This manual documents the Eplain macros, version 3.2, November 2007. Eplain provides functionality for plain TeX that is intended to be useful regardless of how your document is actually formatted.
Most of this manual is in the public domain, like most of the Eplain code. It was originally written by Karl Berry, starting in 1989. Steven Smith wrote the documentation for the commutative diagram macros; this chapter is under the GNU General Public License. Adam Lewenberg has made additions and corrections. Oleg Katsitadze wrote the section on LaTeX packages and the chapter on hyperlinks, and updates throughout.
--- The Detailed Node Listing ---
User definitions
Citations
Displays
Lists
Contents
Cross-references
Equation references
Indexing
Indexing terms
Loading LaTeX packages
Hyperlinks
Implicit hyperlinks
Index hyperlinks
Hyperlink drivers
Hyperlink driver hypertex::
Hyperlink drivers pdftex and dvipdfm::
Setting hyperlink types and options
Turning hyperlinks on/off
Arrow theoretic diagrams
Commutative diagrams
Programming definitions
Expansion
Demo files
The Eplain macro package expands on and extends the definitions in plain TeX. This manual describes the definitions that you, as either an author or a macro writer, might like to use. It doesn't discuss the implementation; see comments in the source code (xeplain.tex) for that.
Eplain is not intended to provide “generic” typesetting capabilities, as do LaTeX (written by Leslie Lamport) or Texinfo (written by Richard Stallman and others). Instead, it provides definitions that are intended to be useful regardless of the high-level commands that you use when you actually prepare your manuscript.
For example, Eplain does not have a command \section, which
would format section headings in an “appropriate” way, such as
LaTeX's \section. The philosophy of Eplain is
that some people will always need or want to go beyond the macro
designer's idea of “appropriate”. Such canned macros are
fine—as long as you are willing to accept the resulting output. If
you don't like the results, or if you are trying to match a different
format, you are out of luck.
On the other hand, almost everyone would like capabilities such as cross-referencing by labels, so that you don't have to put actual page numbers in the manuscript. The author of Eplain is not aware of any generally available macro packages that (1) do not force their typographic style on an author, and yet (2) provide such capabilities.
Besides such generic macros as cross-referencing, Eplain contains another set of definitions: ones that change the conventions of plain TeX's output. For example, math displays in TeX are, by default, centered. If you want your displays to come out left-justified, you have to plow through The TeXbook to find some way to do it, and then adapt the code to your own needs. Eplain tries to take care of the messy details of such things, while still leaving the detailed appearance of the output up to you.
Finally, numerous definitions turned out to be useful as Eplain was developed. They are also documented in this manual, on the chance that people writing other macros will be able to use them.
You can send bug reports or suggestions to
tex-eplain@tug.org. The current version number of Eplain is
defined as the macro \fmtversion
at the end of the source file eplain.tex. When corresponding,
please refer to it.
To get on this mailing list yourself, email tex-eplain-request@tug.org with a message whose body contains a line
subscribe you@your.preferred.address
or visit http://tug.org/mailman/listinfo/tex-eplain.
David Walden had reported his experience with Eplain as a new user. The article is available online at http://tug.org/pracjourn/2005-4/walden. An introductory article (written for TUGboat) is also available online at http://tug.org/eplain/misc/tb84katsi.pdf.
Your TeX installation should already contain a version of Eplain (eplain.tex) in its main texmf tree (usually under /usr/share/texmf/tex/eplain/ on Unix systems). To install a newer version of Eplain, put the new eplain.tex (included in Eplain distributions) in the tex/eplain/ subdirectory of your local texmf tree. The newer version you install in the local tree should override the older one in the main tree.
The location of the local texmf tree obviously depends on your operating system and TeX installation. On Unix systems the usual location is /usr/local/share/texmf/. If you don't have write permissions for /usr/local/share/texmf/, many installations read the texmf tree in the user's home directory; eplain.tex then should go under ~/texmf/tex/eplain/. For more information about TeX directory structure, please see http://www.tex.ac.uk/cgi-bin/texfaq2html?label=tds.
If you prefer to install eplain.tex in a non-standard place, set
an environment variable (TEXINPUTS for the Web2C port of TeX
to Unix) to tell TeX how to find it.
If you want, you can also create a format (.fmt)
file for Eplain, which will eliminate the time spent reading the macro
source file with \input. You do this by issuing a sequence of
Unix commands something like this:
prompt$ touch eplain.aux
prompt$ initex
This is TeX, ...
**&plain eplain
(eplain.tex)
*\dump
... messages ...
You must make sure that eplain.aux exists before you run initex; otherwise, warning messages about undefined labels will never be issued.
You then have to install the resulting eplain.fmt in your local
texmf tree or set an environment variable to tell TeX how to
find it. For the Web2C port of TeX to Unix, format files are usually
installed under /usr/local/share/texmf/web2c/ or
/var/lib/texmf/web2c/; the environment variable is
TEXFORMATS.
The simplest way to use Eplain is simply to put:
\input eplain
at the beginning of your input file. The macro file is small enough that reading it does not take an unbearably long time—at least on contemporary machines.
In addition, if a format (.fmt)
file has been created for Eplain (see the previous section), you can
eliminate the time spent reading the macro source file. You do this
by responding &eplain to TeX's `**' prompt. For
example:
initex
This is TeX, ...
**&eplain myfile
Depending on the implementation of TeX which you are using, you might also be able to invoke TeX as eplain and have the format file automatically read.
If you write something which you will be distributing to others, you
won't know if the Eplain format will be loaded already. If it is, then
doing \input eplain will waste time; if it isn't, then you must
load it. To solve this, Eplain defines the control sequence
\eplain
to be the letter t (a convention borrowed from Lisp; it doesn't
actually matter what the definition is, only that the definition
exists). Therefore, you can do the following:
\ifx\eplain\undefined \input eplain \fi
where \undefined must never acquire a definition.
Eplain consists of several source files:
\ifpdf, which can be used to
detect pdfTeX in PDF mode (see Checking for PDF output), written by Heiko Oberdiek;
The file eplain.tex is all of these files merged together, with comments removed.
All of these files except xeplain.tex can be input individually, if all you want are the definitions in that file.
Also, since the bibliography macros are fairly extensive, you might not
want to load them, to conserve TeX's memory. Therefore, if the
control sequence \nobibtex
is defined, then the bibliography definitions are skipped. You must set
\nobibtex before eplain.tex is read, naturally. For
example, you could start your input file like this:
\let\nobibtex = t
\input eplain
By default, \nobibtex is undefined, and so the
bibliography definitions are made.
Likewise, define \noarrow if you don't want to include the
commutative diagram macros from arrow.tex, perhaps because you
already have conflicting ones.
If you don't want to read or write an aux file at all, for any
kind of cross-referencing, define \noauxfile
before reading eplain.tex. This also turns off all warnings
about undefined labels.
Eplain conflicts with AMSTeX (to be precise, with amsppt.sty):
the macros \cite and \ref are defined by both.
If you want to use AMSTeX's \cite, the solution is to define
\nobibtex before reading Eplain, as described above.
If you have amsppt.sty loaded and use \ref, Eplain writes
a warning on your terminal. If you want to use the AMSTeX
\ref, do \let\ref = \amsref after reading Eplain.
To avoid the warning, do \let\ref = \eplainref after reading
Eplain and before using \ref.
Sometimes you may need to run TeX more then once on your .tex file in order to produce and typeset indexes, resolve undefined cross-references and/or citations. The shell script texi2dvi from the Texinfo documentation system (see http://www.gnu.org/software/texinfo) can automate this process: it runs BibTeX, MakeIndex and TeX as many times as needed to complete the compilation process. You will need to set the LATEX environment variable to `tex'. For example, in a Bourne-compatible shell, the following command will do all the work:
prompt$ LATEX=tex texi2dvi file.tex
(Despite the name, texi2dvi can also produce .pdf files; just set `LATEX=pdftex'.) See the output from the command texi2dvi --help for invoking information and a full list of options.
This chapter describes definitions that are meant to be used directly in a document. When appropriate, ways to change the default formatting are described in subsections.
Plain TeX provides the \tracingall command, to turn on the
maximum amount of tracing possible in TeX. The (usually voluminous)
output from \tracingall goes both on the terminal and into the
transcript file.
It is sometimes easier to have the output go only to the transcript
file, so you can peruse it at your leisure and not obscure other output
to the terminal. So, Eplain provides the command \loggingall.
(For some reason, this command is available in Metafont, but not in
TeX.)
It is also sometimes useful to see the complete contents of boxes.
\tracingboxes does this. (It doesn't affect whether or not the
contents
are shown on the terminal.)
You can turn off all tracing with \tracingoff.
You can also turn logging on and off globally, so you don't have to
worry about whether or not you're inside a group at the time of command.
These variants are named \gloggingall
and \gtracingall.
Finally, if you write your own help messages (see \newhelp
in The TeXbook), you want a convenient way to break lines in
them. This is what TeX's \newlinechar parameter is for;
however, plain TeX doesn't set \newlinechar. Therefore,
Eplain defines it to be the character ^^J.
For example, one of Eplain's own error messages is defined as follows:
\newhelp\envhelp{Perhaps you forgot to end the previous^^J%
environment? I'm finishing off the current group,^^J%
hoping that will fix it.}%
The default dimensions of rules are defined in chapter 21 of the
The TeXbook. To sum up what is given there, the “thickness”
of rules is
0.4pt by default. Eplain defines three parameters that let you change
this dimension: \hruledefaultheight, \hruledefaultdepth,
and \vruledefaultwidth. By default, they are defined as
The TeXbook describes.
But it would be wrong to redefine \hrule and \vrule. For
one thing, some macros in plain TeX depend on the default dimensions
being used; for another, rules are used quite heavily, and the
performance impact of making it a macro can be noticeable. Therefore,
to take advantage of the default rule parameters, you must use
\ehrule
and \evrule.
Bibliographies are part of almost every technical document. To handle them easily, you need two things: a program to do the tedious formatting, and a way to cite references by labels, rather than by numbers. The BibTeX program, written by Oren Patashnik, takes care of the first item; the citation commands in LaTeX, written to be used with BibTeX, take care of the second. Therefore, Eplain adopts the use of BibTeX, and virtually the same interface as LaTeX.
The general idea is that you put citation commands in the text of your document, and commands saying where the bibliography data is. When you run TeX, these commands produce output on the file with the same root name as your document (by default) and the extension .aux. BibTeX reads this file. You should put the bibliography data in a file or files with the extension .bib. BibTeX writes out a file with the same root name as your document and extension .bbl. Eplain reads this file the next time you run your document through TeX. (It takes multiple passes to get everything straight, because usually after seeing your bibliography typeset, you want to make changes in the .bib file, which means you have to run BibTeX again, which means you have to run TeX again...) An annotated example of the whole process is given below.
If your document has more than one bibliography—for example, if it
is a collection of papers—you can tell Eplain to use a different root
name for the .bbl file by defining the control sequence
\bblfilebasename.
The default definition is simply \jobname.
See the document BibTeXing (whose text is in the file btxdoc.tex, which should be in the Eplain distribution you got) for information on how to write your .bib files. Both the BibTeX and the Eplain distributions contain several examples, also.
The \cite
command produces a citation in the text of your document. The exact
printed form the citation will take is under your control
(see Formatting citations). \cite takes one
required argument, a comma-separated list of cross-reference labels
(see Cross-references, for exactly what characters are allowed in
such labels). Warning: spaces in this list are taken as part of the
following label name, which is probably not what you expect.
The \cite command also produces a command
in the .aux file that tells BibTeX to retrieve the given reference(s)
from the .bib file. \cite also takes one optional argument,
which you specify within square brackets, as in LaTeX. This text is
simply typeset after the citations. (See the example below.)
Eplain can create hypertext links for citations pointing to the relevant bibliography entries (see Citation hyperlinks).
Another command, \nocite,
puts the given reference(s) into the bibliography, but produces nothing
in the text.
The \bibliography
command is next. It serves two purposes: producing the typeset
bibliography, and telling BibTeX the root names of the .bib
files. Therefore, the argument to \bibliography is a comma
separated list of the .bib files (without the `.bib'). Again,
spaces in this list are significant.
You tell BibTeX the particular style in which you want your
bibliography typeset with one more command:
\bibliographystyle.
The argument to this is a single filename style, which tells
BibTeX to look for a file style.bst.
See the document Designing BibTeX styles (whose text is in the
btxhak.tex) for information on how to write your own styles.
Eplain automatically reads the citations from the .aux file when your job starts.
If you don't want to see the messages about undefined citations, you
can say \xrefwarningfalse before making any citations.
Eplain automatically does this if the .aux file does not exist.
You can restore the default by saying \xrefwarningtrue.
Here is a TeX input file that illustrates the various commands.
\input eplain % Reads the .aux file.
Two citations to Knuthian works:
\cite[note]{surreal,concrete-math}.
\beginsection{References.}\par % Title for the bibliography.
\bibliography{knuth} % Use knuth.bib for the labels.
\bibliographystyle{plain} % Number the references.
\end % End of the document.
If we suppose that this file was named citex.tex and that the
bibliography data is in knuth.bib (as the \bibliography
command says), the following commands do what's required. (`$ '
represents the shell prompt.)
$ tex citex (produces undefined citation messages)
$ bibtex citex (read knuth.bib and citex.aux, write citex.bbl)
$ tex citex (read citex.bbl, still have undefined citations)
$ tex citex (one more time, to resolve the references)
The texi2dvi program can help you automate this process (see Invoking Eplain).
The output looks something like (because we used the plain bibliography style):
Two citations to Knuthian works: [2,1 note].References
[1] Ronald L. Graham, Donald E. Knuth, and Oren Patashnik. Concrete Mathematics. Addison-Wesley, Reading, Massachusetts, 1989.
[2] Donald E. Knuth. Surreal Numbers. Addison-Wesley, Reading, Massachusetts, 1974.
See the BibTeX documentation for information on how to write the
bibliography databases, and the bibliography styles that are available.
(If you want your references printed with names, as in [Knu74], instead
of numbered, the bibliography style is alpha.)
You may wish to change Eplain's formatting of citations; i.e., the
result of your \cite commands. By default, the citation labels
are printed one after another, separated by commas and enclosed in
brackets, using the main text font. Some formats require other styles,
such as superscripted labels. You can accomodate such formats by
redefining the following macros.
\printcitestart\printcitefinish\cite command. By default, they produce a
`[' and `]', respectively.
\printbetweencitations\cite command has multiple citations, as in
\cite{acp,texbook}, Eplain expands this macro in between each
pair of citations. By default, it produces a comma followed by a space.
\printcitenote\cite command. If the \cite command had no note, this
macro isn't used. Otherwise, it should print the note. By default, the
note is preceded with a comma and a space.
Here is an example, showing you could produce citations as superscripted labels, with the optional notes in parentheses.
\def\printcitestart{\unskip $^\bgroup}
\def\printbetweencitations{,}
\def\printcitefinish{\egroup$}
\def\printcitenote#1{\hbox{\sevenrm\space (#1)}}
You may wish to change Eplain's formatting of the bibliography, especially with respect to the fonts that are used. Therefore, Eplain provides the following control sequences:
\biblabelwidth\dimen register, and its value
is the width of the widest label in the bibliography. Although it is
unlikely you will ever want to redefine it, you might want
to use it if you redefine \biblabelprint, below.
\biblabelprint\biblabelwidth, and is followed by
an enspace. When you want to change the spacing around the labels, this
is the right macro to redefine.
\biblabelcontents\bblrm (below), and enclosed in
brackets. When you want to change the appearance of the label, but not
the spacing around it, this is the right macro to redefine.
\bblrm\bblem\bblsc\bblnewblock\biblabelextraspace\biblabelwidth plus this. The default is .5em, where the
em width is taken from the \bblrm font. If you want to change
this, you should do it inside \bblhook.
\bblhook\bblhook:
\parskip, which produces extra space between the items; and
\biblabelextraspace, which is described above.
(By the way, \hookappend won't work with \bblhook, despite
the names. Just use \def.)
If you are really desperate, you can also hand-edit the .bbl file that BibTeX produces to do anything you wish.
By default, TeX centers displayed material. (Displayed material is
just whatever you put between $$'s—it's not necessarily
mathematics.) Many layouts would be better served if the displayed
material was left-justified. Therefore, Eplain provides the command
\leftdisplays,
which indents displayed material by \parindent plus
\leftskip, plus \leftdisplayindent.
You can go back to centering displays with \centereddisplays.
(It is usually poor typography to have both centered and left-justified
displays in a single publication, though.)
\leftdisplays also changes the plain TeX commands that deal
with alignments inside math displays,
\displaylines,
\eqalignno,
and \leqalignno,
to produce left-justified text. You can still override this formatting
by inserting \hfill glue, as explained in The TeXbook.
If you want some other kind of formatting, you can write a definition
of your own, analogous to \leftdisplays. You need only make sure
that \leftdisplaysetup
is called at the beginning of every display (presumably by invoking it
in TeX's \everydisplay parameter), and to define
\generaldisplay.
\leftdisplays expands the old value of \everydisplay
before calling \leftdisplaysetup, so that any changes you have made
to it won't be lost. That old token list as available as the value of
the token register \previouseverydisplay.
TeX provides the day, month, and year as numeric quantities (unless your TeX implementation is woefully deficient). Eplain provides some control sequences to make them a little more friendly to humans.
\monthname
produces the name of the current month, abbreviated to three letters.
\fullmonthname
produces the name of the current month, unabbreviated (in English).
\timestring
produces the current time, as in `1:14 p.m.'
\timestamp
produces the current date and time, as in `23 Apr 64 1:14 p.m.'.
(Except the spacing is slightly different.)
\today
produces the current date, as in `23 April 1964'.
Many documents require lists of items, either numbered or simply
enumerated. Plain TeX defines one macro to help with creating lists,
\item, but that is insufficient in many cases. Therefore,
Eplain provides two pairs of commands:
\numberedlist ... \endnumberedlist\orderedlist ... \endorderedlist\numberedlist
labels the items with lowercase letters, starting with `a'. Another
nested \numberedlist labels the items with roman numerals. Yet
more deeply nested numbered lists label items with `*'.
\unorderedlist ... \endunorderedlist\unorderedlist labels items with
em-dashes. Doubly (and deeper) nested unordered lists label items with
`*'s.
The two kinds of lists can be nested within each other, as well.
In both kinds of lists, you begin an item with \li.
An item may continue for several paragraphs. Each item starts a
paragraph.
You can give \li an optional argument, a cross-reference label.
It's defined to be the “marker” for the current item. This is useful
if the list items are numbered. You can produce the value of the label
with \xrefn. See Cross-references.
Eplain can create hypertext links for the markers produced by
\xrefn pointing to the relevant list item (see List hyperlinks).
You can also say \listcompact
right after \numberedlist or \unorderedlist. The items in
the list will then not have any extra space between them
(see Formatting lists). You might want to do this if the items in
this particular list are short.
Here is an example:
\numberedlist\listcompact
\li The first item.
\li The second item.
The second paragraph of the second item.
\endnumberedlist
Several registers define the spacing associated with lists. It is likely that their default values won't suit your particular layout.
\abovelistskipamount, \belowlistskipamount\interitemskipamount\listcompact resets this to zero, as mentioned above.
\listleftindent, \listrightindent\listrightindent is the amount of space by which the list is
indented on the right; i.e., it is added to \rightskip.
\listleftindent is the amount of space, relative to
\parindent, by which the list is indented on the left. Why treat
the two parameters differently? Because (a) it is more useful to
make the list indentation depend on the paragraph indentation;
(b) footnotes aren't formatted right if \parindent is reset
to zero.
The three vertical glues are inserted by macros, and preceded by
penalties: \abovelistskip
does \vpenalty\abovelistpenalty
and then \vskip\abovelistskip. \belowlistskip
and \interitemskip
are analogous.
In addition, the macro \listmarkerspace
is called to separate the item label from the item text. This is set to
\enspace by default.
If you want to change the labels on the items, you can redefine these
macros:
\numberedmarker
or \unorderedmarker.
The following registers might be useful if you do:
\numberedlistdepth, \unorderedlistdepth\itemnumber, \itemletter\itemnumber starts at one, and \itemletter starts at 97,
i.e., lowercase `a'.
You can also redefine the control sequences that are used internally,
if you want to do something radically different: \beginlist
is invoked to begin both kinds of lists; \printitem
is invoked to print the label (and space following the label) for each
item; and \endlist
is invoked to end both kinds of
lists.
It is sometimes useful to include a file verbatim in your document;
for example, part of a computer program. The \listing
command is given one argument, a filename, and produces the contents of
that file in your document. \listing expands \listingfont
to set the current font. The default value of \listingfont
is \tt.
You can take arbitrary actions before reading the file by defining the macro
\setuplistinghook.
This is expanded just before the file is input.
If you want to have line numbers on the output, you can say
\let\setuplistinghook = \linenumberedlisting.
The line numbers are stored in the count register \lineno while
the file is being read. You can redefine the macro
\printlistinglineno
to change how they are printed.
Normally, the \listing command will add a final empty line at
the end of the output, even if the file does not end in a newline. To
suppress this final line, you can say
\let\setuplistinghook = \nolastlinelisting.
This also works with line numbers (say
\def\setuplistinghook{\linenumberedlisting \nolastlinelisting}),
but only if \printlistinglineno
consists exclusively of boxes at the top level (i.e., any
\kerns or glue should be wrapped up in a box).
You can use the form feed control character (ASCII code 12, typed as CTRL-L) in the file to force a page break in the output.
You can produce in-line verbatim text in your document with \verbatim.
End the text with |endverbatim. If you need a `|' in the text,
double it. If the first character of the verbatim text is a space, use
| . (| will work elsewhere in the argument, too, but
isn't necessary.)
For example:
\verbatim| ||\#%&!|endverbatim
produces |\#%&!.
Line breaks and spaces in the verbatim text are preserved.
You can change the verbatim escape character from the default `|'
with \verbatimescapechar char; for example, this changes
it to `@'.
\verbatimescapechar \@
The backslash is not necessary in some cases, but is in
others, depending on the catcode of the character. The argument to
\verbatimescapechar is used as \catcode `char, so
the exact rules follow that for \catcode.
Because \verbatim must change the category code of special
characters, calling inside a macro definition of your own does not work
properly. For example:
\def\mymacro{\verbatim &#%|endverbatim}% Doesn't work!
To accomplish this, you must change the category codes yourself before
making the macro definition. Perhaps \uncatcodespecials will
help you (see Category codes).
Producing a table of contents that is both useful and aesthetic is one of the most difficult design problems in any work. Naturally, Eplain does not pretend to solve the design problem. Collecting the raw data for a table of contents, however, is much the same across documents. Eplain uses an auxiliary file with extension .toc (and the same root name as your document) to save the information.
To write an entry for the table of contents, you say
\writetocentry{part}{text},
where part is the type of part this entry is, e.g.,
`chapter', and text is the text of the title.
\writetocentry puts an entry into the .toc file that looks
like \tocpartentry{text}{page number}
(unless part is an integer, see below). The text is
written unexpanded.
A related command, \writenumberedtocentry, takes one additional
argument, the first token of which is expanded at the point of the
\writenumberedtocentry, but the rest of the argument is not
expanded. The usual application is when the parts of the document are
numbered. On the other hand, the one-level expansion allows you to use
the argument for other things as well (author's names in a proceedings,
say), and not have accents or other control sequences expanded. The
downside is that if you want full expansion of the third
argument, you don't get it—you must expand it yourself, before you
call \writenumberedtocentry.
For example:
\writenumberedtocentry{chapter}{A $\sin$ wave}{\the\chapno}
\writetocentry{section}{A section title}
Supposing \the\chapno expanded to `3' and that the
\write's occurred on pages eight and nine, respectively, the
above writes the following to the .toc file:
\tocchapterentry{A $\sin$ wave}{3}{8}
\tocsectionentry{A section title}{9}
A variation on \writenumberedtocentry is
\writenumberedtocline, differing only in the order of the
parameters it takes and writes for the \tocpartentry
control sequences. To continue the previous example:
\writenumberedtocline{chapter}{\the\chapno}{A $\sin$ wave}
writes the following to the .toc file:
\tocchapterentry{3}{A $\sin$ wave}{8}
Such ordering of the parameters allows the
\tocpartentry macros to typeset the text of the entry
without actually reading it as an argument. This is required for
entries which need to change character catcodes, e.g., to produce
verbatim text (see Verbatim listing).
Each of \writetocentry, \writenumberedtocentry and
\writenumberedtocline processes a numeric part argument
specially. If you pass part expanding to an integer, these
macros write into the .toc file an entry that starts with
\tocentry{part}. Thus, you can define a single
\tocentry macro which formats all entries for a table of
contents. To continue the previous examples:
\writenumberedtocentry{1}{A $\sin$ wave}{\the\chapno}
\writenumberedtocline{1}{\the\chapno}{A $\sin$ wave}
\writetocentry{2}{A section title}
writes the following to the .toc file:
\tocentry{1}{A $\sin$ wave}{3}{8}
\tocentry{1}{3}{A $\sin$ wave}{8}
\tocentry{2}{A section title}{9}
You read the .toc file with the command \readtocfile.
Naturally, whatever \toc... entry commands that were written
to the file must be defined when \readtocfile is invoked. Eplain
has minimal definitions for \tocchapterentry,
\tocsectionentry, and \tocsubsectionentry, just to prevent
undefined control sequence errors in common cases. They aren't suitable
for anything but preliminary proofs.
Each of \writetocentry, \writenumberedtocentry and
\writenumberedtocline opens
the .toc file for writing, thereby deleting the information from the
previous run. You should therefore arrange that \readtocfile be
called before the first call to a \writetoc... macro.
\readtocfile does not itself delete the information
from the .toc file, so that you can call it several times,
e.g., to create both a short
and normal table of contents. (To produce this in particular, define
\tocsectionentry to produce nothing while you are reading
.toc file for a short table of contents (see Macro arguments).)
On the other hand, if you don't want to rewrite the .toc file at
all, perhaps because you are only running TeX on part of your
manuscript, you can set \rewritetocfilefalse.
By default, the .toc file has the root \jobname. If your
document has more than one contents—for example, if it is a collection
of papers, some of which have their own contents—you can tell Eplain
to use a different root name by defining the control sequence
\tocfilebasename.
Note that \writetocentry, \writenumberedtocentry and
\writenumberedtocline will
open the contents file for writing only at the first call, using the
value of \tocfilebasename at that time. Changing the value of
\tocfilebasename afterwards will not affect which file gets
written, although it will affect which file gets read by
\readcontentsfile. In case you need to write several contents
files from a single TeX job, use \definecontentsfile
(see Alternative contents files).
In addition to the usual table of contents, you may want to have a list
of figures, list of tables, or other such contents-like list. You can do
this with \definecontentsfile{abbrev}. All of the
above commands are actually a special case that Eplain predefines with
\definecontentsfile{toc}
The abbrev is used both for the file extension and in the control sequence names.
It is often useful to refer the reader to other parts of your document; but putting literal page, section, equation, or whatever numbers in the text is certainly a bad thing.
Eplain therefore provides commands for symbolic cross-references. It uses an auxiliary file with extension .aux (and the same root name as your document) to keep track of the information. Therefore, it takes two passes to get the cross-references right—one to write them out, and one to read them in. Eplain automatically reads the .aux file at the first reference; after reading it, Eplain reopens it for writing.
You can control whether or not Eplain warns you about undefined labels. See Citations.
Labels in Eplain's cross-reference commands can use characters of category code eleven (letter), twelve (other), ten (space), three (math shift), four (alignment tab), seven (superscript), or eight (subscript). For example, `(a1 $&^_' is a valid label (assuming the category codes of plain TeX), but `%#\{' has no valid characters.
You can also do symbolic cross-references for bibliographic citations and list items. See Citations, and Lists.
Eplain can create hypertext links for the cross-references (see Cross-reference hyperlinks).
Eplain provides the command \definexref for general
cross-references. It takes three arguments: the name of the label (see
section above for valid label names), the value of the label (which can
be anything), and the “class” of the reference—whether it's a
section, or theorem, or what. For example:
\definexref{sec-intro}{3.1}{section}
Of course, the label value is usually generated by another macro using TeX count registers or some such.
\definexref doesn't actually define label; instead, it
writes out the definition to the .aux file, where Eplain will read
it on the next TeX run.
The class argument is used by the \ref and \refs
commands. See the next section.
To retrieve the value of the label defined via \definexref (see
the previous section), Eplain provides the following macros:
\refn{label}\xrefn{label}\refn and \xrefn (they are synonyms) produce the bare
definition of label. If label isn't defined, issue a
warning, and produce label itself instead, in typewriter. (The
warning isn't given if \xrefwarningfalse.)
\ref{label}\definexref in the previous section), expand the control sequence
\c word (if it's defined) followed by a tie. Then call
\refn on label. (Example below.)
\refs{label}\ref, but append the letter `s' to the
\...word.
The purpose of the \...word macro is to produce the word
`Section' or `Figure' or whatever that usually precedes the actual
reference number.
Here is an example:
\def\sectionword{Section}
\definexref{sec-intro}{3.1}{section}
\definexref{sec-next}{3.2}{section}
See \refs{sec-intro} and \refn{sec-next} ...
This produces `See Sections 3.1 and 3.2 ...'
Eplain provides two commands for handling references to page numbers, one for definition and one for use.
\xrdef{label}\xref{label}\xrefpageword.
Its default definition is p.\thinspace.
Eplain can create hypertext links for the page references (see Page reference hyperlinks).
Instead of referring to pages, it's most useful if equation labels
refer to equation numbers. Therefore, Eplain reserves a \count
register, \eqnumber,
for the current equation number, and increments it at each
numbered equation.
Here are the commands to define equation labels and then refer to them:
\eqdef{label}\eqnumber,
and, if the current context is not inner, then produces a \eqno
command. (The condition makes it possible to use \eqdef in an
\eqalignno construction, for example.) The text of the equation
number is produced using \eqprint. See Formatting equation references.
If label is empty, you still get an equation number (although naturally you can't reliably refer to it). This is useful if you want to put numbers on all equations in your document, and you don't want to think up unique labels.
To refer to the last equation with the empty label, you just use the
empty label in one of the equation reference macros (see below). This
can be handy when you want to refer to an equation shortly after its
definition, say, in the sentence following the displayed equation, and
do not intend to refer to the equation later. But use this trick with
extreme caution: if later you change the text and insert another
empty definition between the original definition and the reference,
the reference will start to refer to the new empty-labeled equation.
\eqdefn{label}\eqdef, except it always omits the \eqno
command. It can therefore be used in places where \eqdef can't;
for example, in a non-displayed equation. The text of the equation
number is not produced, so you can also use it in the (admittedly
unusual) circumstance when you want to define an equation label but not
print that label.
\eqref{label}\eqprint.
\eqrefn{label}\eqref, except it doesn't call \eqprint.
Equation labels can contain the same characters that are valid in general cross-references.
Eplain can create hypertext links for the equation references (see Equation reference hyperlinks).
Both defining an equation label and referring to it should usually
produce output. This output is produced with the \eqprint macro,
which takes one argument, the equation number being defined or referred
to. By default, this just produces `(number)', where
number is the equation number. To produce the equation number in
a different font, or with different surrounding symbols, or whatever,
you can redefine \eqprint.
For example, the following definition would print all equation numbers
in italics. (The extra braces define a group, to keep the font change
from affecting surrounding text.)
\def\eqprint#1{{\it (#1)}}
In addition to changing the formatting of equation numbers, you might want
to add more structure to the equation number; for example, you might
want to include the chapter number, to get equation numbers like
`(1.2)'. To achieve this, you redefine \eqconstruct.
For example:
\def\eqconstruct#1{\the\chapternumber.#1}
(If you are keeping the chapter number in a count register named
\chapternumber, naturally.)
The reason for having both \eqconstruct and \eqprint may
not be immediately apparent. The difference is that \eqconstruct
affects the text that cross-reference label is defined to be, while
\eqprint affects only what is typeset on the page. The example
just below might help.
Usually, you want equation labels to refer to equation numbers. But sometimes you might want a more complicated text. For example, you might have an equation `(1)', and then have a variation several pages later which you want to refer to as `(1*)'.
Therefore, Eplain allows you to give an optional argument (i.e.,
arbitrary text in square brackets) before the cross-reference label to
\eqdef. Then, when you refer to the equation, that text is
produced. Here's how to get the example just mentioned:
$$...\eqdef{a-eq}$$
...
$$...\eqdef[\eqrefn{a-eq}*]{a-eq-var}$$
In \eqref{a-eq-var}, we expand on \eqref{a-eq}, ...
We use \eqrefn in the cross-reference text, not
\eqref, so that \eqprint is called only once.
As another example, consider the following requirement: we want to
include chapter number in all equation references, and additionally we
want to include the part number when referencing an equation from any
part other than the one where the equation appears. For example,
references to the third equation in chapter 2 of part 1
should be typeset as `(2.3)' throughout part 1, but as `(I.2.3)'
in any other part. Let's assume we have the current chapter and part
numbers in count registers \chapnum and \partnum,
respectively.
The idea is to have \eqconstruct store the part number of the
equation (that is, the part number at the time of definition),
so that later \eqprint can compare the stored number with the
current part number (that is, the part number at the time of
reference). The complicating factor is that internally, the result
of \eqconstruct is both expanded and written out to the
.aux file, and used to typeset the equation number, so
the commands that store the part number should behave correctly in
both situations. This is difficult to achieve with expandable
commands; therefore, to avoid expansion problems, we are going to use
only TeX primitives, which are non-expandable:
\newcount\eqpartnum
\def\eqconstruct#1{%
\global\eqpartnum=\the\partnum\relax
\number\chapnum.#1%
}
\def\eqprint#1{%
\setbox0=\hbox{#1}%
(\ifnum\partnum=\eqpartnum \else
\uppercase\expandafter{\romannumeral\eqpartnum}.%
\fi
\box0)%
}%
In \eqconstruct, besides constructing the base equation number
(e.g., `1.2'), we also store the part number of the equation in the
count register \eqpartnum (\the\partnum is expanded when
the equation number is written to the .aux file, so the
equation label definition in the .aux file will contain the
actual part number). In \eqprint, we need to know the
equation's part number before we typeset the base equation number,
therefore we first put the argument in a box, thus causing
\eqpartnum to be set.
Eplain also provides for one level of substructure for equations. That is, you might want to define a related group of equations with numbers like `2.1' and `2.2', and then be able to refer to the group as a whole: “... in the system of equations (2)...”.
The commands to do this are \eqsubdef and
\eqsubdefn.
They take one label argument like their counterparts above,
and generally behave in the same way. The difference is in how they
construct the equation number: instead of using just \eqnumber,
they also use another counter, \subeqnumber.
This counter is advanced by one at every \eqsubdef or
\eqsubdefn, and reset to zero at every \eqdef or
\eqdefn.
You use \eqref to refer to subequations as well as main
equations.
To put the two together to construct the text that the label will
produce, they use a macro \eqsubreftext.
This macros takes two arguments, the “main” equation number (which,
because the equation label can be defined as arbitrary text, as
described in the previous section, might be anything at all) and the
“sub” equation number (which is always just a number). Eplain's
default definition just puts a period between them:
\def\eqsubreftext#1#2{#1.#2}%
You can redefine \eqsubreftext to print however you
like. For example, this definition makes the labels print as `2a',
`2b', and so on.
\newcount\subref
\def\eqsubreftext#1#2{%
\subref = #2 % The space stops a <number>.
\advance\subref by 96 % `a' is character code 97.
#1\char\subref
}
Sadly, we must define a new count register, \subref,
instead of using the scratch count register \count255, because
`#1' might include other macro calls which use \count255.
Eplain provides support for generating raw material for an index, and for typesetting a sorted index. A separate program must do the actual collection and sorting of terms, because TeX itself has no support for sorting.
Eplain can create hypertext links pointing from the index to the index terms (see Index hyperlinks).
Eplain's indexing commands were designed to work with the program MakeIndex, available from CTAN hosts in tex-archive/indexing/makeindex; MakeIndex is also commonly included in prepackaged TeX distributions. It is beyond the scope of this manual to explain how to run MakeIndex, and all of its many options. See http://www.ctan.org/tex-archive/indexing/makeindex.
The basic strategy for indexing works like this:
\idx; see the section `Indexing terms' below) write the raw index
material to foo.idx.
The texi2dvi program can help you automate this process (see Invoking Eplain).
If your document needs more than one index, each must have its own
file. Therefore, Eplain provides the command \defineindex, which
takes an argument that is a single letter, which replaces `i' in
the filenames and in the indexing command names described below. For
example,
\defineindex{m}
defines the command \mdx to write to the file
foo.mdx. Eplain simply does \defineindex{i} to define
the default commands.
For each index defined with \defineindex{n}, Eplain
provides a switch \ifndx which controls whether indexing
commands write index entries to the corresponding index file. However,
even when index term writing is disabled, indexing commands still do
all other processing of their arguments, including typesetting of
proof index terms (see Proofing index terms.
For example, if you write \idxfalse near the beginning of a
document foo.tex (before the first indexing command), Eplain
will not open the default index file (foo.idx) and the
corresponding indexing commands (\idx, \sidx, etc.)
will not write index entries there. This may be useful for draft
compilations of a manuscript, e.g., to avoid the overhead of index
file input/output.
Indexing commands in Eplain come in pairs: one command that only writes the index entry to the `.idx' file (see above section), and one that also typesets the term being indexed. The former always starts with `s' (for “silent”). In either case, the name always includes `Idx', where I is the index letter, also described above. Eplain defines the index `i' itself, so that's what we'll use in the names below.
The silent form of the commands take a subterm as a trailing optional
argument. For example, \sidx{truth}[definition of] on page 75
makes an index entry that will eventually be typeset (by default) as
truth
definition of, 75
Also, the silent commands ignore trailing spaces. The non-silent ones do not.
\sidx{term}[subterm] makes an index entry for
term, optionally with subterm subterm.
\idx{term} also produces term as output. Example:
\sidx{truth}[beauty of]
The beauty of truth is \idx{death}.
Subterms at the second and further levels can also be
specified in subterm, using the
\idxsubentryseparator character to separate
them. This character is by default `!'.
\sidxname{First M.}{von Last}[subterm]
makes an index
entry for `von Last, First M.'. You can change the
`, ' by redefining \idxnameseparator.
\idxname{First M.}{von Last} also produces First M. von
Last as output. (These commands are useful special cases of \idx
and \sidx.) Example:
\sidxname{Richard}{Stark}
\idxname{Donald}{Westlake} has written many kinds of novels, under
almost as many names.
\sidxmarked\cs{term}[subterm] makes an index
entry for term[subterm], but term will be put
in the index as \cs{term}, but still sorted as just
term. \idxmarked\cs{term} also typesets
\cs{term}. This provides for the usual ways of changing
the typesetting of index entries. Example:
\def\article#1{``#1''}
\sidxmarked\article{Miss Elsa and Aunt Sophie}
Peter Drucker's \idxmarked\article{The Polanyis} is a remarkable
essay about a remarkable family.
\sidxsubmarked{term}\cs{subterm} makes an index
entry for term, subterm as usual, but also puts subterm in
the index as \cs{term}.
\idxsubmarked{term}\cs{subterm} also typesets
term \cs{subterm}, in the unlikely event that your
syntax is convoluted enough to make this useful. Example:
\def\title#1{{\sl #1}}
\sidxsubmarked{Anderson, Laurie}\title{Strange Angels}
The \idxsubmarked{Anderson}\title{Carmen} is a strange twist.
The commands above rely on MakeIndex's feature for separating sorting of
an index entry's from its typesetting. You can use this directly by
specifying an index entry as sort@typeset. For
example:
\sidx{Ap-weight@$A_\pi$-weight}
will sort as Ap-weight, but print with the proper math.
The @ here is MakeIndex's default character for this purpose.
See http://www.ctan.org/tex-archive/indexing/makeindex. To make an index
entry with an @ in it, you have to escape it with a backslash;
Eplain provides no macros for doing this.
After any index command, Eplain runs
\hookaction{afterindexterm}. Because the index commands always
add a whatsit item to the current list, you may wish to preserve a
penalty or space past the new item. For example, given a conditional
\if@aftersctnhead set true when you're at a section heading, you
could do:
\hookaction{afterindexterm}{\if@aftersctnhead \nobreak \fi}
All the index commands described in the previous section take an initial
optional argument before the index term, which modify the index entry's
meaning in various ways. You can specify only one of the following in
any given command, except that begin and end can be
specified together with pagemarkup=cs (separate them with
a comma without a following space, like this:
[begin,pagemarkup=defn]).
These work via MakeIndex's “encapsulation” feature. See Customizing indexing, if you're not using the default characters for the MakeIndex operators. The other optional argument (specifying a subterm) is independent of these.
Here are the possibilities:
beginend \sidx[begin]{future}[Cohen, Leonard]
...
\sidx[end]{future}[Cohen, Leonard]
will typeset as something like
future,
Cohen, Leonard, 65–94
see \sidx[see]{analysis}[archetypal]{archetypal criticism}
becomes
analysis,
archetypal, see archetypal criticism
seealsosee (the previous item), but also allows for normal
index entries of the referencing term. The normal index entries have
to be created separately—seealso does not contribute a
page number to the index entry. For example, if you have indexed a
term on pages 75, 97 and 114, and then add a seealso
entry for the term:
\sidx[seealso]{archetypal criticism}[elements of]{dichotomies}
the index will contain
archetypal criticism,
elements of, 75, 97, 114, see also dichotomies
(Aside for the academically curious: The archetypally critical book I
took these dichotomous examples from is Laurence Berman's The
Musical Image, which I happened to co-design and typeset.)
pagemarkup=cs\cs before the page number in the typeset index,
thus allowing you to underline definitive entries, italicize examples,
and the like. You do not precede the control sequence cs
with a backslash. (That just leads to expansive difficulties.) Naturally
it is up to you to define the control sequences you want to use. Example:
\def\defn#1{{\sl #1}}
\sidx[pagemarkeup=defn]{indexing}
becomes something like
indexing, \defn{75}
Indexing terms with special characters can become quite cumbersome because you have to keep both TeX and MakeIndex happy at the same time. For example, while `!' has no special meaning for TeX, it is a subentry separator for MakeIndex, therefore you'd have to escape occurrences of literal `!' in index terms. Things get even more interesting with characters which are special in both TeX and MakeIndex.
This in turn has some implications for the non-silent forms of the indexing commands (see Indexing terms), since TeX and MakeIndex use different conventions for escaping characters. For example, this will not typeset the exclamation point correctly within the text, while it will come out right inside the index, after MakeIndex strips the quoting character (`"'):
\idx{"!}
This would have to be rewritten using the silent command:
!\sidx{"!}
In general, it is a good idea to eschew the non-silent commands whenever index term contains anything unusual.
To understand this keep in mind that indexing commands read the terms verbatim so that the terms can embed almost any character, and that's what gets written into the .idx file. The non-silent forms then typeset the term by rescanning the verbatim copy, therefore for the non-silent commands the term, besides being a valid MakeIndex input, must also represent a valid TeX input. The silent commands don't have this restriction—their terms only need to become valid TeX input after MakeIndex processes the .idx file and writes the .ind file. This is what makes the non-silent commands less powerful and more troublesome when dealing with special characters.
Here's an example showing that terms for the silent commands can contain almost any character:
\sidx[see]{comments}[with %@with \verbatim %"|endverbatim]
{commenting with \verbatim %"|endverbatim}
We didn't have to escape `%' in the sort string for MakeIndex, while we had to put it inside the verbatim environment (see Verbatim listing) in the part which MakeIndex will pass back to TeX. Also, we had to escape the `|' character because it is special for MakeIndex. If you have trouble understanding the reasons for the different types of escaping used, it is best to examine the .idx and .ind files resulting from processing the above input.
As was mentioned, index terms can embed “almost any character”, so now we'll describe the exceptions.
The following characters are reset to their usual meanings because they are not useful verbatim: multiple consequent spaces are converted into a single space; ASCII tab characters are treated as spaces; ASCII return is treated as end-of-line (this means, among other things, that long terms can be broken across several lines).
You have to be careful with the begin- and end-group characters (`{' and `}' by default). If they are matched, you don't have to do anything special. For example:
\sidx {braces {, }@braces
\verbatim {"|endverbatim, \verbatim }"|endverbatim}
However, if they are not matched you have two problems on
hand. The first one is TeX—you have to instruct TeX to use
something else as begin- and/or end-group characters. Eplain provides
an easy way to do this: just define
\idxargopen and/or
\idxargclose to the begin- and end-group characters you are
going to use with indexing macros, and use braces inside index terms
without any restrictions. Here's an example:
\def\idxargopen{`\<}
\def\idxargclose{`\>}
\sidx <left brace "{@left brace \verbatim "{"|endverbatim>
In this example we've also dealt with the second problem—braces are MakeIndex's grouping characters as well (by default), so we have escaped unmatched braces with `"'.
And the final note: if you need a subentry containing brackets
(`[' and `]'), avoid the optional argument of \sidx
and friends, and use instead MakeIndex's subentry separator to create
the subentry with the brackets in it:
\sidx{entry!subentry with a bracket [}
As you are reading through a manuscript, it is helpful to see what terms have been indexed, so you can add others, catch miscellaneous errors, etc. (Speaking from bitter experience, I can say it is extremely error-prone to leave all indexing to the end of the writing, since it involves adding many TeX commands to the source files.)
So Eplain puts index terms in the margin of each page, if you
set \indexproofingtrue. It is false by default. The terms
are typeset by the macro \indexproofterm, which takes a single
argument, the term to be typeset. Eplain's definition of
\indexproofterm just puts it into an \hbox, first doing
\indexprooffont, which Eplain defines to select the font
cmtt8. With this definition long terms run off the page, but
since this is just for proofreading anyway, it seems acceptable.
On the other hand, we certainly don't want the index term to run into
the text of the page, so Eplain uses the right-hand side of the page
rather than the left-hand page (assuming a language read left to right
here). So \ifodd\pageno, Eplain kerns by \outsidemargin,
otherwise by \insidemargin. If those macros are undefined,
\indexsetmargins defines them to be one inch plus \hoffset.
To get the proofing index entries on the proper page, Eplain defines a new
insertion class \@indexproof. To unbox any index proofing
material, Eplain redefines \makeheadline to call
\index