On 06/03/2011 14:03, Arno Trautmann wrote:
> Well, first, it would show the development. I know several people who
> have lost interest in following LaTeX3 development because there is no
> visible progress.

Unfortunately, 'progress' in this regard I suspect means 'an actual
LaTeX3 format'. Can't promise that any time soon, I'm afraid.
>
Also, it is unpleasant to tell someone about how great
> LaTeX3 syntax is but having to tell they have to load a package on top
> of l2ε.

Well that's also true of any other change - LaTeX2e will not alter, so
it's not as though they'll simply be able to expect 'LaTeX' to suddenly
gain new abilities without either (1) loading packages or (2) using a
different format.

> Second, the loading of expl3 stuff would be much faster which would be
> favorable especially for short tests – often the preamble takes much
> longer to load than the actual compilation time of the document.

That is true :-)

I'm not opposed to the idea, so here's something to think about. My view
is that for a LaTeX3 format the approach should be rather more similar
to ConTeXt than LaTeX2e when it comes to engines and output mode. So
I've imagined

  $ latex3 --dvi --pdftex <file>
  $ latex3 --pdf --luatex <file>

an so forth (with --xetex ignoring --dvi for the obvious reasons).  Does
a similar scheme make sense for a hypothetical 'latex2x'? (I'm going
with 'x' for 'extended', and also for 'like LaTeX2e, but clearly a bit
further along. Of course, there would need to be some defaults for the
above - I guess I'd favour pdfTeX in PDF mode at present.

Second question: anything else that should be included that is not in
the combined 'release' material (expl3, xparse, xtemplate, xcoffins)?
These do load various bits and pieces (for example, graphicx), but I'd
like to at least add fixltx2e to the above.
-- 
Joseph Wright