% pdccode.dtx -- documentation & source for pdccode.tex -*-tex-*-
        
%%%@TeX-document-file {
%%% title 	= "PDCCODE -- Documented TeX Macros",
%%% filename 	= "$texmf/doc/plain/pdcmac/pdccode.dtx",
%%% version 	= "$Revision: 1.6 $",
%%% package 	= "pdcmac 1.0",
%%% date 	= "$Date: 1995/03/30 16:08:32 $",
%%% author	= "P. Damian Cugley",
%%% email	= "damian.cugley@comlab.ox.ac.uk",
%%% address	= "Oxford University Computing Laboratory,"
%%%                Parks Road, Oxford  OX1 3QD, UK",
%%% abstract	= "This document describes and is the source code for 
%%%                the TeX definitions file pdccode.tex.
%%%                Running plain TeX on this file produces both a copy of the
%%%                definitions file and the printed documentation.",
%%% copyright	= "Copyright (c) 1991-1995 P. Damian Cugley",
%%% copying	= "This program is free software; you can redistribute
%%%		   it and/or modify it under the terms of the GNU
%%%		   General Public License as published by the Free
%%%		   Software Foundation; either version 2 of the License,
%%%		   or (at your option) any later version.",
%%% notice	= "This program is distributed in the hope that it will
%%%		   be useful, but WITHOUT ANY WARRANTY; without even the
%%%		   implied warranty of MERCHANTABILITY or FITNESS FOR A
%%%		   PARTICULAR PURPOSE.  See the GNU General Public
%%%		   License for more details.",
%%% notice	= "You should have received a copy of the GNU General
%%%		   Public License along with this program; if not, write
%%%		   to the Free Software Foundation, Inc., 675 Mass Ave,
%%%		   Cambridge, MA 02139, USA.",
%%% codetable	= "USASCII",
%%% dependencies = "pdccode.tex"
%%%}

%{{{ pdccode
%{{{  preamble

\relax
\input pdccode
\document
\rcs$Id: pdccode.dtx,v 1.6 1995/03/30 16:08:32 pdc Exp $\endrcs

\codefile{tmp.tex}

\defverbatim\"

%}}}  preamble
%{{{  introduction
	
\author{P. Damian Cugley}
\title{PDCCODE---Documented \TeX\ Macros}
\section{Introduction}

	PDCCODE is a simpler and less featureful system than the
	\LaTeX~2e Docstrip package, but uses the same file suffix
	`|dtx|'.  Lines between |\code| and |\endcode| markers are
	written to a file (set with the |\codefile| command), as well as
	being printed `verbatim' in document.  Thus one \TeX\ pass
	generates both the macro file(s) and a |dvi| file.

	This document---generated from |pdccode.dtx|---is a |dtx| file
	for PDCCODE itself, the file of macros used by my |dtx| files.
	It cannot be used to generate |pdccode.tex| directly (\TeX\
	might choke halfway through overwriting |pdccode.tex| with a new
	version, leaving no |pdccode.tex| to generate |pdccode.tex|
	with).  Instead it produces its copy as `|tmp.tex|', and
	`|tmp.tex|' can be compared to the supplied |pdccode.tex|.

\subsec{How to read this document}

	The |dtx| files are programs (in \TeX's macro language) that
	generate new programs (the macro files that will be used in
	documents).  The macros described in {\it this} file are used
	when generating macro files.  So, for example, the |\rcs| and
	|\fileversion| macros used to extract identification information
	from the |dtx| file are not defined in any of the generated
	macro files.

\subsec{Background}

	For some time I~have been using a collection of macro files to
	typeset various types of document with \TeX.  After using them
	in Malvern\footnote*{Malvern is a sanserif font family
	implemented in METAFONT.  The current version is release $1.2$.}
	releases I have discovered that they had been merged with the
	Unix\TeX\ macros directory.  Originally I~had never intended
	that these files be used by anyone but me---they were full of
	hacks and messy code and definitions I~don't use anymore.  Since
	then I~thought of tidying up the macros enough that I'm not
	embarassed by them, but the task of doing this {\it and}
	documenting their use was daunting.

	After installing \LaTeX~2e on my new computer, I~decided to have
	a go at adapting the idea of `documented \TeX\ macros' files to
	plain \TeX.  This way I~could clean up the macros and document
	them in a systematic fashion.

	The \LaTeX\ Docstrip/Doc-option approach evolved through several
	different models and so is a complex format.  Docstrip (a
	program written in \TeX's macro language) was used to strip
	comments from |doc| files making |sty| files, and the \LaTeX\
	Doc-option separately extracted \LaTeX\ code from comments in
	|doc| files to make printed documentation.  Both Doc and
	Docstrip required the user to type in file names whenever they
	ran---which made them useless for batch processing.  So separate
	driver files (with suffixes |drv| and |ins|) were added to
	distributions.  Finally, the |doc| files, which used to be
	usable as |sty| files (just a little slower because of the
	comments) were have been changed to be documents instead (making
	|drv| files redundant, once the |ltxdoc.cfg| configuration file
	was added to the conventions) and given a new suffix |dtx|.

	The system I~use---called PDCCODE---is simpler and faster, but
	omits indexes, special change-log commands, and all the features
	that come with \LaTeX\ formatting.  It doesn't have Docstrip's
	{\it options}\/ feature, which allows several largely-similar
	files to be generated from the same source.\footnote\dag{I do
	have an experimental version |pdccode2.tex| which does this but
	I~am not particularly happy with it yet.}  It doesn't print
	statistics at the end, or do checksums, or even print comforting
	lines of asterisks and percent signs in the log file.

	My |dtx| files use a single macro file |pdccode.tex| which
	defines a `code mode' where lines are written to the macro file
	as well as appearing in the printed documentation.  One pass
	generates both the macro file and the documentation in |dvi|
	form.  The |pdccode.tex| file also includes enough layout
	commands to typeset the comments.
%}}}
%{{{  using pdccode
\section{Summary of using PDCCODE}

	These commands are explained in more detail below.
	
\subsec{A skeleton  code file}

	A code file using |pdccode.tex| will look something like this:
$$\it
\halign{\indent#\hfil\cr
	"%" name "--" one line description\cr
\noalign{\smallskip}
	"%%%@TeX-document-file {"\cr
	header comments for the document\cr
	"%%%}"\cr
\noalign{\smallskip}
	"\relax"\cr
	"\input pdccode"\cr
	"\document"\cr
	"\rcs$" RCS identification string "$\endrcs"\cr
	"\codefile{" file name "}"\cr
	"\author{" author name "}"\cr
	"\title{" title for the document "}"\cr
\noalign{\smallbreak}
	"\section{Introduction}"\cr
	introduction\cr
\noalign{\smallbreak}
	"\section{File identification}"\cr
	"\code"\cr
	"  \|\% " file name "\fileversion~\filedate~--" one line desc\cr
	"  |"\cr
	"  |%%%@TeX-definition-file {"\cr
	"  " header comments for macro file\cr
	"  |%%%}"\cr
	"  |"\cr
	"  \|\\message\{\fileversion~<\fileauthor~\filedate>\}"\cr
	"  |"\cr
	"\endcode"\cr
\noalign{\smallbreak}%
	definitions and commentary\cr
\noalign{\smallbreak}%
	"\endcodefile"\cr
	"\enddocument"\cr
	"\bye"\cr
}$$

\subsec{Commands in code sections}

	Code sections start with |\code| and end with |\endcode|.  Before
	the first |\code| you must do |\codefile{|\<file name>|}| to
	specify the file.

	Vertical mode material may be used between lines (e.g.,
	|\smallbreaks|).  Unexpanded code lines are introduced with "|",
	expanded lines with "\|".  In expanded lines, use "\{", "\}",
	"\\", "\%", "\~" for those characters, "~" for a single space.

\subsec{Commands in commentary}

	Most of the commentary will be typeset using the conventions of
	plain \TeX.

	|\author{|\<text>|}|  (must precede |\title|)

	|\title{|\<heading>|}|

	|\section{|\<heading>|}|

	|\subsec{|\<heading>|}|

	|\noindenttrue|

	|\footnote|

	|\defverbatim|\<one-character control sequence>

	|$|\<number>|\mm$| |$|\<number>|\em$| |$|\<number>|\pt$|
	|$|\<number>|\inch$| 

	|\<|\<name>|>| |\cs{|\<name>|}|
	|\arg{|\<argument name>|}|

	|\TeX| |\LaTeX| |\LaTeX~2e|

	|\notepar| \<paragraph> |\endnotepar|
	
	
%}}}  using pdccode
%{{{  file identification
\section{File identification}

	Nowadays, macro files start with some comments identifying the
	file.
\code \catcode`\"=12
	\|\% pdccode.tex \fileversion~\filedate~-- self-documenting TeX macros
	|
	|%%%@TeX-definition-file {
	|%%% filename       = "$texmf/tex/plain/pdcmac/pdccode.tex",
	\|\%\%\% version~~~~~~~~= "\fileversion",
	\|\%\%\% date~~~~~~~~~~~= "\filedate",
	|%%% package        = "pdcmac 1.0",
	|%%% author         = "P. Damian Cugley",
	|%%% email          = "damian.cugley@comlab.ox.ac.uk",
	|%%% address        = "Oxford University Computing Laboratory,
	|%%%                   Parks Road, Oxford  OX1 3QD, UK",
	|%%% codetable      = "USASCII",
	|%%% keywords       = "TeX, plain TeX, macros",
	|%%% supported      = "Maybe",
	|%%% abstract       = "Macros for documented TeX macros (DTX files).
	|%%%                   This file was generated by running
	\|\%\%\%~~~~~~~~~~~~~~~~~~~plain TeX on \jobname.dtx",
	|%%% copyright      = "Copyright (c) 1991-1995 P. Damian Cugley",
	|%%% copying        = "DO NOT DISTRIBUTE THIS FILE.
	\|\%\%\%~~~~~~~~~~~~~~~~~~~Distribute \jobname.dtx only as part of the
	|%%%                   package it came in.",
	|%%% dependencies   = "",
	|%%% }
	|
	\|\\message\{\fileversion~<pdc \filedate>\}
	|
\endcode
%}}}  file id

\bigskip\noindent
	We start by making a nickname for a scratch register.  This is
	|\toks@| in plain \TeX; this definition gives it a name I~can
	use without catcode hackery.
\code
	|\toksdef\toksa=0
\endcode

%{{{  rcs support
	
\section{RCS Support}

	RCS (Revision Control System) is a free source code control
	system which is available from the Free Software Foundation.
	It inserts keyword-value pairs enclosed in dollar signs into the
	source code.  This command extracts the file's version, date,
	author and locker (if there is one).  It is used like this:
$$\centerline{\vbox{\tt\def~{\char32 \ignorespaces}
		\halign{#\hfil\cr
	\cs{rcs}\$Id:~\<name>,v~\<version>~\<date>~\<ignored>\quad\cr
	\hfill~\<author>~\<ignored>~{\rm[}\<locker>~{\rm]}\$\cs{endrcs}\cr
}}} $$
 	The \<locker> field is non-empty iff the file is `locked'
	(meaning it has been edited and the changes are not checked in
	yet).  The |\filename| macro is defined, even though it will
	almost certainly be identical to |\jobname|.
\code
	|\def\rcs#1: #2,v #3 #4 #5 #6 #7 #8$#9\endrcs{
	|    \def\filename{#2}
	|    \def\fileversion{#3}
	|    \def\filedate{#4}
	|    \def\fileauthor{#6}
	|    \def\filelocker{#8}
	|}
\endcode
	
%}}}  rcs support
%{{{  verbatim text
\section{Verbatim text}

	Verbatim text is text that is printed identically to the
	characters in the manuscript file.  In the code sections, a
	vertical bar is used to make lines for the code file.  In the
	documentation, the vertical bar is used as a self-matching
	verbatim mode character.

	These definitions should be the same as the corresponding code
	in |pdcfmt2.dtx|.

\subsec{Lists of active characters}

	The verbatim text macros need to know if any characters have
	been made special apart from the usual ones.  This command tells
	the verbatim macros that |#1| is an active char; in addition to
	setting its catcode, this makes sure that the character will be
	re-catcoded appropriately when in verbatim mode.  |#1| should be
	a something that will work after `|`|'; usually this means a
	one-character control sequence.
\code
	|\def\declareactivechar#1{%
	|    \toksa\expandafter{\verbatimplains\do#1}%
	|    \edef\verbatimplains{\the\toksa }%
	|    \catcode`#1\active
	|}
\endcode

	Two list macros in the style of |\dospecials| are defined.  The
	first is for characters that must be set to catcode~12, such as
	`|\|', `|{|', `|}|', `|_|', `|$|', `|#|', `|&|', and `|%|'.  The
	other is for a few characters that are given catcode~13, such as
	`|`|', `|'|', `|~|', `|^|' and SP.
\code
	|\def\verbatimplains{\do\\\do\{\do\}\do\_\do\$\do\#\do\&\do\%}
	|\def\verbatimactives{\do\`\do\'\do\~\do\^\do\ }
\smallbreak
	|\chardef\other=12
\endcode

\subsec{Setting up verbatim mode}

	This is a macro common to various verbatim text modes.  The
	default spacing of whatever font is used, without any
	adjustments of spacefactor for punctuation.  The characters
	`|`|', `|'|', `|~|', and `|^|' produce slightly lowered glyphs.
\code
	|\newtoks\everyverbatim
\smallbreak
	|\bgroup \catcode`\^=13 \catcode`\'=13 \catcode`\`=13 \toksa={\egroup
	|    \def\setupverbatim{%
	|        \frenchspacing
	|        \spaceskip0pt \xspaceskip0pt % use spacing of font
	|        \def\do##1{\catcode\lq##112 }\verbatimplains
	|        \def\do##1{\catcode\lq##1\active }\verbatimactives
	|        \let`\ttlq \let'\ttrq 
	|        \let~\tttilde \let^\ttcircum
	|        \the\everyverbatim
	|    }
	|}\the\toksa
\endcode

	Here are the definitions for the munged glyphs.  These are only
	necessary in CM-Typewriter-like fonts; using something
	like Courier, these definitions could probably be omitted.

	The names `rq' and `lq' are based on the use of |\lq| and |\rq|
	in plain \TeX.  An alternative definition for |\ttrq| would be
	|\chardef\ttrq=13|, producing `{\tt\char13 }'.
\code
	|\def\ttlq{\lower0.125ex \hbox{\char18 }}
	|\def\ttrq{\lower0.125ex \hbox{\char19 }}
	|\def\tttilde{\lower0.5ex \hbox{\char`\~ }}
	|\def\ttcircum{\lower0.5ex \hbox{\char`\^ }}
\endcode
	This makes the difference between |`'^~| and
	{\tt`'\char`\^\char`\~}.\footnote*{It turns out that on some
	\TeX\ systems, the CMTT font has been hacked to have an
	ASCII-style reverse-apostrohe-cum-grave-accent as character~96
	and a matching apostrophe-cum-acute-accent as charcter~39,
	thereby correcting the defects that my definitions of |\ttlq|
	and |\ttrq| address.  But such modified fonts are
	nonstandard---they should {\it not} be called CMTT---and should
	be replaced.}

\subsec{Defining a self-matching character for verbatim text}

	A macro for defining a self-matching verbatim character (I~got
	the idea of a macro for declaring verbatim characters from the
	\LaTeX~2e documentation macros).  |#1| must be a one-character
	control sequence.  If it isn't active already, then
	|\declareactivechar| is applied to it first.
\code
	|\def\defverbatim#1{%
	|    \ifcat\noexpand#1\noexpand~\else \declareactivechar#1 \fi
	|    \begingroup \uccode`\~=`#1 \uppercase{\toksa={\endgroup
	|        \def~{%
	|            \leavevmode
	|            \begingroup \tt \setupverbatim
	|            \catcode`#1\active \let~\endgroup
	|       }%
	|    }}\the\toksa
	|}
\endcode
	Note that there must be a font called |\tt| for it to switch to.
	(The catcode of |#1| is set to 13 after calling |\setupverbatim|
	because |#1| will be included in |\verbatimplains|.)

	Finally, we define vertical bar as the conventional
	self-matching verbatim-mode character.
\code
	|\defverbatim\|
\smallbreak
	|\hyphenchar\tentt=-1
\endcode
	Both |\defverbatim| and |\defineactivechar| have local scope.

%}}}  verbatim text
%{{{  code sections
%{{{   intro
\section{Code sections}

	The markers |\code| \dots |\endcode| mark code to be written to
	the definition file.  Each line starts with either "|" or "\|"
	and finishes with  the end of the line in the manuscript file.

	Lines that start with "|" are written out as-is; no macro
	expansion is allowed.  This is how the vast majority of lines
	are done.

	Lines starting "\|" are expanded.  This is so that things like
	the current version ID (extracted from the RCS identification
	line or set with |\def|) may be included in the code file.
	Within such an expanded line, the following macros are defined:
    $$\halign{\indent#\hfil&&\quad#\hfil\cr
	|\%|&	|%|&	per cent sign&
	|\{|&	|{|&	left brace&
	|\}|&	|}|&	right brace\cr
	|\\|&	|\|&	backslash&
	|~|&	\tt\char32 & space&
	|\~|&	|~|&	tilde\cr
	}
    $$
	(This list is probably incomplete.)
%}}}   intro
%{{{   code file name
	
\subsec{Setting the codefile name}

	Before the first code section, the |\codefile{|\<file name>|}|
	command specifies the code file.  The last code section should
	be followed by |\endcodefile|.  Only one code file may be open
	at a time.
\code
	|\newwrite\CODEwrite
\smallbreak
	|\def\codefile#1{
	|    \ifeof\CODEwrite\else \endcodefile \fi	
	|    \def\thecodefile{#1} \openout\CODEwrite=#1
	|}
\smallbreak
	|\def\endcodefile{
	|    \message{Code written on \thecodefile.}
	|    \closeout\CODEwrite
	|}	
\endcode
%}}}   code file name
%{{{   code sections
	
\subsec{Code sections}

	Switching to `code mode' is easy, because most of the work is
	really done by the "|" and "\|" commands.

\code
	|\def\code{
	|    \smallskip
	|    \begingroup
	|    \let|\CODEline \let\|\CODElinex
	|    \let\{=\CODElb \let\}=\CODErb \let\\=\CODEbslash 
	|    \def~{ } \let\~\CODEtilde
	|    \the\everycode
	|    \obeylines
	|}
\smallbreak
	|\def\endcode{
	|    \smallbreak
	|    \endgroup
	|}
\endcode
%}}}   code sections
%{{{   code lines
\subsec{Code lines}

	This is the definition for unexpanded code lines. 
\code
	|\def\CODEline{%
	|    \begingroup	% matched by \CODElinei
	|    \setupverbatim \catcode`\|\other
	|    \CODElinei
	|}
\endcode

	In |\CODElinei|, We must redefine the characters (such as |`|)
	that are active within verbatim text to expand to their
	catcode-12 equivalents when they are written to the code file
	currectly.

	Since one of these is |^|, we temporarily give |@| the meaning
	of |^| so that we may safely change the catcode of |^|.
\code
	|\begingroup\catcode`\'=\active \catcode`\`=\active
	|            \catcode\lq\@=\catcode\lq\^ \catcode\lq\^=\active
 	|            \catcode\lq\@@M=\active \toksa={\endgroup %
	|    \def\CODElinei#1@@M{%
	|        \begingroup\let`\lq \let'\rq %
	|            \let^\CODEcircum \let~ \CODEtilde%
	|            \edef\tmp{#1}%
	|            \write\CODEwrite\expandafter{\tmp}%
	|        \endgroup %
	|        \indent\the\everycodeline #1\par %
	|        \endgroup % matches \CODEline
	|    }%
	|}\the\toksa
\endcode

	The definition for |\CODElinex| (expanded code lines) is much
	simpler: 
\code
	|\begingroup\catcode`\^^M=13 \toksa={\endgroup %
	|    \def\CODElinex#1^^M{%
	|        \edef\tmp{\write\CODEwrite{#1}}\tmp %
	|        {\indent\the\everycodeline #1\par}%
	|    }%
	|} \the\toksa
\endcode
%}}}   code lines
%{{{   catcode hacks
\subsec{Macros for making category-12 characters}

	For characters that are active in verbatim mode (like |`|), or
	which must be specially generated in expanded code lines (such
	as braces generated with {\tt\cs{\char123 }}), we have a
	`|\CODE|-' macro that generates an explicit character token with
	catcode~12.

\code
	|\begingroup\catcode`\/=0 \catcode`\\=\other /toksa={/endgroup
	|    /def/CODEbslash{\}
	|}/the/toksa
\smallbreak
	|\begingroup\catcode`\^=\other \toksa={\endgroup
	|    \def\CODEcircum{^}
	|}\the\toksa
\smallbreak
	|\begingroup\catcode`\~=\other \toksa={\endgroup
	|    \def\CODEtilde{~}
	|}\the\toksa
\smallbreak
	|\begingroup\catcode`\%=\other \toksa={\endgroup
	|    \def\%{%}
	|}\the\toksa
\smallbreak
	|\begingroup\catcode`\~=\other \toksa={\endgroup
	|    \def\CODEtilde{~}
	|}\the\toksa
\smallbreak
	|\begingroup\catcode`\[=1 \catcode`\]=2 
	|        \catcode`\{=\other \catcode`\}=\other 
	|        \toksa=[\endgroup
	|    \def\CODElb[{] \def\CODErb[}]
	|]\the\toksa
\endcode
%}}}   catcode hacks
%{{{   customization
\subsec{Customization}

	The token list register |\everycode| is expanded at the start of
	every code section, the token list register |\everycodeline| at
	the start of every code line.  The default definitions select
	typewriter font for the code listings, with each line numbered.
\code
	|\newtoks\everycode
	|\newtoks\everycodeline
\smallbreak
	|\everycode{\tt \baselineskip=0.8\baselineskip \lineskiplimit=0pt }
\smallbreak
	|\newcount\lineno
	|\everycodeline{\llap{$ \global\advance\lineno1 
	|        \scriptstyle\the\lineno $\enspace}%
	|        \mark\expandafter{\the\lineno}}
\endcode
%}}}   customization
%}}}  code sections
%{{{  formating
%{{{   intro
\section{Layout}

	The PDCCODE macros include enough basic formatting commands to
	make a simple document---since this file needs to be
	self-contained. The formatting is fairly primitive, so that
	|pdccode.tex| will not be excessively long.

	The text is set in the default plain \TeX\ fonts: 10-point
	Computer Modern Roman.  Titles are in CM Sans Serif Bold
	Extended, headings in CMSS Demibold Condensed, and subheadings
	in CMSS Italic: all of these are standard fonts which should be
	available on all \TeX\ systems.

	The line width is wide enough that code lines do not need to be
	printed in a reduced font (which looks silly).

	the first tokenm of a document is |\relax| and everything except
	the initial |\input| and final |\bye| in the document is
	enclosed in |\document|\dots|\enddocument|.  This is so that
	documents using the PDCCODE conventions can be embedded within a
	another document which by redefining these tokens.
\code
	|\let\document\relax \let\enddocument\relax
\endcode

	I have to set some \TeX~3 parameters because we don't seem to
	have definitions for them in the local version of |plain.tex|.
\code
	|\lefthyphenmin=3 \righthyphenmin=2 \errorcontextlines10000
\endcode

%}}}   intro
%{{{   layout
\subsec{Layout}

	The text width is wide enough for 72-character code lines.    
\code
	|\parindent=2em \hsize=36\fontdimen6\tentt \advance\hsize\parindent
\endcode
	A largish left margin is used set---perhaps readers will find it
	useful for writing comments.
\code	
	|\hoffset=210mm \advance\hoffset-\hsize \advance\hoffset-25mm
	|\newdimen\leftmargin \leftmargin=\hoffset \advance\leftmargin-25mm
	|\advance\hoffset-1in
\endcode

	The vertical margins are set so that if the file is formatted
	for A4 ($297\mm$) paper and fed into a driver which thinks it is
	formatted for American Letter ($11\inch$) paper no information
	will be lost.  We allow for $7\mm$ unprintable area at the edges
	of paper (this is reasonable for a laser printer), and for the
	headline being $24\pt + 6\pt - 10\pt = 20\pt$ above the top of
	the text.  (The $6\pt$ allows for the height of the headline,
	which is in 7-pt type.)
\code
	|\voffset=297mm \advance\voffset-11in 
	|\advance\voffset7mm \advance\voffset20pt
	|\vsize=297mm \advance\vsize-2\voffset \advance\vsize-2\baselineskip
	|\advance\voffset-1in
\endcode
	
	We use ragged-right (ranged-left) printing.  (Readability
	studies have indicated that ragged-right printing is, if
	anything, easier to read than with both sides aligned.)  To do
	this we remove the stretch and shrink components of the font we
	use, and make the right margin stretchable.
\code
	|\def\do#1{\fontdimen3#1=0pt \fontdimen4#1=0pt }
	|\do\tenrm \do\tenit \do\tenbf
\smallbreak
	|\rightskip=0pt plus 2em minus 2pt \adjdemerits=0 \tolerance=500
	|\parfillskip=0pt plus 0.8\hsize
\endcode
%}}}   layout
%{{{   title and headline
\subsec{Title and headline}

	The author and title info are printed at the top of the first
	page---in lieu of a separate title page---and the title goes in
	the headline of subsequent pages, followed by the version ID and
	date as extracted from the RCS identification string.  If the
	file locker is non-empty (i.e., the file is being edited), an
	`e' is appended to the version ID in the headline.
\code
	|\def\author#1{\def\theauthor{#1}} \author{P. Damian Cugley}
\smallbreak
	|\font\titlefont=cmssbx10 scaled \magstep3
\smallbreak
	|\def\title#1{
	|    \vglue2\baselineskip
	|    \begingroup \parindent-\leftmargin \hyphenpenalty10000 \titlefont 
	|	#1\smallskip
	|    \endgroup
	|    \leftline{Version \fileversion, 
	|        last changed \fileauthor\space \mungedate\filedate}
	|    \ifx\filelocker\empty\else
	|        \leftline{Edited by \filelocker\unskip\ \today}
	|    \fi
	|    \medskip 
	|    \hrule \vskip1pt \hrule \vskip1pt \hrule \vskip1pt \hrule 
	|    \medskip
	|    \rightline{\theauthor}
	|    \bigskip
	|    \headline{\hfil\global\headline{%
	|        \sevenbf \hskip-\leftmargin \themarks \hfil #1 
	|        \fileversion
	|        \ifx\filelocker\empty  
	|            \space\mungedate\filedate
	|        \else 
	|            e\space\today
	|        \fi
	|    }}
	|}
\smallbreak
	|\def\themarks{%
	|    \ifx\firstmark\empty\else
	|        \firstmark
	|        \ifx\firstmark\botmark \else --\botmark \fi
	|    \fi
	|}
\endcode

	Some macros for printing the date.  The |\mungedate| macro
	converts from RCS's format (using slashes) to ISO format (using
	short dashes).
\code
	|\def\today{\mungedate{\number\year/\twodigits\month/\twodigits\day}}
	|\def\twodigits#1{\ifnum#1<10 0\fi \number#1}
	|\def\mungedate#1{$ \textfont0=\font \mathcode`/="7B #1 $}
\endcode

%}}}   title and author
%{{{   sectioning
\subsec{Sectioning}

	The sectioning commands do no more than generate numbered
	headings---there are no tables of contents or marks required.
\code
	|\font\secfont=cmssdc10 at 12pt
	|\font\subsecfont=cmssi10
\smallbreak
	|\newcount\seccount
	|\newcount\subseccount
\smallbreak
	|\def\section#1{%
	|    \advance\seccount1 \subseccount=0
	|    \bigbreak
	|    \begingroup
	|        \secfont \hyphenpenalty=10000 \tolerance=1000
	|        \rightskip=0pt plus 2em
	|        \noindent\llap{\number\seccount.\enspace}#1
	|        \medskip
	|    \endgroup
	|    \noindenttrue
	|}
\smallbreak
	|\def\subsec#1{%
	|    \advance\subseccount1
	|    \medbreak
	|    \begingroup
	|        \subsecfont \hyphenpenalty=10000 \tolerance=1000
	|        \rightskip=0pt plus 2em
	|        \noindent\llap{\number\seccount.%
	|            \number\subseccount.\enspace}#1
	|        \nobreak\smallskip
	|    \endgroup
	|    \noindenttrue
	|}
\endcode

	Here's a simple hack to remove the indentation of the first
	paragraph following a section heading.
\code
	|\newif\ifnoindent
\smallbreak
	|\everypar{%
	|    \ifnoindent
	|        {\setbox0=\lastbox}\global\noindentfalse
	|    \fi
	|}
\endcode
	
%}}}   sectioning
%{{{   footnotes
\subsec{Footnotes}

	Plain \TeX's footnote command is adjusted to use a new parameter
	to specify the value of |\rightskip| in footnotes.\footnote*{The
	default definition sets |\rightskip| to $0\pt$, which is
	moderately disastrous given that spaces cannot stretch or
	shrink.}  It was the trickiness involved in changing definitions
	like this that persuaded me to not use `|@|' characters in macro
	files.
\code
	|\newskip\footnoterightskip \footnoterightskip=\rightskip
\smallbreak
	|\catcode`@=11
	|\def\vfootnote#1{\insert\footins\bgroup
	|  \interlinepenalty\interfootnotelinepenalty
	|  \splittopskip\ht\strutbox % top baseline for broken footnotes
	|  \splitmaxdepth\dp\strutbox \floatingpenalty\@MM
	|  \leftskip\z@skip \rightskip\footnoterightskip 
	|  \spaceskip\z@skip \xspaceskip\z@skip \rm
	|  \textindent{#1}\footstrut\futurelet\next\fo@t}
	|\catcode`@=11
\endcode
	This definition is a drop-in replacement for the one in {\it\TeX
	book} Appendix~B (p.\thinspace363 in my copy).

%}}}   footnotes
%{{{   abbrevs
\subsec{Miscellaneous abbreviations} 

	The |\<| macro for writing the names of syntactic entities (as
	\<foo>) is like in {\it\TeX book} Appendix~E.  The |\cs| macro
	prints a control sequence name; I use this where the easier
	dodge of writing "|\foo|" won't work (e.g., in the argument to
	another macro).  Similarly |\arg| places its argument in
	typewriter-style braces.  The |\pt| macro works in maths mode.
\code
	|\def\<#1>{\leavevmode\hbox{$\langle$\it#1\/$\rangle$}}
	|\def\cs#1{\leavevmode\hbox{\tt\char`\\#1}}
	|\def\arg#1{\leavevmode\hbox{{\tt\char`\{}#1{\tt\char`\}}}}
	|\def\pt{\,{\rm pt}} \def\em{\,{\rm em}} \def\mm{\,{\rm mm}}
	|\def\inch{''}
\endcode

	To make a version of the \LaTeX\ logo that should work
	acceptably with a range of fonts ({\it\LaTeX}, {\bf\LaTeX},
	{\subsecfont\LaTeX}), I~first define a macro for raising letters
	until they are flush with the prevailing cap height.  I also
	loosen up the `{\it\TeX}\/' logo (so that it also works with
	sanserif fonts) and make it set the spacefactor to 1000 (to
	override \TeX's assumption that a full stop following a capital
	$X$ isn't the end of a sentence).
\code	
	|\def\flushtop#1{%
	|    \leavevmode
	|    \begingroup
	|        \setbox0=\hbox{#1}\setbox2=\hbox{X}%
	|        \dimen0=\ht2 \advance\dimen0-\ht0
	|        \raise\dimen0 \box0
	|    \endgroup
	|}
\smallbreak
	|\def\TeX{T\kern-0.1em \lower0.5ex \hbox{E}X\spacefactor1000 }
	|\def\LaTeX{L\kern-0.1em \flushtop{a}\TeX}
\endcode

	Occasional paragraphs are written in italics and marked with
	`note'.
\notepar
	The command says `notepar' not just `note' because in my macro
	files I~use \/|\note| for numbered footnotes (\/{\rm \TeX book},
	Exercise~15.12).\footnote*{Footnotes in |dtx| files are produced
	with the plain |\footnote| command.}

	Notes may occasionally be more than one paragraph long.
\endnotepar
\code
	|\def\notepar{
	|    \smallskip
	|    \begingroup
	|    \leftskip2\parindent \advance\rightskip\parindent minus \parindent
	|    \noindent\kern-\parindent{\bf Note}\quad\it
	|    \ignorespaces
	|}
\smallbreak
	|\def\endnotepar{
	|    \smallbreak
	|    \endgroup
	|}
\endcode
%}}}   abbrevs
%}}}  formatting

\bigskip\noindent
	The file is finished with an |\endinput| command; this can be
	used by other prgorams to check that this file hasn't been
	truncated by \TeX\ coming to a halt part-way through processing
	the |dtx| file.
\code
	|\endinput
\endcode
\endcodefile
\enddocument
\bye
%}}} pdccode
	
%Local variables:
%fill-prefix: "	"
%fold-folded-p: t
%End: