MIME-version: 1.0
Content-type: TEXT/PLAIN; CHARSET=US-ASCII
Lines: 77
References: <199910311329.OAA00567@istrati.zdv.uni-mainz.de>
Message-ID:  <199911011409.PAA16801@relay.urz.uni-heidelberg.de>
Reply-To: Mailing list for the LaTeX3 project
              <LATEX-L@URZ.UNI-HEIDELBERG.DE>
Comments:     Resent-From: mjd@ams.org
Comments:     Originally-From: Michael John Downes <mjd@ams.org>
In-Reply-To:  Frank Mittelbach's message of Sun, 31 Oct 1999 14:29:20 +0100
Date:         Mon, 1 Nov 1999 09:09:14 -0500
From: "Michael J. Downes" <mjd@AMS.ORG>
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:      Combining sequences of identical elements
Status: R

Frank Mittelbach <frank.mittelbach@LATEX-PROJECT.ORG> writes:

>  > >  text text\footnote{first}\footnote{second}
>  > >
>  > >in most case this would be supposed to come out as
>  > >
>  > >            1, 2
>  > >  text text

> okay, i've convinced myself too that this boolean would be useful whether or
> not it is passed onto the user syntax level (instead one could think, for
> example, of extending xparse to be able to check for a certain command
> following and if so setting such a boolean)

I would just like to give some more examples where the idea of combined
formatting for consecutive elements of the same type might be
applicable:

  \cite{foo}\cite{bar}  -> [19,35] instead of [19] [35]

The above usage would allow each \cite to have its own optional arg,
something that is not possible with current syntax.

  \newcommand{\secref}[1]{Section~\ref{#1}}
  ...
  \secref{foo}\secref{bar}

Here one might like to have the pair of refs expand to

  Sections~\ref{foo} and \ref{bar}

or for three or more

  Sections~\ref{foo}\textendash\ref{baz}

A similar idea: author names in a bibitem (if bibitems were done with
logical markup instead of preprocessed by BibTeX):

  \bibitem{foo} \author{First Author} \author{Second Author} ...

where one might like to have consecutive author names automatically
combined into a list form "A, B, and C" or whatever. Such a reformatting
is already done by amsart.cls for the author names printed in
\maketitle but that is a different mechanism because accumulating the
data and printing happen in two different steps and there is a
definite stopping point (\maketitle) where it is known that the list
of author elements is complete.

Then consider

  \[...\]
  \[...\]
  etc

which an author might reasonably expect to be combined into a group of
equations (with differences in the vertical spacing and page-break
penalties). One might even say that the equations in a consecutive group
should by default get aligned on the relation symbols---but then it
becomes clear that in some cases the author will want to override the
default whichever way it goes: i.e., need to specify an option for the
whole group. This suggests I think that group markup would be a good
idea after all:

  \begin{eqngroup}[align=false]
  \[...\]
  \[...\]
  ...
  \end{eqngroup}

\citegroup[maybe-some-options]{\cite{...}\cite{...}...}

\footnotegroup{\footnote{...}\footnote{...}...}

(One might like to have a smart editor that automatically adds the group
markup when you put in the second footnote :-)

Michael Downes