style.doc

TeX code from the Z Reference Manual.
-------------------------------------

The Z Reference Manual uses a LaTeX document style called `phi', and
style options `fuzz' -- the standard Z macros -- and `zrm' -- some
special stuff.  The phi style is based on a design document from
Prentice-Hall, but I implemented only the design elements that
occurred in the ZRM, and added a couple of my own: for example, the
`manual page' layout.

There are three sections to these notes: one on phi.sty, one on
zrm.sty, and a final one on the three index-generating programs
`index', `symdex' and `gloss'.


PHI.STY -- A LaTeX style for the Prentice-Hall red-and-white series
-------    - ----- ----- --- --- ------------- ------------- ------

This style is modelled on the `book' style of LaTeX, but has quite a
lot of differences ...

No base size style option (10pt, 11pt, 12pt) is supported. The only
available base size is 12pt.

The following sectioning commands are supported, in both unstarred
and starred forms:

	\chapter
	\section
	\subsection
	\paragraph

Section numbering applies to chapters, sections and subsections.
Paragraphs are not numbered.  The right-hand running head is taken
from the last subhead on the page (not the first as in ordinary
LaTeX).

A special `thema' environment allows a decorative quotation to be
added to a chapter opening.  Here is an example from the ZRM:

    \chapter*{Preface}

    \begin{thema}
	\begin{playlet}
	\item[Jack] You're quite perfect\index{perfection}, 
		Miss Fairfax.
	\item[Gwendolen] Oh! I hope I am not that. It would 
		leave no room for developments, and I intend 
		to develop in many~directions.
		\source{Oscar Wilde, {\em The 
				Importance of Being Earnest}}
	\end{playlet}
    \end{thema}

    \noindent
    The Z notation for specifying and designing software 
    has evolved ...

The following variants of the list environment are supported (page
numbers refer to examples in the ZRM):

* enumerate
* itemize (p. 2)
* quotation -- smaller type, paragraphs indented, whole environment
  deeply indented (p. xi).
* playlet -- with names of speakers set in caps and small caps (see
  above for example from p. vii).
* oblig -- list of proof obligations, numbered (a1) (a2) (a3), etc,
  where the `a' is an argument to the environment (p. 126).
* manpage -- UNIX-style manual page (p. 89).
* description (p. ix).

There is a simple titlepage environment for making a cover sheet --
P-H generally set the real title-page themselves.

The \tableofcontents command is supported. There are environments
called theindex and theglossary; in the index, there is a special
command \indexspace for the space between items with different
initial letters. There is an \indexnote{...} command for specifying
a descriptive note to be included at the beginning of the index. Use
it before the start of the theindex environment.

Another environment, thesymdex, allows an index of symbols to be
compiled.  This index is set in smaller type and three equal
columns. If it won't fit on one page, you lose.  As well as the
commands \makeindex, \index and \makeglossary, \glossary, there is a
new pair \makesymdex and \symdex which write entries on the file
jobname.sdx.

The float placement parameters for figures and tables are unchanged
from the LaTeX defaults, and there is no provision for any special
kinds or formats of floats.


ZRM.STY -- Special little hacks for the Z Reference Manual
-------    ------- ------ ----- --- --- - --------- ------

* \source{...} works like this:

    \begin{quotation}
	Anyone can write a new set of TeX macros; they only need to
	memorize 360 pages of arcane gobbledegook and it's easy!
	\source{Mike Spivey}
    \end{quotation}

The source of the quotation is ranged right, and put on a line of
its own if necessary.

* \because{...} puts an explanation on its own line in an `argue'
  environment:

    \begin{argue}
	\dom birthday' \\
    \t1	= known' &			invariant after \\
    \t1	= \{\,i: 1 \upto hwm' \spot names'(i)\,\} & 
					from $Abs'$ \\
    \t1	= \{\,i: 1 \upto hwm \spot names'(i)\,\} 
				\cup \{names'(hwm')\} \\
	    \because{since $hwm' = hwm + 1$}
    \t1	= \{\,i: 1 \upto hwm \spot names(i)\,\} 
				\cup \{name?\} \\
	    \because{since $names' = names \oplus 
				\{hwm' \mapsto name?\}$}
    \t1	= known \cup \{name?\} &	from $Abs$ \\
    \t1	= \dom birthday \cup \{name?\} & invariant before
    \end{argue}

* There is a `display' environment for one-off math displays: it
  behaves like all the other Z displays.

* A - in math mode makes a hyphen; for a minus sign, use \minus.

* \lopt and \ropt make slanting square brackets.

* a name environment is provided for the `name' section of a
  manpage. Also a laws environment for the `laws' section. Goodness
  knows what the \- command does.

* There is a menu environment for the little indexes of symbols at
  the beginning of chapter 4.

* The syntax environment is modified to set everything in the \sf
  font.

* There is an `equations' environment: what for?

* \symlist{a b c} makes a list of math symbols.

* There's a whole lot of stuff for making fragments of schema boxes
  to use in syntax charts.

* \(...\) sets its argument as a math formula in sans-serif.

As well as these commands and environments, there are a few
adjustments to parameters: 

* \em is redefined to use slanted type instead of italics.

* Page breaking is allowed before all math displays, not just
  schemas.

* \also gives only 3pt (\jot) of extra vertical space, not
  half a line (\medskipamount) as in fuzz.sty


Indexing
--------

The ZRM has three indexes: a general index, an index of symbols, and
a glossary. Each is compiled from annotations in the text of the
book.  When LaTeX finds one of these annotations as it processes the
book, it writes a record to one of three output files, one for each index.
There are three programs that process these output files -- zrm.idx,
zrm.sdx, and zrm.glo -- into the three input files -- index.tex,
symdex.tex, and gloss.tex -- for the next run of LaTeX.  As with
LaTeX cross-references, it is necessary to run the book twice to get
everything up to date. The three programs are run between the first
LaTeX run and the second one.