LATEX-L Archives

Mailing list for the LaTeX3 project


Options: Use Classic View

Use Monospaced Font
Show Text Part by Default
Condense Mail Headers

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

Print Reply
Sender: Mailing list for the LaTeX3 project <[log in to unmask]>
From: Frank Mittelbach <[log in to unmask]>
Date: Sat, 17 Feb 2001 18:10:18 +0100
In-Reply-To: <[log in to unmask]>
Reply-To: Mailing list for the LaTeX3 project <[log in to unmask]>
Parts/Attachments: text/plain (114 lines)
Marcel Oliver wrote:
 > David, some question for clarification:

since I can't find his answers here are mine (and he can disagree with me:-)

 > David Carlisle writes:
 >  > If latex switched to use omega (only) then
 >  > a) this might require omega to be more stable than omega users would
 >  > wish, ie it might prematurely limit addition of new features.
 > - Is the envisioned time frame for Omega to stabilize really longer than
 >   a realistic time frame for LaTeX3 to be completed?

probably not, but you can't build on something if it changes below you (or at
least that is pretty difficult).

 >  > b) it would cut out people using tex systems that don't include omega.
 >  > You might say they should all switch to web2c tex, but that's like
 >  > saying that everyone should use emacs on linux. Clearly it's true, but
 >  > it doesn't happen that way.
 > - What are the platforms that can compile TeX but cannot compile
 >   Omega?

it's not only the question of the platform it is also the question of the
system (eg the commercial ones). While I don't think this is really important
if you come up with good reasons for changing so that the people on free
distributions change it is a factor not to overlook.

 >  > c) special case of (b) it would (at present, I think) cut out pdflatex.
 > - How much of an advantage is pdftex compared to creating pdf via DVI?
 >   I have only done the latter without any problems, but of course it
 >   involves more file format conversions.

not only that, but "rich" as produced via pdftex with hyperef might be  more
difficult to provide (I might get corrected here), eg can you easily produce
references etc via the dvi -> ps -> pdf route right now? that might become a
non issue but I guess there is an advantage when you can control what is
produced for the target format directly.

 >  > d) It would require reasonably major surgery to LaTeX internals. It
 > Now it's getting interesting...
 > - Is it basically clear to the LaTeX experts what needs to be done, or
 >   will major conceptual work be necessary?

difficult to say. this partly depends on the underlying features you finally
build upon. See our discussion of LICR viz OICR1/2. while I consider Omega far
ahead of TeX in what ocps can be used for, they operate on OICR2 while there
isn't enough control for OICR1 which is sort of the equivalent of LICR. So
LaTeX right now is ahead there and finding concepts that work on top of the
current Omega model, would be difficult. That might vanish if the model is
changed in this respect.

 > - How do such changes compare with what is being done anyway for
 >   LaTeX3?

for a lot of areas where we try to make progress for LaTeX3 Omega doesn't
concern itself with, so we would struggle there as well as within the bounds
of TeX. eTeX has a few items that would help but eTeX as now isn't an engine
you could build upon

 > - Will things get harder or easier with Omega?

neither, as of now, essentially just different in a few areas

 >  > would be possible to make documents and packages using "documented
 >  > interfaces" still work with a new internal character handling, but
 >  > ctan will reveal a lot of heavily used packages that for good (or bad)
 >  > reasons don't use documented interfaces, but just redefine arbitrary
 >  > macros. (Often because there isn't a documented interface).
 >  > A lot of these would break.
 > Again, it may be good to assess the extent of potential damage.  The
 > switch to 2e has broken lots of stuff, but in the long run it was the
 > right thing to do.  So one should make a strong case why it should be
 > otherwise now.  (And maybe some packages actually deserve to die...)

actually most stuff did work after the switch from 209 to 2e, but i expect
more things to not work after a switch from 2e to X

I personally think that that switch needs to be accompanied by a heavy
organised rewrite of external packages (perhaps even by financially supported
rewrite initially)

 >  > So in short to medium term it seems there have to be two versions
 >  > latex/omega and latex/tex. How compatible they can be as latex/omega
 >  > uses more omega features I am not sure.
 > This is a situation we should REALLY avoid!

you can't at this stage avoid the fact that Omega can do things that TeX can't
just like pdftex or etex can do things TeX and Omega can't do.
And some of the things Omega can do require kernel level internal bits being
implemented quite differently.

as long as such features are kept at bay (eg pdftex features are essentially
hidden by the package hyperef) the situation is relatively relaxed, ie
documents not using \usepackage[pdf]{hyperref} can run on LaTeX based on TeX,

But the moment you start providing alternate syntax for the source document
you end up making perfectly valid document  sources impossible to use with
different LaTeX implementation and that creates very unnecesary islands. This
is comparable to producing a document class that is essentially encoding the
functionality of article.cls but changing \section \subsection etc to