%		DEPROCDOC.TeX
%
%  This file is the preliminary version of a talk given at the Fall '85
%  DECUS in Anaheim by Barbara Beeton, American Mathematical Society.
%  It describes the method by which an author of an article for the DECUS
%  Proceedings can prepare his camera copy using TeX.
%
%  This file, when run through TeX, will be formatted as described below.
%  It uses the macro package DEPROC.TeX plus a few locally defined macros,
%  all of which are given at the top of the file.
%
%  One page in the output will be blank, reserved for a table.  The
%  contents of the table are in the file DEPROCDEV.TeX, which is to be
%  run through TeX as a separate job.
%
%  Although essentially complete, and to the best of my knowledge correct,
%  some details can be expected to change prior to publication.
%
%  Barbara Beeton
%  American Mathematical Society
%  P O Box 6248
%  Providence, RI 02940
%    (401)272-9500
%
%  5 Dec 85


\input deproc

%\SpeechFormat			% this will print smaller (one-column) pages
				% in larger type for reading aloud

%	macros needed for this article

\define\AMS{American Mathematical Society}
\define\Proc{Proceedings}
\define\DP{{\tt DEPROC}}
\define\DProc{{\sl DECUS \Proc}}
\define\POBox{P.\thinspace O.~Box }
\define\TB{{\sl The \TeX book}}
\define\TUB{TUGboat}
\define\TUG{\TeX\ Users Group}
\define\VAX{\leavevmode\hbox{V\kern-.12em A\kern-.1em X}}

\define\<#1>{$\langle${#1}$\rangle$}
%\define\cs{\<cs>}
%\define\css{\cs es}
\define\cs{{\tt cs}}
\define\css{\cs-es}
\define\CR{\<\smc cr>}
\define\tab{\<\smc tab>}
\def\ttvert{{\tt\char'174 }\ignorespaces}

\title{Typesetting Articles for the DECUS
    Proceedings with \TeX\footnotemark[1]}
\author{Barbara N. Beeton\\
    \AMS\\
    Providence, Rhode Island}

\begin{abstract}
The \DProc\/ have traditionally been published from copy supplied by
the authors, prepared according to rules devised for typewritten
material.  The power of the computer typesetting language \TeX\ has
now been applied to this task, and a formatting package, named \DP,
has been submitted to the DECUS Program Library for use by authors who
have access to a working \TeX\ system.  (The \TeX\ program and related
software, created by Donald Knuth of Stanford, are in the public domain.)

This paper presents the important features of \DP\ and, through
examples, shows how it is to be used.  Use of \DP, which is
encouraged, will produce the author's work, nicely typeset, in the
standard {\sl\Proc\/} format.  There is a general description
of how the package works and of the mechanical requirements for camera
copy of {\sl\Proc\/} articles, which will be created on the
author's local output device.

No prior knowledge of \TeX\ is required, but authors using \DP\ will
be expected to learn some rudiments, especially if their papers
contain special notation or formats such as tables.
\end{abstract}

\begin{document}

\maketitle

\footnotetext[1]{\TeX\ is a trademark of the \AMS.}

The \DProc, like the conference proceedings of many other organizations,
is rushed to publication as quickly as possible so that the material
will reach the conference participants and other interested readers
before its value is diminished by time.  Reproducing author-prepared
copy eliminates the considerable bother and expense of typesetting,
proofreading and corrections.  The published document should be
compact, uniform in appearance, and readable, regardless of the kind
or quality of printing device available to the author.  For these
reasons, instructions to authors have heretofore assumed that nothing
more elaborate is available than an ordinary typewriter or dot matrix
printer.

To enforce uniformity, the author is provided with ``model paper'', on
which are printed (in non-reproducing ink) column and page borders,
alignment marks, and instructions for placement of title, author, and
the other parts of a proceedings article.  The dimensions of the model
paper are almost always larger than those of the published \Proc\Dash
this permits more text to be packed onto each page, and also improves
its appearance or ``quality'' when photographically reduced, smoothing
out the rough edges of letters and symbols generated by a typewriter,
dot-matrix printer or other ``low-resolution'' device.

Within the past few years, advances in laser-printer technology have
made good-quality output accessible to a growing number of users,
through a widening selection of low-cost output systems based on print
engines with 300 dot-per-inch resolution and (relatively) easy-to-use
interfaces.  Such devices have been attached to most kinds of DEC
computers, and drivers now exist to print the output from such programs
as Scribe\footnote[2]{Scribe is a trademark of Unilogic Ltd.}
\TeX\ and Troff.  Most low-end laser printers cannot use paper wider
than $8\frac1/2''$, however, so even if both a good composition
program and output printer had been available, until now an author
would have been discouraged from using them for mechanical reasons.

The editor of the \DProc\/ has now agreed to accept typeset copy
printed on such a system at 100\% on $8\frac1/2\times11''$ paper,
provided it conforms to the published format.  This article (which has
itself been produced by the technique it describes) introduces a
package, \DP, designed to prepare {\sl\Proc\/} articles using \TeX.


\section What is \TeX?

\TeX\ is a public-domain typesetting language created by Donald Knuth
of Stanford University.  His original aim was to typeset his own books,
in particular {\sl The Art of Computer Programming\/} \bibref{ACP}, with
a quality equal to that produced by the best traditional composition
methods.  The technical content of these books assured that full
attention was given to the niceties of formatting mathematical
expressions, as well as to the structures of documents commonly
encountered in technical publishing.

\TeX\ deals with low-level concepts familiar to typesetters\Dash type
size, leading, interword spacing and kerning.  It does not incorporate
directly the structures an author encounters when writing a paper\Dash
title, figure references, bibliographic entries.  However, \TeX\ is
essentially a macro compiler, and provides a full vocabulary of low-level
functions that can be manipulated by knowledgeable users to create 
higher-level packages to support the casual user.

Two such macro packages are \LaTeX\ and \AMSTeX.  \LaTeX\ \bibref{LT} is
a powerful document formatter, providing the capability to format
books and reports, with functionality similar to that provided by
Scribe.  \AMSTeX\ \bibref{Joy} supports only one format (for preprints),
but it incorporates complex formatting instructions for aligned
displays, theorems and proofs, and other structures which occur in
mathematical research papers.  (Both of these macro packages will be
made available through the DECUS Program Library.)

\TeX\ actually comes with its own macro package\Dash |PLAIN|\null.  |PLAIN|
contains general formatting instructions to produce ``report-style''
output (one $6\frac1/2''$-wide column to be printed on letter-size
paper, with page number centered at the bottom and macros for
creating simple running heads).  Most other macro packages, including
\DP, use many of |PLAIN|'s techniques, either by incorporating
the code explicitly, or by making the assumption (as \DP\ does) that
|PLAIN| is preloaded.

A good introduction to |PLAIN| and to \TeX\ itself is {\sl First Grade
\TeX\/} \bibref{FG}, by Arthur Samuel.  The whole story can be found in
\TB\/ \bibref{TB}, by Donald Knuth.


\reservefigurepages[1]{Table of \TeX\ computer/output device combinations}

\section The \DP\ macro package

In order to use the \DP\ macro package, the author of a
\DProc\/ article must have available a working \TeX\ system.
\TeX\ has been implemented on \VAX es and DECsystem-10s and -20s
under the standard operating systems.
There is also a good selection of output devices available, capable
of production output of quality suitable for the {\sl\Proc\/}; Table~1
shows the computer/output device combinations known to the \TUG.
(\TeX\ has not, however, been implemented on {\ninerm}PDP-11s, since
it requires a larger address space than is supported on those machines.)

Although most of the document structures (section headings, figures, 
and the like) needed for the \DProc\/ are already supported in \LaTeX,
\LaTeX\ is not available to all \TeX\ users.  (\TeX\ is a very large
program by itself, and routinely adding a large macro package can put 
unwelcome strain on an already overloaded machine.  Some system
administrators prefer not to give their users that opportunity.)
For that reason, the first implementation of \DP\ will not require
\LaTeX, and it will not automatically do all the useful things that
\LaTeX\ users have come to expect.  But since the \LaTeX\ input
language has been so well designed, the same vocabulary and syntax
will be used whenever practicable.  Later, I intend to create a
\LaTeX-based style file, so users will have a choice.

The first version of this macro package to be distributed
is called |DEPROC.TeX|, for ``\DProc\/ TeX definitions''.  It is
an ordinary ASCII file, and is on the latest DECUS Program Library
tapes for Languages \& Tools, Large Systems, and \VAX.


\section Some preliminary \TeX nical information

An author who intends to use \DP\ should preferably have used
\TeX\ already, either in its |PLAIN| wrapper or through some macro
package.  Nonetheless, a few basic concepts are worth repeating.

\subsection Spacing

\TeX\ uses different spacing rules in text (paragraphs) and math.
Paragraphs are set so that interword spacing is as uniform as possible.
Wider spaces are set after punctuation that indicates the ends of
sentences (period, !\ and ?).  Within math, the best traditions for
arranging symbols in two dimensions, including proper spacing, are
observed.  Thus input spacing is largely ignored, except
for its functions of separating words and marking the boundaries of
certain kinds of expressions.  \TeX\ considers
multiple spaces in an input file to be equivalent to a single space.
The carriage return \CR\ and the tab character \tab\ are equivalent
to ordinary spaces, except in special environments (noted below).
And all spaces at the beginning of any line are ignored.

\subsection Paragraph breaks

A blank line in the input file indicates a paragraph break.  A line is
blank if it contains only a \CR\ or spaces and a \CR.  Multiple blank
lines are equivalent to a single blank line.  (A paragraph can also be
indicated by |\par|; terms beginning with |\| are described below.)

\subsection Comments

A comment may be entered on any line; a comment begins with a |%| sign:
\begintt
%  This line contains nothing but a comment.
\def\cs{...}%  explanatory comment
... Smythe % ***** check spelling *****
\endtt
\TeX\ will ignore the |%| and everything following it, including the \CR.
Thus, the space ordinarily indicated by the \CR\ will be suppressed,
and if a space is really wanted between the last item before a comment
and the first item on the next line, it must be input before the |%|.
Conversely, if no space is wanted between the last item on a line and
the first item on the next, a |%| can be used to suppress it intentionally.

\subsection Control sequences, also called macros

A ``control sequence'' \cs\ is an instruction for \TeX\ to perform
some action or to produce a particular symbol.  A \cs\ begins with a
backslash, |\|.  There are two types of \css:
\item[--] A ``control word'' consists of |\| followed by one or more
	letters.  It is terminated by any non-letter, including a space;
	multiple spaces follow the usual compression rule, so a
	special technique (see next paragraph) is required to create an
	output space after a control word.  |\TeX| is an example of
	a control word; it produces the \TeX\ logo.
\item[--] A ``control symbol'' consists of |\| followed by exactly one
	non-letter.  Since its length is known, no special terminator
	is required.  |\&| is a control symbol to produce an \&.
	{\tt\bs\SPC} (|\| followed by a space) is an explicit space, to
	be used where an output space should follow an element
	input as a control word.

\noindent
New \css\ can be defined within a document to make input easier or clearer.
A few rules governing \cs\ names should be observed carefully.
\item[--] Case matters; |\csname| is not the same as |\Csname| or |\CSName|.
	Try to pick a name that means something to you, and is easy to type.
\item[--] A new definition will replace an existing one.  If you specify
	a \cs\ name that performs an important function in the document
	formatting, results, as they say ``may be unpredictable''.

\smallskip
\noindent
You can use the |PLAIN| method of defining a \cs:
\begintt
\def\csname{...something...}
\endtt
but if you are not really familiar with \TeX\ or with \DP, you might
choose a |\csname| that already exists.  \DP\ provides a simple
alternative, that checks to make sure your |\csname| isn't already
reserved for something else:
\begintt
\define\csname{...something...}
\endtt
If this name has been used before, \TeX\ will stop to give you a
warning, which you may ignore by responding with a \CR, after which the
definition will be made as requested.  You may continue to ignore the
warning, if you know the prior use won't affect you, but it's usually
better to change |\csname| to avoid the interruption.

The control symbols |\0,...,\9| always start out undefined, so they are
available for transient use without checking.

\subsection Math

Mathematical expressions are input between |$...$|.  Display math is begun
and ended with |$$...$$|.  For details of math input, see \bibref{FG},
\bibref{TB} or \bibref{Joy}, in order of increasing complexity of
expressions handled.


\section Starting a \DProc\/ article

The first step in preparing an article is to create a file.
The first line in this file should be
\begintt
\input deproc
\endtt
This will cause the formatting definitions to be loaded when the file
is input to \TeX.

Next, enter the ``top matter''.  This consists of such
things as the title of the article, the author(s) and their addresses,
and the abstract.

\subsection Title and authors

For an article with a short title and one author, they are input like this:
\begintt
\title{A One-Line Title}
\author{Author Name\\
    Author's Organization\\
    City, State}
\endtt
The double backslashes |\\| indicate line breaks.  This technique is
also used to break up long titles:
\begintt
\title{Here We Have a Particularly
    Long Title\\That Can't Possibly
    Fit on a Single Line}
\endtt
This will be set (in a boldface font slightly larger than text size) as\strut
\par{\vskip\abovedisplayskip
	\raggedstretch=.2\hsize \raggedcenter \bf
	Here We Have a Particularly Long Title\break
	That Can't Possibly Fit on a Single Line\endgraf
\vskip\belowdisplayskip}
\noindent
Notice that the way the lines are broken in the input file is not how
they appear in the output\Dash only |\\| matters to \TeX.
Actually, \TeX\ will break long titles into lines short enough to fit
on the page, but a multi-line title usually makes more sense to the reader
if the author decides where the line breaks should occur.

For multiple authors, the same |\author| tag is used with |\and| or |\And|:
\begintt
\author{First Author
  \and
    Second Author\\
    Common Organization\\
    City, State}
\endtt
or
\begintt
\author{First Author\\
    First Organization\\
    City, State
  \And
    Second Author\\
    Second Organization\\
    City, State}
\endtt
and so forth, which will appear thus in the output:\strut
\par{\vskip\abovedisplayskip
	\raggedstretch=.3\hsize \raggedcenter
	{\bf First Author\enspace{\rm and}\enspace Second Author}\break
	Common Organization\break
	City, State\endgraf
\vskip\belowdisplayskip}
\par\noindent
or\strut
\par{\vskip\abovedisplayskip
	\interlinepenalty=10000
	\raggedstretch=.3\hsize \raggedcenter
	{\bf First Author}\break
	First Organization\break
	City, State\break
	\null\hfil\break
	{\bf Second Author}\break
	Second Organization\break
	City, State\endgraf
\vskip\belowdisplayskip}
\noindent
Authors' names (the first line, and the first line after |\And|) are
printed in boldface; if an author name is to appear on any other line,
begin that line with |\bf| (the \TeX\ instruction for boldface type).

The title and author of the present paper look like this in the file:
\begintt
\title{Typesetting Articles for the DECUS
    Proceedings with \TeX\footnotemark[1]}
\author{Barbara N. Beeton\\
    \AMS\\
    Providence, Rhode Island}
\endtt
A couple of additional points can be noted in these few lines.
First is the |\footnotemark[1]| which follows |\TeX|.
Where is the footnote text?  (After all, \TeX\ is supposed to take
care of formatting, leaving the author free to worry about content.)
For \TeX nical reasons too complicated to explain here, footnotes on
items in the top matter are lost unless extraordinary measures are
taken, and I hadn't developed a reliable solution in time to install
it in the first release.  Fortunately, there is a simple alternative,
to enter the footnote marker (as shown here) and the footnote text
separately.  This is covered in more detail in the section on footnotes.

The other item to look at is |\AMS|, which becomes \AMS\ in the output.
This is an example of a ``local definition'', something that is not likely
to be useful to anyone else, but can save the author a lot of time correcting
typing errors.  Local definitions that are used throughout an article are
best input right after the request to load \DP:
\begintt
\input deproc
\define\AMS{American Mathematical Society}
...
\endtt

\subsection Abstract

The abstract is the final part of the top matter.
\begintt
\begin{abstract}
This is a short summary of what
the article is about.
\end{abstract}
\endtt
\par\noindent
The heading ``{\bf Abstract}'' is provided automatically; don't input it.
The abstract may contain more than one paragraph.  Paragraphs are
separated by a blank line or by |\par|, as usual.

The top matter is now complete.  The body of the article follows.
\begintt
\begin{document}
\maketitle
|bgroup|rm(|it|0Text of footnotes to the top matter is given here|/|rm)|egroup

This is the first sentence of article text.
...
\end{document}
\endtt


\section The body of the article

An article can start out with text or with a heading.
Three levels of headings are provided by \DP:
\begintt
\section Section heading\par
\subsection Subsection heading\par
\subsubsection Subsubsection heading\par
\endtt
(As usual, a blank line is equivalent to |\par|.)
These produce headings (with extra space above and below, not shown here)
in the following styles:
$$\vbox{\advance\baselineskip by 2pt
\centerline{\bf Section heading}
\leftline{\bf Subsection heading}
\leftline{\it Subsubsection heading}}
$$

The first paragraph following a heading will not be indented in the
default style.  This may be changed if you prefer, so that all
paragraphs will be indented, by specifying
\begintt
\NormalParIndent
\endtt
on the line after |\begin{document}|.

Paragraph indentation can be suppressed throughout by specifying
|\NoParIndent| but that would probably make the article too hard to read
without other alterations to the style; for an example of unindented
style, see the article by Richard Southall \bibref{TD} in \TUB.

To suppress indentation on a single paragraph, precede it by |\noindent|.

\subsection Footnotes

The use of |\footnotemark| in a title has already been illustrated.  A
footnote consists of two parts, the mark and the text.  These
are usually entered as a unit\footnote[3]{Like this.}:
\begintt
... as a unit\footnote[3]{Like this.}:
\endtt
\par\noindent
When the marks and text must be entered separately\footnotemark[4,5]%
\footnotetext[4]{Two footnotes assigned to one item must be marked and
entered separately}\footnotetext[5]{This situation is slightly
different from what happens in the top matter.}, two statements are
needed:
\begintt
... entered separately\footnotemark[4,5]%
\footnotetext[4]{Two footnotes ...
... separately.}\footnotetext[5]{This ...}
\endtt
\par\noindent
Note the use of the |%| after the |\footnotemark|\Dash care must be taken
to avoid spaces that might creep unwanted into the text, and suppressing
the \CR\ at the end of the line is probably the easiest way.

Marks are set as superscripts in the default style, but this can be changed
by specifying
\begintt
\footnotemarkstext
\endtt
just after |\input deproc|.

\subsection Quotations

Short quotations, of less than a paragraph, are set with
\begintt
\begin{quote}
If you can't fix it, ... {\em Button}
\end{quote}
\endtt
and look like this:
\begin{quote}
If you can't fix it, call it a feature. \ {\em Button}
\end{quote}

\noindent
For longer quotations, use
\begintt
\begin{quotation}
...
\end{quotation}
\endtt
in a similar manner, separating paragraphs with blank lines as usual.


\subsection Lists

Itemized and enumerated lists occur in many \DProc\/ articles.
\LaTeX\ provides automatic counters and up to four levels of nesting.
\DP\ users will have to make do with two levels, and no automatic
counters.  The general input structure is this:
\begintt
\item[label] text
\itemitem[label] text
\endtt
Any desired labels can be used;
\item[\bullet] |\bullet|,
\item[\circ] |\circ|, and
\item[--] en-dash (input as |--|)
\par\noindent are common choices, as are numerals and single letters.
Here is a short example of a two-level list.
\begintt
\item[\bullet] first item
\item[\bullet] second item
\itemitem[\circ] new level
\itemitem[\circ] one more
\item[\bullet] back a level
\endtt
\par\noindent
Here's what this looks like, after padding out the text a bit to show
how longer items look.
\item[\bullet] The first item in this list isn't particularly interesting,
	but it has to be long enough to make two lines.
\item[\bullet] The second item isn't either.
\itemitem[\circ] Even going to a new level doesn't add very much
	excitement to this exercise.
\itemitem[\circ] We'll do one more at this level.
\item[\bullet] Then we'll go back a level to finish things off.
\smallskip
\noindent
Each item comprises one paragraph; an unlabeled paragraph can be
produced by specifying an empty label.  Extra space above and below a
list can be obtained by specifying |\smallskip| or |\medskip.| It is
usually advisable either to insert extra space after a list or to
apply |\noindent| to the following paragraph, for clarity.  For
example, the sample list above ended this way:
\begintt
...
... finish things off.
\smallskip
\noindent
\endtt


\subsection Figures

Figures come in several sizes and shapes:
\item[--] small figures which can be set in place, i.e., in the same
	relative position where they occur in the input file;
\item[--] one-column figures to be set at the top or bottom of the
	first available column;
\item[--] double-column figures to be set at the top or bottom of the
	first available page;
\item[--] full-page figures.
\smallskip
\noindent
Not all of these formats are supported yet by \DP.  In particular,
two-column figures cannot be incorporated into text pages, although
there is a mechanism for leaving space at the bottom of a page, so
that separately-prepared figures can be pasted in later.  Similarly,
space (i.e.\ blank pages) can be reserved for full-page figures.

\subsubsection One-column figures

To get a single-column, in-line figure, enter
\begintt
\begin{figure}
|MTH|hbox|bgroup|it|0content of figure|egroup|MTH
\caption{Figure |MTH|0n|MTH.| Caption text}
\end{figure}
\endtt
\par\noindent
``Figure'' and the figure number must be input.  (This is not
necessary in \LaTeX, where figures are numbered by a counter reserved
for that purpose.)  If the figure will be prepared separately and
pasted in, space can be reserved:
\begintt
\begin{figure}
\vspace{2.5in}
\caption{Figure |MTH|0n|MTH.| Caption text}
\end{figure}
\endtt
\par\noindent
Space equivalent to a blank line is skipped above and below an in-line
figure, and a half-line between the figure and the caption, so the
dimension given with |\vspace| should be precisely the size of the
item to be pasted in.

If insufficient space remains in the column to accommodate the figure
in-line, it will automatically be shifted to the top of the next
available column.  Figures can be placed at the tops of columns
explicitly by specifying
\begintt
\begin{topfigure}
|    ...
\end{topfigure}
\endtt
\par\noindent
A |topfigure| will be set at the top of the current column, if space
is available, otherwise at the top of the next available column.
When both regular |figure|s and |topfigure|s are used concurrently,
they may possibly be set out of order, depending on their sizes and
the other contents of the article; it is safer to use only one style
or the other within a single article, and even then, to check the
results carefully before sending your article off to be published.


\subsubsection Two-column figures

Double-column figures, as mentioned earlier, cannot yet be
incorporated into text pages by \DP.  However, space can be reserved
for figures prepared separately; see the next section, full-page
figures.

To reserve space, you must know the page number, or the relative
page, on which you want to place the figure; you must also know the
size of the figure, including its caption.  Space will be reserved
by shortening the text area on the specified page, so double-column
figures at present will be placed only at the bottom of a page.

The command to shorten a page must appear on a line by itself.
\begintt
\ShortenPage |MTH|0n|MTH by |MTH|0dimen|MTH.
\endtt
or
\begintt
\ShortenPage +|MTH|0n|MTH by |MTH|0dimen|MTH.
\endtt
(The spaces and the period at the end are necessary.)
The first form would be used if you know the page number; otherwise,
|+|$n$ equals the number of pages {\em after the one on which the command
is given}.  For example, if you give the command on page 1, and the
figure is to appear on page 3, $n=2$.  For {\it dimen}, enter the
height of the figure as measured, plus about $.2''$ for appearance
(extra space is not added, unlike |figure| and |topfigure|).
This height may also be expressed in terms of number of text lines:
\begintt
... by 12\lines.
\endtt
will reserve the same space occupied by 12 lines of text, about $2''$.

|\ShortenPage| will not shorten the first page of an article, and
only one |\ShortenPage| command can be in effect at once.
But a page that is being shortened can also include a command
|\ShortenPage +1| to adjust the next page.


\subsubsection Full-page figures

Full-page figures are partially supported by \DP.  Pages are
reserved during the main \TeX\ run, and figures can be prepared
separately, to replace the reserved pages.  To reserve a page for
a figure, on the page before the one to be reserved (this will most
likely be near the first reference to the figure), enter a line
of the form
\begintt
\reservefigurepages[1]{Replace by Figure N}
\endtt
\par\noindent
Multiple pages may be reserved at once, say for a program listing:
\begintt
\reservefigurepages[3]{Listing of XYPROG}
\endtt
\par\noindent
Each reserved page will contain only a page number and a two-line
message:\strut
\par{\vskip\abovedisplayskip
	\raggedstretch=.3\hsize \raggedcenter
	Page reserved for figures\break
	Text input with |\reservefigurepages|\endgraf
\vskip\belowdisplayskip}

A separate file can be created for figure input, if you wish to use
any of the \DP\ facilities.  It should begin with these three lines:
\begintt
\input deproc
\figurepages
\pageno=|MTH|0nn|MTH
\endtt
The value assigned to |\pageno| should be the page number to appear on
the first figure page; subsequent page numbers can be reset in the
same manner as appropriate.  |\figurepages| will redefine the output
format to be one column $5''$ wide.  This page format is called
|\onenarrow|.  Two other single-column page formats are available:
\item[ ] |\onecol| is the same width as the two-column page;
\item[ ] |\onemedium| is $6''$ wide.
\par\noindent (The normal output format is called |\twocol|.)
Any one of these output formats can be specified just after a page
break, before anything has been set on the new page.  A page break
can be forced by the command |\newpage|.  (In two-column format, a
column break is forced by |\newcol|; in one-column format, |\newcol|
is equivalent to |\newpage|.)

Within the figure input file, most \DP\ options are available for use.
Exceptions are the top matter commands (|\author|, |\title| and
|\abstract|), |\begin{document}| and |\end{document}|,
and |\maketitle|.  Almost all plain \TeX\ facilities can be used,
although it is advisable to check the \DP\ definitions if you
intend to make changes in page format.

When you have completed the file of figure input, end it with the
command |\bye|.


\subsection Tables

There is no support yet in \DP\ for tables.  Tables can be coded
using plain \TeX\ rules for tabbed or |\halign| environments.  See
\TB\/ for details.

\subsection Verbatim

Verbatim items are printed in so-called ``typewriter'' style, using \TeX's
|\tt| font.  In-text verbatim items are enclosed in vertical bars
{\tt\ttvert...\ttvert}; blocks of verbatim code are delimited by
\begintt
|bs|0begintt
...
|bs|0endtt
\endtt
|\begintt| and |\endtt| should be on lines by themselves.  Within verbatim
mode, \CR s are obeyed as line breaks, not spaces.  An input line that is too
long for the current column width will be broken at a space if possible,
and the remainder of the line hanging indented on the next output line; 
since this may change the meaning of the verbatim passage, such passages
should be checked with special care in the output.  Overlong lines also
frequently result in overfull |\hbox|es, which are indicated clearly
on the output by black boxes:
{\vrule height\ht\strutbox depth\dp\strutbox width 5pt }\thinspace.

A passage between |\begintt...\endtt| is treated as a unit by \TeX\Dash
if it is too long for the vertical space available, it either will be
carried over as a unit to the next column or page,
or will result in an overfull |\vbox|, which will be noted only in the
transcript of the \TeX\ run.  In such a case, the best remedy is to
break the passage in two, by inserting another |\endtt \begintt|.

Verbatim mode is suitable for program listings, indicating keyboarding
instructions, file names, and similar uses.

\subsection References, bibliography

References in text to items in the bibliography are input as
\begintt
\bibref{|MTH|0label|MTH}
\endtt
where the label that will be used in the bibliography must be entered;
there is no automatic association between the two as in \LaTeX.  For
a label ``ABC'', the text reference will be rendered \bibref{ABC}.

Before you start to input the reference list, some housekeeping is
required\Dash you must decide what you want the list to look like.
This is what the input looks like for one of the items in the reference
list at the end of this article:
\begintt
\bibitem[TB] Knuth, Donald E., \TB,
    Addison-Wesley and \AMS, 1979.
\endtt
(|\TB| and |\AMS| are among the local definitions for this article.)
Default output looks like this:
\smallskip
{\SetUpBiblio{TB}
\bibitem[TB] Knuth, Donald E., \TB, Addison-Wesley and \AMS, 1979.\endgraf
}
\smallskip

If you prefer the reference labels to be enclosed in parentheses
instead of square brackets, |\bibbrackets()| will replace them.
(|\bibbrackets[]| will restore the square brackets.)
Brackets may be eliminated by |\omitbibbrackets|.  If your labels are
simply numbers, $1,\dots,n$, brackets will be omitted automatically
(see below).

Now give the command to begin the reference list
\begintt
\Bibliography{|MTH|hbox|bgroup|it|0widest label|egroup|MTH}
\endtt
entering the widest label that will actually be used in the list;
this will be used to control the formatting.

If you do not wish to use labels, substitute |\omit| for the widest
label.  (The |[...]| are still required in the context of |\bibitem|.)
\begintt
\bibitem[] Knuth, Donald E., ...
\endtt
will result in
\par\nobreak\smallskip
{\SetUpBiblio{\omit}
\bibitem[] Knuth, Donald E., \TB, Addison-Wesley and \AMS, 1979.\endgraf
}
\smallskip

If your labels are numeric, substitute |\numeric| for the widest label.
The input
\begintt
\bibitem[2] Knuth, Donald E., ...
\endtt
will now look like this:
\par\nobreak\smallskip
{\SetUpBiblio{\numeric}
\bibitem[2] Knuth, Donald E., \TB, Addison-Wesley and \AMS, 1979.\endgraf
}


\section Caveats

\DP\ and this article were created on a DECSYSTEM-20 at the
\AMS, running \TeX\ version~1.5.  The AMS installation is standard
in all ways except that a few memory cells have been increased for
reasons not relevant to, and not affecting the performance of, \DP.

With one exception, none of the changes to the \TeX\ program since
version~1.0 should have any noticeable effect on an article produced
with \DP.  The exception is large, complex tables\Dash
tables incorporating many boxes and rules require large amounts of
\TeX\ memory.  Memory management was radically changed in version~1.3
to make more memory available to the user without actually changing
the physical memory allotment.  (Otherwise, if you run out of memory,
the most likely cause is an input error.)

Although thorough testing has been attempted, no one outside the AMS has
tried to use \DP\ yet, so bugs are sure to be found.  In fact,
the version of \DP\ first placed in the Program Library should best
be considered a beta test version.  If you find a bug, please
communicate it to the author, accompanied by an example which
demonstrates the bug as simply as possible.  Suggestions for
improvements are also welcome.  Send everything to\break
{\obeylines %
	\indent Barbara Beeton
	\AMS
	\POBox 6248
	Providence, RI 02940
}


\Bibliography{ACP}

\bibitem[ACP] Knuth, Donald E., {\sl The Art of Computer Programming},
	Addison-Wesley, Vol.~2, second edition, 1981.

\bibitem[TB] Knuth, Donald E., \TB, Addison-Wesley and \AMS, 1979.

\bibitem[LT] Lamport, Leslie, {\sl L\kern-.36em\raise.3ex\hbox{\sixi A}%
	\kern-.15em\TeX, A document preparation system}, Addison-Wesley,
	1985.

\bibitem[TD] Southall, Richard, First principles of typographic design
	for document production, {\sl\TUB\/} Vol.~5 (1984), No.~2, 79--90;
	Corrigenda, Vol.~6 (1985), No.~1, p.~6.

\bibitem[FG] Samuel, Arthur, {\sl First Grade \TeX}, \TUG, 1984.

\bibitem[Joy] Spivak, Michael, {\sl The Joy of \TeX}, \AMS, 1980;
	new edition in preparation, 1985.

\bibitem[TUB] {\sl\TUB, the Newsletter of the \TUG}, \TUG, \careof\AMS,
	\POBox 9506, Providence, RI, 02940.



\end{document}

