LATEX-L Archives

Mailing list for the LaTeX3 project


Options: Use Forum View

Use Monospaced Font
Show Text Part by Default
Condense Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Mailing list for the LaTeX3 project <[log in to unmask]>
Wed, 9 Mar 2011 11:38:40 +0100
Mailing list for the LaTeX3 project <[log in to unmask]>
text/plain; charset=unknown
Frank Mittelbach <[log in to unmask]>
text/plain (128 lines)
Arno Trautmann writes:
 > I guess this question is raised from time to time $,1rs(B now I am asking ;)
 > As far as I understand, there is still no own LaTeX3 format because the
 > code just is not ready. But expl3 is more or less stable and some
 > important packages (e.g. fontspec) make heavy use of it causing many
 > people to load expl3 at the beginning of their document. As this is
 > somewhat time-consuming, Philipp Stephani suggested about a month ago on
 > the lualatex-dev list the introduction of a "LaTeX 2.2" format with
 > expl3 and xpackages (and fontspec) preloaded.
 > So what is the "official" meaning of the LaTeX3 team? Are there any
 > plans for a test format? Or is it not intended/possible/ ?

There has been already some discussion on this with various suggestions and
I'm a little late in joining the club, but here we go.

In my opinion there are a number of factors that need to be weight against
each other.

speed was mentioned as a factor

well, on the old machine I'm currently using running a manual load of expl3,
xparse, template and xcoffins takes something like

real    0m0.511s
user    0m0.449s
sys     0m0.027s

give or take, ie around half a second. Now I can remember using TeX when took
half a minute to typeset a single page so I can't really get very much excited
saving half a second but ... :-)

visual sign of activity was mentioned

now to me that has a higher rating. The question though is, would an
additional format actually provide this?

In my opinion only if the format is more than just a random collection of
additional packages to speed up usage for testing and if it is "relatively"
stable so that people actually "use" it.

I said "relatively" stable. By that I mean that overall the additional
functionality provided is something we feel is "likely" to available in LaTeX3
even if we currently make no guarantee and occasionally update/change things
so that packages based on expl3 need to follow to work.

In my opinion base "expl3" by now provides this stability at least if we are
careful enough not to start adding arbitrarily add functionality without
initial experimentation as xpackage (as there was recently some tendency). So
in this respect a format that provides LaTeX2e + expl3 + a few carefully
chosen xpackages (eg xbase) would fit this bill.

adding additional external package was mentioned

sorry no. in my eyes that would be counter productive. any package that we add
here should pass the test:

 - we feel that this is a candidate (from interface perspective) for LaTeX3 
   (the occasional change of mind not withstanding)

 - it is fully implemented in expl3 and it is properly tested and documented
   (ignoring that to my standard that rules out all or most of our code right
   now :-))

And none of the suggested packages mentioned fit that bill, not even fontspec
I would claim at this point in time.

I would probably make an exception for fixltx2e because this is fixing the 2e
kernel but that's about it (if at all).

processability of documents was not mentioned

one point I was missing so far in the dicussion. By that I mean that I
wouldn't want to force people to have to use this format to run documents and
in the opposite direction I wouldn't want documents which have been written
using the format not to work on an installation that doesn't have this format

I'm a strong believer that a document should state what it needs, 2e has
proven successful because it stopped all that mess with a dozen different
formats all having different code loaded and where you had to deduce from the
content of a document what format to use to process it.

Now you can say that if somebody starts of  the document with


say then everything will be fine whether or not the document will be produced
by ordinary 2e format or 2x.

But history will tell us that what is likely to happen is that people start
writing documents that will not \RequirePackage/\usepackage

And I don't like to see a repeat of what happened when we introduced iLaTeX
nfssLaTeX, nfss2latex ... in the early '90

there are a couple of possibilities how to guard against this at least to some
extend, but it would need some thought and preparation

a 2x format as a promoted default wasn't mentioned

a step further to the discussion would be to start promoting a 2x format that
is 2e + expl3 but expl3 only sitting dormant usable only if explicitly
requested by the document. There is a danger of 2e becoming instable this way
but this could probably be mitigated. Something to think about.


in summary I think that the time is ripe to do something like this, but I
still maintain that it needs to be carefully done to not convey the wrong
message or lead to a mess incompatible formats and unprocessable documents