpars| counter within the |section| counter won't work. \begin{lstlisting}[firstnumber=last] \newcounter{sectionpars} \let\dccp@old@ep\everypar \newtoks\everypar \dccp@old@ep{\the\everypar\stepcounter{sectionpars}} \end{lstlisting} We need to manually reset |sectionpars| when |\section| is called. Also, the normal font size is 12pt/14.5pt, while |\Large| is 17pt/22pt; so the |\Large| line height = 1.5172 $\times$ normal line height. Nevertheless it seems to work better if we let the heading eat 0.528|\baselineskip| into the 2|\baselineskip| of padding above it. \begin{lstlisting}[firstnumber=last] \titleformat{\section} [block] {% \vspace{2\baselineskip}% \nobreak \vspace*{-0.528\baselineskip}% \setcounter{sectionpars}{0}% \filcenter\normalfont\Large\bfseries } {\thesection} {1em} {} \end{lstlisting} The others use a |\normalsize| font so that makes life easier. The format for |\subsection| command includes conditional spacing: if the |sectionpars| counter equals 2, this means the heading immediately follows a |\section|, so less white space is needed. \begin{lstlisting}[firstnumber=last] \titleformat{\subsection} {% \ifnum\thesectionpars>2% \vspace{2\baselineskip}% \else \vspace{\baselineskip}% \fi\nobreak \vspace*{-0.6\baselineskip}% \normalfont\normalsize\bfseries } {\thesubsection} {1em} {} \titleformat{\subsubsection} [block] {\normalfont\normalsize\bfseries} {\thesubsubsection} {1em} {} \titleformat{\paragraph} [block] {\normalfont\normalsize\bfseries\itshape} {\thesubsubsection} {1em} {} \titleformat{\subparagraph} [block] {\normalfont\normalsize\itshape} {\thesubsubsection} {1em} {} \end{lstlisting} DCC papers do not typically number their sections. \begin{lstlisting}[firstnumber=last] \setcounter{secnumdepth}{0} \end{lstlisting} To help with the display of tables we load the \textsf{array} and \textsf{booktabs} packages. As we don't like lines between rows in the table body, we stretch them out a bit so that white space does the job instead. \begin{lstlisting}[firstnumber=last] \RequirePackage{array,booktabs} \renewcommand{\arraystretch}{1.25} \end{lstlisting} We use the \textsf{caption} package to give captions the right format. \begin{lstlisting}[firstnumber=last] \RequirePackage [ format=hang , labelsep=period , font=small , labelfont=bf , figureposition=bottom , tableposition=top ]{caption} \end{lstlisting} Footnotes should be set right up against the left margin. They should be set hung and in the same half-ragged style as the main text. They should also, for neatness, be at the bottom of the page regardless of how short it is. The \textsf{footmisc} package helps here. \begin{lstlisting}[firstnumber=last] \RequirePackage[hang,bottom]{footmisc} \settowidth{\footnotemargin}{\footnotesize\textsuperscript{99}\space} \renewcommand{\footnotelayout}{\raggedyright} \end{lstlisting} Also, if multiple footnotes are set at once, the markers should be separated with superscript commas. The \textsf{footmisc} package should help here but its solution is clobbered by \textsf{hyperref}. So after a footnote is set, we check to see if the next token is also a footnote, and if so, slip a comma in before it.\footnote{This solution was provided at \url{http://tex.stackexchange.com/q/40072}} This tweak needs to be done late, |\AtBeginDocument|. Note that the \textsf{newtx} superior figures are a bit lower than normal superscript text. \begin{lstlisting}[firstnumber=last] \AtBeginDocument{ \let\dccp@footnote\footnote \def\dccp@next@token{\relax}% \def\dccp@supercomma{\textsuperscript{,}}% \IfFileExists{newtxtext.sty}% {\def\dccp@supercomma{\raisebox{-0.2ex}{\textsuperscript{,}}}}% {} \newcommand\dccp@check@for@footnote{% \ifx\footnote\dccp@next@token \dccp@supercomma \fi } \renewcommand\footnote[1]{% \dccp@footnote{#1}% \futurelet\dccp@next@token\dccp@check@for@footnote } } \end{lstlisting} By default lists are quite loose. These settings help to tighten them. \begin{lstlisting}[firstnumber=last] \topsep = \z@ \partopsep = \z@ \appto{\enumerate}{\itemsep = 0.5ex plus 0.25ex minus 0.25ex} \appto{\itemize}{\itemsep = 0.5ex plus 0.25ex minus 0.25ex} \end{lstlisting} A DCC paper should either be using \textsf{biblatex} or \textsf{apacite} for references. If \textsf{biblatex} is used, we need to ensure that the reference list heading is a normal section rather than a starred one so it appears in the PDF bookmarks. \begin{lstlisting}[firstnumber=last] \AtBeginDocument{ \@ifpackageloaded{biblatex}{% \defbibheading{bibliography}[\refname]{\section{#1}}% \end{lstlisting} We also move the ‘doi:’ portion of a DOI inside the hyperlink. \begin{lstlisting}[firstnumber=last] \DeclareFieldFormat{doi}{% \ifhyperref{% \href{http://dx.doi.org/#1}{\nolinkurl{doi:#1}}% }{% \nolinkurl{doi:#1}% }% } }{% \end{lstlisting} If \textsf{apacite} is used, there are a few other adaptations we need to make. \begin{lstlisting}[firstnumber=last] \@ifpackageloaded{apacite}{% \end{lstlisting} With \textsf{hyperref} loaded, \textsf{apacite} makes the whole of a citation a link to the reference list item. We patch |\@ifauthorsunequalc@de| so only the year portion gets linked. \begin{lstlisting}[firstnumber=last] \def\@ifauthorsunequalc@de#1{% \if@F@cite \@F@citefalse \else \if@Y@cite {\@BAY}% \fi {\@BBC}% \fi \edef\@cite@undefined{?}% \def\BBA{\@BBA}% \if@A@cite %%\hyper@natlinkstart{#1}% We remove this line... {\csname b@\@citeb\APAC@extra@b@citeb\endcsname}% %%\hyper@natlinkend% ...and this one. \if@Y@cite {\@BBAY}% \fi \fi \if@Y@cite \hyper@natlinkstart{#1}% {\csname Y@\@citeb\APAC@extra@b@citeb\endcsname}% \hyper@natlinkend \fi \let\BBA\relax } \end{lstlisting} The Spanish language support file defines a different version of |\@ifauthorsunequalc@de|, which might override the patch we have just introduced. So we employ the same test that \textsf{apacite} uses when deciding whether to load that file; if successful, we patch the Spanish version. Note that as \textsf{apacite} loads language support files |\AtBeginDocument|, we have to do our thing after that, |\AfterEndPreamble|. (Note that as we set the language to British English earlier, this should never be needed, but we try to be resilient to tinkering!) \begin{lstlisting}[firstnumber=last] \AfterEndPreamble{% \@ifundefined{iflanguage}{% \relax }{% \edef\APAC@tmp{nohyphenation}% \ifx\languagename\APAC@tmp \else \edef\APAC@tmp{spanish}% \ifx\languagename\APAC@tmp \def\@ifauthorsunequalc@de#1{% \if@F@cite \@F@citefalse \else \if@Y@cite {\@BAY}% \fi {\@BBC}% \fi \edef\@cite@undefined{?}% \def\BBA{\@BBA}% \@ifundefined{spanishe@\@citeb\APAC@extra@b@citeb}% {}% skip {{% Use `e' instead of `y' in Spanish \global\let\oldBBA\BBA \global\def\BBA{e\global\let\BBA\oldBBA}% }}% \if@A@cite %%\hyper@natlinkstart{#1}% We remove this line... {\csname b@\@citeb\APAC@extra@b@citeb\endcsname}% %%\hyper@natlinkend% ...and this one. \if@Y@cite {\@BBAY}% \fi \fi \if@Y@cite \hyper@natlinkstart{#1}% {\csname Y@\@citeb\APAC@extra@b@citeb\endcsname}% \hyper@natlinkend \fi \let\BBA\relax }% \fi \fi }% \end{lstlisting} Another thing \textsf{apacite} does |\AtBeginDocument| is set the URL style to monospaced. So we reset it back to normal roman type |\AfterEndPreamble|. \begin{lstlisting}[firstnumber=last] \urlstyle{APACrm} }% \end{lstlisting} We pre-empt \textsf{apacite}'s |\providecommand| of |\doi| with our own definition that includes the `doi' URI scheme label in the link, remembering to remove the one inserted by |\doiprefix|. \begin{lstlisting}[firstnumber=last] \newcommand{\doi}[1]{\href{http://dx.doi.org/#1}{\nolinkurl{doi:#1}}}% \renewcommand{\doiprefix}{\unskip}% }{}% }% \end{lstlisting} Both \textsf{biblatex} and \textsf{apacite} use |\bibitemsep| for the space between bibliography items. Just in case they haven't been loaded, though, we protect our setting of that length with an |\ifx| test. \begin{lstlisting}[firstnumber=last] \ifx\undefined\bibitemsep \else \setlength{\bibitemsep}{1em plus 1ex minus 1ex}% \fi } \end{lstlisting} As mentioned above, if \textsf{apacite} is used, we can use a package option to ensure that the reference list heading appears in the PDF bookmarks. \begin{lstlisting}[firstnumber=last] \PassOptionsToPackage{numberedbib}{apacite} \end{lstlisting} We, of course, use \textsf{hyperref} for enhancing the PDF with working links, bookmarks, metadata, etc. \begin{lstlisting}[firstnumber=last] \usepackage [ colorlinks=true , linkcolor=black , anchorcolor=black , citecolor=links , filecolor=black , menucolor=black , runcolor=black , urlcolor=links ]{hyperref} \end{lstlisting} Links should be in roman type, not monospaced. \begin{lstlisting}[firstnumber=last] \urlstyle{rm} \end{lstlisting} We provide an |\email| command for displaying the email address of the corresponding author. \begin{lstlisting}[firstnumber=last] \newcommand*{\email}[1]{\href{mailto:#1}{#1}} \end{lstlisting} Once the user has had a chance to provide the metadata, we can add it to the PDF metadata. \begin{lstlisting}[firstnumber=last] \AtBeginDocument{% \hypersetup { pdftitle={\thetitle} , pdfauthor={\dccp@author} , pdfsubject={\dccp@subject} } \end{lstlisting} The APA has its own style for line breaks in URLs. The \textsf{apacite} package provides the code for this, but in case \textsf{biblatex} is used instead, we repeat the settings (from 2013/07/21 v6.03) here. \begin{lstlisting}[firstnumber=last] \@ifundefined{Url@force@Tilde}{\def\Url@force@Tilde{\relax}}{}% \def\url@apa@dot{\mathchar"2E }% \def\url@apa@comma{\mathchar"2C }% \def\url@apa@questionmark{\mathchar"3F }% \def\url@apa@exclamation{\mathchar"21 }% \def\url@apa@hyphen{\mathchar"2D }% \def\url@apa@underscore{\_}% \def\UrlBreaks{\do\@\do\\\do\|\do\;\do\>\do\]\do\)\do\'\do+\do\=\do\#}% \def\UrlBigBreaks{\do\/\do\:\do@url@hyp}% \def\UrlNoBreaks{\do\(\do\[\do\{\do\<}% \)} \def\UrlOrds{\do\*\do\~\do\'\do\"}% \def\UrlSpecials{% \do\.{\mathbin{}\url@apa@dot }% \do\,{\mathbin{}\url@apa@comma }% \do\-{\mathbin{}\url@apa@hyphen }% \do\?{\mathbin{}\url@apa@questionmark }% \do\!{\mathbin{}\url@apa@exclamation }% \do\_{\mathbin{}\url@apa@underscore }% \do\ {\Url@space}\do\%{\Url@percent}\do\^^M{\Url@space}% \Url@force@Tilde}% \def\Url@OTnonTT{\do\<{\langle}\do\>{\mathbin{\rangle}}\do \_{\mathbin{}\_}\do\|{\mid}\do\{{\lbrace}\do\}{\mathbin{\rbrace}}\do \\{\mathbin{\backslash}}\UrlTildeSpecial} } \end{lstlisting} We now embed the Creative Commons licence information in the PDF using an XMP packet. To do this, we employ the same technique as Scott Pakin's \textsf{hyperxmp} (2014/01/02 v2.4). In order to avoid avoid a bug whereby Adobe Acrobat confuses the XMP author information and the regular author information, though, we \emph{only} embed the licence information. We need to make sure that any characters to appear verbatim in the XMP packet are treated as ordinary characters and not active ones. The likely active characters are symbols and punctuation, so should be treated as `other' (category 12). \begin{lstlisting}[firstnumber=last] \begingroup \catcode`\"=12 \catcode`\&=12 \catcode`\#=12 \catcode`\<=12 \catcode`\>=12 \catcode`\_=12 \end{lstlisting} We construct the XMP packet as the document begins. \begin{lstlisting}[firstnumber=last] \AtBeginDocument{% \end{lstlisting} For convenience we define |\sp| to be a level of indent, translating to three spaces. \begin{lstlisting}[firstnumber=last] \def\sp{\space\space\space} \end{lstlisting} The text of the XMP packet is recorded in |\cc@xmp@packet|. We use |^^J| to break lines. \begin{lstlisting}[firstnumber=last] \long\gdef\cc@xmp@packet{% ^^J% ^^J% ^^J% \sp^^J% \sp\spTrue^^J% \sp^^J% \sp^^J% \sp\sp^^J% \sp\sp\sp^^J% \sp\sp\sp\spThis work is licensed under a Creative Commons Attribution 4.0 International Licence.^^J% \sp\sp\sp^^J% \sp\sp^^J% \sp^^J% \sp^^J% \sp\sp^^J% \sp^^J% ^^J% ^^J% ^^J% }% } \endgroup \end{lstlisting} Different workflows require the XMP packet to be embedded in different ways. Pdf\TeX\ can inject objects into PDFs natively. \begin{lstlisting}[firstnumber=last] \newcommand*{\ccxmp@embed@packet@pdftex}{% \bgroup \pdfcompresslevel=0 \immediate\pdfobj stream attr {% /Type /Metadata /Subtype /XML }{\cc@xmp@packet}% \pdfcatalog {/Metadata \the\pdflastobj\space 0 R}% \egroup } \end{lstlisting} The |\pdfmark| command defined by \textsf{hyperref} is respected by tools such as Dvipdf, Dvips, Dvipsone, etc. \begin{lstlisting}[firstnumber=last] \newcommand*{\ccxmp@embed@packet@pdfmark}{% \pdfmark{% pdfmark=/NamespacePush }% \pdfmark{% pdfmark=/OBJ, Raw={/_objdef \string{ccxmp@packet\string} /type /stream}% }% \pdfmark{% pdfmark=/PUT, Raw={\string{ccxmp@packet\string} 2 dict begin /Type /Metadata def /Subtype /XML def currentdict end }% }% \pdfmark{% pdfmark=/PUT, Raw={\string{ccxmp@packet\string} (\cc@xmp@packet)}% }% \pdfmark{% pdfmark=/Metadata, Raw={\string{Catalog\string} \string{ccxmp@packet\string}}% }% \pdfmark{% pdfmark=/NamespacePop }% } \end{lstlisting} Dvipdfm has its own |\special| command for inserting PDF objects, but it is a bit basic and requires advance knowledge of how long (in characters) the object is. The |\ccxmp@count@spaces| macro counts the number of spaces in its parameter through a process of iteration, adding this figure to |\@tempcnta|. \begin{lstlisting}[firstnumber=last] \def\ccxmp@count@spaces#1 {% \def\ccxmp@one@token{#1}% \ifx\ccxmp@one@token\@empty \advance\@tempcnta by -1 \else \advance\@tempcnta by 1 \expandafter\ccxmp@count@spaces \fi } \end{lstlisting} The |\ccxmp@count@non@spaces| command counts the number of non-spaces in its argument through a process of iteration, adding this figure to |\@tempcnta|. \begin{lstlisting}[firstnumber=last] \newcommand*{\ccxmp@count@non@spaces}[1]{% \def\ccxmp@one@token{#1}% \ifx\ccxmp@one@token\@empty \else \advance\@tempcnta by 1 \expandafter\ccxmp@count@non@spaces \fi } \end{lstlisting} The |\ccxmp@string@len| command sets |\@tempcnta| to the number of characters (spaces + non-spaces) in its argument. \begin{lstlisting}[firstnumber=last] \newcommand*{\ccxmp@string@len}[1]{% \@tempcnta=0 \expandafter\ccxmp@count@spaces#1 {} % \expandafter\ccxmp@count@non@spaces#1{}% } \end{lstlisting} So now, finally, is the command for embedding the packet using Dvipdfm. \begin{lstlisting}[firstnumber=last] \newcommand*{\ccxmp@embed@packet@dvipdfm}{% \ccxmp@string@len{\cc@xmp@packet}% \special{pdf: object @ccxmp@packet << /Type /Metadata /Subtype /XML /Length \the\@tempcnta >> stream^^J\cc@xmp@packet endstream% }% \special{pdf: docview << /Metadata @ccxmp@packet >> }% } \end{lstlisting} \XeTeX\ creates PDFs with Xdvipdfmx, which supports a simpler |\special| for inserting objects that does not require us to count characters. \begin{lstlisting}[firstnumber=last] \newcommand*{\ccxmp@embed@packet@xetex}{% \special{pdf:stream @ccxmp@packet (\cc@xmp@packet) << /Type /Metadata /Subtype /XML >> }% \special{pdf:put @catalog << /Metadata @ccxmp@packet >> }% } \end{lstlisting} We rely on \textsf{hyperref} to tell us how the PDF will be generated (after all, it may not be done in the current pass) and use the respective technique to embed the XMP packet. \begin{lstlisting}[firstnumber=last] \AtBeginDocument{% \begingroup \def\ccxmp@driver{hpdftex}% \ifx\ccxmp@driver\Hy@driver \ccxmp@embed@packet@pdftex \else \def\ccxmp@driver{hdvipdfm}% \ifx\ccxmp@driver\Hy@driver \ccxmp@embed@packet@dvipdfm \else \def\ccxmp@driver{hxetex}% \ifx\ccxmp@driver\Hy@driver \ccxmp@embed@packet@xetex \else \@ifundefined{pdfmark}{}{% \ccxmp@embed@packet@pdfmark }% \fi \fi \fi \endgroup } %% %% Copyright (C) 2015 Digital Curation Centre, University of Edinburgh %% %% %% End of file `dccpaper-base.tex'. \end{lstlisting}