MIME-Version: 1.0
References: <v03110702b268dfad082e@[130.237.37.120]>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Message-ID:  <364340A4.3A5CE73C@na.uni-tuebingen.de>
Reply-To: Mailing list for the LaTeX3 project
              <LATEX-L@URZ.UNI-HEIDELBERG.DE>
Date:         Fri, 6 Nov 1998 19:32:04 +0100
From: Marcel Oliver <oliver@NA.MATHEMATIK.UNI-TUEBINGEN.DE>
Sender: Mailing list for the LaTeX3 project
              <LATEX-L@URZ.UNI-HEIDELBERG.DE>
To: Multiple recipients of list LATEX-L
              <LATEX-L@URZ.UNI-HEIDELBERG.DE>
Subject:      Re: pattern matching in LaTeX
Status: R

Hans Aberg wrote:
>   The whole problem is more complicated, because it is not only a question
> of expansion, but also when things should be expanded: Sometimes
> immediately, sometimes later. The correct way around this would be to
> define a stricter input syntax which separates the elements authoring
> semantics, typesetting style elements and typesetting fine-tuning, but

This is basically what I was thinking of.  I couln't find an argument
for needing more than 2 tiers in the expansion process, but what I was
describing is certainly generalizable.  So precisely why do you think
one needs to distinguish "authoring semantics", "typesetting style" and
"fine-tuning"?

> there is no way to enforce such a syntax in TeX.

This is clear.  Question: how difficult would it be to extend TeX to
allow for dual/multiple tier expansion?  What are the draw-backs?

Marcel