Message-Id: <199207030809.AA29179@mail.cs.tu-berlin.de>
Reply-To: Mailing list for the LaTeX3 project <LATEX-L%DHDURZ1@tubvm.cs.tu-berlin.de>
Date:         Fri, 3 Jul 92 10:04:58 CET
From: Frank Mittelbach <MITTELBACH@MZDMZA.ZDV.UNI-MAINZ.DE>
Sender: Mailing list for the LaTeX3 project <LATEX-L%DHDURZ1@tubvm.cs.tu-berlin.de>
To: "Rainer M. Schoepf" <Schoepf@SC.ZIB-Berlin.DE>
Subject:      resend re: radical changes needed?
Status: R

> Subj:	radical changes needed?

> What is wrong with \footnote{text text text}?
>
> Once after the argument of \footnote is read, the catcodes are fixed. This
> produces problems in mixed language documents, e. g.
> \footnote{text \begin{greektext}text\end{greektext} text}
> if the catcodes differ for the different languages. This is, however,
> common. Therefore a syntax like
> \begin{footnote}
> text text text
> \end{footnote}
> is preferable.

As already remarked in the Cork proceeding we intend to allow for
every ``text producing command'' an environment form and vice versa.


> I know, that this kind of change produces compatibility problems between
> LaTeX2.09 and LaTeX3, but I don't see any other way out of the dilemma
> described above.

Such a change wouldn't produce compatibility problems there are other
areas which are much more difficult to handle.

> Consequently, any text producing command (those with the famous moving
> arguments) need to be changed. There are so many and so common ones ...
> \chapter \caption \footnote...

However, in case of so called moving arguments there is simply no
perfect solution that avoids fixing catcodes and at the same time
allows \verb etc. The need to scan the argument of a heading, etc.,
into a tokenlist without real processing makes this nearly a must. Of
course there are ideas all over place ranging from writing to and from
a file, etc., but all such ideas that I know of have defects in one
way or the other.

Anybody out there who thinks he/she has a new piece of could that is
really stable? We would certainly be interested to know about.

As an alternative one can prevent the use of catcode changes outside a
well-defined interface and this may be the solution we will finally go
for.

Frank Mittelbach