DocBook style sheet 0.87

[Norman Walsh <norm@xxxxxxxxxxxxx>]

[Apologies if this goes out twice, I sent it yesterday but find
no record of receiving it.]

At http://www.berkshire.net/~norm/dsssl/db087.zip

Jon recently asked me when db08*.zip might be ready to publish
on sunsite.  I think that 0.87 is pretty close to ready.  If you
have time, please test it on some of your documents and let me
know if you find any problems.  If there aren't any
show-stoppers, I'll probably send it along to Jon sometime
around the end of the week.

README for DocBook Style Sheet version 0.87, 1997-07-13

Changed method for configuring generated text; this changed a bunch of
  definitions in the localization module and rippled through many other
  files.  (See below)
Localization is now done by pointing to the complete style sheet containing
  localization, rather than by entity reference.  This allows the localization
  layer to include parameter entities (e.g., ISO entity sets).
Use gentext for labels and label seperators in dbsect.dsl
Added localization for start and end quotation marks
Fixed bug where admonitions with titles got formatted oddly
Support APPENDIXes, GLOSSARYs, BIBLIOGRAPHYs, and INDEXes occuring after
  PARTs in the autotoc
Changed a whole slew of inlines from (empty-sosofo) to ($charseq$)
Added limited support for REVHISTORY
Consolidated table-caption-mode (in dbtable.dsl) and figure-caption-mode
  into a single block-caption mode.
Support TITLE on EXAMPLE with block-caption-mode
Added %two-side% to configure running heads appropriately (James Bostock)
Redefined several page-attributes (is attributes the right word?) as
  (declare-initial-value) instad of using a (define)d variable everywhere.
  This method was suggested by Paul Prescod on dssslist.  In the future,
  more attributes may be initialized this way; I did the easy ones this time.
Made ARTHEADER (process-children) instead of (empty-sosofo) so that
  (ARTHEADER TITLE) fires. (Ulrich Drepper)
Pass FORMAT notation on graphics to the back end (for TeX) (Ulrich Drepper)
Added support for GLOSSTERM as an inline
Tried to make INLINEGRAPHIC actually inline. Doesn't work in RTF, but looks
  like it should, IMHO ;-)
Added support for XREF.  Works for many TITLE'd components.  Supports both
  lableled (enumerated) sections and unlabeled.  Some generated
  text in dblink.dsl will have to be changed for non-English
  styles using unlabled sections, but it's not simply a text
  question, it's a grammar question, so I haven't tried to
  extract it out into the localization files.  In the non-labeled
  style, XREFs produce text like this "the section called
  [section title] in Chapter [#]".
Added support for NUMERATION on ORDEREDLIST
Incorporated patch for VARLISTENTRY LISTITEM spacing: in VARLISTENTRY 
  there was a problem with the way indentation was calculated when a 
  VARLISTENTRY included an ITEMIZEDLIST.  (Ulrich Drepper)
Procedure STEPs are no longer bold.
Modularized title code in dbtitle.dsl, use that code to generate
  REFNAMEs so they appear the same as other titles (James Bostock)
Made SYNOPSIS a $verbatim-display$, made CMDSYNOPSIS a $paragraph$ (for
  proper vertical spacing).
Added support ARG in CMDSYNOPSIS (James Bostock)
Made FUNCSYNOPSIS a $verbatim-display$.  This is an inadequate solution,
  but it looks a little better than (process-children)
Replaced %gentext-authored-by% and %gentext-copyright-by% with 
  %gentext-by%
Fixed bug: use of the word "Page" in the page number was erroneously
  dependent on the use of %page-number-restarts%

Notes on generated text:

Using a different definition for each piece of generated text
was rapidly becoming unmanagable.  Most generated text is keyed
of the type of element (the word "Chapter" for a CHAPTER, etc.)
so one function takes the name of the element and returns the
appropriate string.  Other functions were added to configure
things like the characters that occur in and between labels.  In
a few cases I've stretched the rules a bit (gentext abuse, I
suppose, to missquote a colleage ;-), but the principle is to
use a function keyed off the element name (or pseudo-element
name).  There's some potential for this to get unmanagable as
well.  Suggestions welcome.


 DSSSList info and archive:  http://www.mulberrytech.com/dsssl/dssslist