Reply-to: Mailing list for the LaTeX3 project <LATEX-L@VM.URZ.UNI-HEIDELBERG.DE>
Message-id: <9506231210.AA01430@atlantis.Mathematik.Uni-Mainz.DE>
MIME-version: 1.0
Content-type: TEXT/PLAIN; CHARSET=US-ASCII
Content-transfer-encoding: 7BIT
Date: Fri, 23 Jun 1995 14:10:14 +0000
From: Klaus Georg Barthelmann <barthel@INFORMATIK.MATHEMATIK.UNI-MAINZ.DE>
Sender: Mailing list for the LaTeX3 project <LATEX-L@VM.URZ.UNI-HEIDELBERG.DE>
To: Multiple recipients of list LATEX-L <LATEX-L@VM.URZ.UNI-HEIDELBERG.DE>
Subject: Discussing LaTeX2e's modguide
Status: R

Hi all!
This contribution addresses the topics raised in the file `modguide.tex', which
is currently distributed with LaTeX2e: May official files be changed or not?
 
Any LaTeX installation is supposed to serve two contradictory demands: It
should be ``standard'' and it should provide the best possible output quality.
The latter contradicts the former because every ``standard'' must be based on
the least common denominator. And this least common denominator is clearly not
acceptable to all the users, probably not even to the majority of them.
 
Among these quality issues is the use of fonts in their design sizes, that is,
without magnification. Font families containing all the point sizes can be
generated by the dc and Sauter packages. To make them accessible, the font
definition (fd) files must be customized. This is currently allowed by
LaTeX2e's distribution policy although the use of these (nicer) fonts gives a
different spacing. Furthermore, dvi files referencing these fonts are no longer
fully ``portable''. But neither are documents using other nonstandard fonts.
 
In my opinion, the present clear distiction between files that must not be
changed and others which may be customized is the right choice. The correct
answer to the general question is neither ``yes'' nor ``no'' but somewhere in
between.
 
The compromise has, however, further consequences. For example, the file
``exscale.sty'' interferes with fd files. ``exscale.sty'' is a style file and
as such protected by the distribution policy. I see two ways to handle the
problem:
- Drop exscale.sty entirely. In fact, it is a site-specific decision to provide
  an extended collection of cmex font sizes, not a document-related.
- Allow to customize exscale.sty. The substitute for exscale.sty should be
  obliged to serve the same purpose as the original (that is, make mathematical
  delimiters grow) but need not access the same set of font sizes.
There is no point in keeping ``exscale.sty'' fixed.
 
By the way, much worse than ``exscale.sty'' is the AMS-LaTeX package. The
``cmex10'' and ``cmex7'' options (one of which is executed) to ``amsfonts.sty''
and even ``amsmath.sty'' freeze the ``standard'' in an intolerable way. This
lack of discipline is in strong contrast to the excellent work in other areas.
 
To conclude, I think that the idea of fully ``portable'' documents is an
illusion. Certain, limited areas should remain open for customization. The last
example shows that package writers have to be aware of the user's freedom and
respect it. Burdening package writers might even be desirable: ``Liberal''
programs tend to be more robust. Remaining independent from font sizes in
particular is as important as font encodings and the like.
 
Best regards
  Klaus
 
Klaus Barthelmann, Universit\"at Mainz, Germany
barthel@informatik.mathematik.uni-mainz.de