LATEX-L Archives

Mailing list for the LaTeX3 project

LATEX-L@LISTSERV.UNI-HEIDELBERG.DE

Options: Use Forum View

Use Proportional Font
Show HTML Part by Default
Show All Mail Headers

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

Print Reply
Subject:
From:
Frank Mittelbach <[log in to unmask]>
Reply To:
Mailing list for the LaTeX3 project <[log in to unmask]>
Date:
Sun, 22 Jun 1997 23:14:04 +0200
Content-Type:
text/plain
Parts/Attachments:
text/plain (253 lines)
I did send this message around midnight yesterday making sure that it
was well below 500 lines (something like 480) but forgot that the
header might get included; so this evening i found it again being
rejected by the listserv.

so here i try again this time in two parts
frank

------- start of forwarded message (RFC 934 encapsulation) -------
To: Mailing list for the LaTeX3 project
              <[log in to unmask]>
Subject: Re: The LaTeX, e-TeX, NTS & Omega projects: a composite response
Date: Sun, 22 Jun 1997 23:05:36 +0200


i promised myself to write some comments to Phil's long message by the
end of this evening, so i better start. He brought up many important
points and opinions and at least with those where i disagree ...

let me first start by saying that one impression i had throughout the
mail was that Phil thought i suggested that e-TeX should incorporate
concepts found in omega. I don't. e-tex is e-tex with its goals its
research and its team and omega is omega with its goals, research and
team and neither project should or probably can jump over its own fences.

Philip Taylor writes:
 > [from SPQR]
 >
 > >> surely the present discussion shows whats missing? a solution to the
 > >> problem of multi-language hyphenation in the same paragraph?
 >
 > That is _one_ element that is missing; my question was intended to
 > provoke a rather fuller response!

well, Sebastian gave a crucial deficiency resolved by Omega --- one
that more or less shuts the door for every potential TeXlikesystem
user who is not using latin based alphabets or who needs more than one
language in the document.

 > There is no rivalry whatsoever.  I have the greatest respect for John & Yannis's
 > work, but I cannot see how trying to merge the projects _at this stage_ would
 > be possible or even desirable: a small team works well together -- as the team
 > gets larger, it gets ever harded to achieve a consensus.

i wasn't asking for enlarging the team. my main claim is that there
are valid and stable solutions within the current omega and different
valid solutions with e-tex and that those could be brought together
(if you like by a completely different team and as small as you wish)
while the omega project as well as the e-tex project can research
further in those areas that are of interest to them and are not
resolved.

 > with Thanh's TeX2pdf/pdftex: I would far sooner that he continued to develop
 > his ideas independently, rather than trying to to "bring him under the e-TeX
 > umbrella" at this stage.

exactly, i don't want to have things under an "e-tex umbrella"

 >  Let us see where the various projects lead, and
 > _then_ (perhaps) think about a synthesis of the best elements of each.  NTS
 > is due to start in 1998: by the time we have a functional implementation,
 > we may be better placed to analyse e-TeX/Omega/TeX2pdf and develop a synthesis.

my claim is that e-tex and omega are at a point where we do have the
first stable results of what they lead to and it would be about time
to bring that what was achieved so far together. As such bringing
together takes time of its own, i'd expect that a stable pdf could be
incorporated as well (which could then be part of the for promotion
the combined version)

 > >> switching eTeX to Unicode would be nice...
 >
 > I do not believe that this will ever happen, nor do I regard it as desirable.

neither do i (unfortunately) --- give or take that we don't mean
unicode but really just a large enough unique internal encoding
(preferably close to unicode as this makes life easier in some respect
but isn't really the essential part of the change)

but again this is seeing the whole situation from the e-tex umbrella.

 > The primary desideratum of e-TeX is 100% compatibility with TeX; I do not
 > believe that that could be accomplished in a system which used Unicode.
 > Unicode is an option for NTS, not for e-TeX.

that is something i really don't understand. perhaps that is something
you really should look at it a little more closely. why is that
different to, say allowing for many mark registers, what kind of
compatibility do you fear you lose?

no, don't answer that one here unless it is short :-) but perhaps we
should go through this in SF.


 > [from Frank Mittelbach]

ah from me ;-)

 > >> but then a frozen version so that one could reliably use the features
 > >> in LaTeX and other formats.
 >
 > It is intended that e-TeX will be updated once a year, with no
 > feature being removed from a later version unless there is are
 > overwhleming reasons so to do.  On that basis I would argue that it
 > is "frozen" enough, since LaTeX comes out twice as frequently.

but this is a different situation (i'm not talking about about real
bug fixes). LaTeX(2e) is coming out twice a year but we have stopped
modifications on the kernel as we did during the first release as at
that time some of the intended features where not in place yet.

others have explained by there is a difference in installing a new
latex version to updating the basis environment on which it is based
on.

so no i don't think that this is frozen enough. i think what you want
is as switch of the basis if at all and then give the users given
breathing space and the project teams (ie e-tex, omega, nts) time to
arrive at the next level of development (if any).

 > >> Phil has asked what features i miss that omega has. i'm not sure that
 > >> this was a serious question (you should know what your competing
 > >> successor is capable of, shouldn't you? :-)
 >
 > What it's capable of is not the issue; what features _you_ (the LaTeX
 > team) need is really the point, but you answer that next:

need for what and by whom? that really is the crucial question here. 99% of what
e-tex offers is something we do welcome but it is not just what we need for
developing it is what is necessary and helpful to the end user (and to the future
of TeX)

 > >>  - full internal 16 bit (ie unicode support)
 > >>  - thus true support for multi-lingual issues, eg proper hyphenation,
 > >>    encoding support, rewriting support
 > >>  - otp's
 >
 > In other words, Omega!  But I find it difficult to reconcile this
 > with your own arguments adduced earlier, about the difficulty of
 > persuading people to use a new LaTeX predicated on e-TeX; I would
 > have thought that it might prove about 100 times harder to persuade
 > people to use a new LaTeX predicated on Omega, given that e-TeX itself
 > is predicated on complete compatibility with TeX whilst Omega rejects
 > compatibility in order to achieve more radical enhancements...

why do you find this so difficult? my main argument is that e-tex on
its own is something where most users don't see benefits in and the
same is true for omega on its own (although in the latter case there
are more visible advantages from day one and completely new groups
that would benefit by being able to use the system). but in
combination they are more likely to have the desired effect.

and much more so if major formats like latex would be ported to such a
combined version with immediate benefits to nearly everybody.

okay, you are wondering about a compromise of compatibility, but i
think this is at least partly a misunderstanding; i don't see that
such a combined version would be much different as far as
compatibility for old documents are concerned to the situation with
e-tex right now. doable (worth doing, for sure) and not a really
problem.

can you put forward what kind of compatibility you see compromised? or
again save it for SF after having looked into it a bit. I really think
you argue a bit from public statements that may have been given by the
omega team rather then from valid "real" data of what that means. In
other words, the fact that somebody may have said that he doesn't care
about compatibility doesn't necessarily mean that his product in fact
results in being incompatible.

so coming back to Sebastian's question of rivalry: have you ever run
omega to prove that point? if not, i think my argument that both
projects actually worked on orthogonal problems and that it is about
time to combine the fruits and use them (carefully and in a compatible
manner) becomes even stronger.

let me say one thing again:

with my latex development hat on (trying to tidy up what we have
right now and producing in some restricted areas (like marks)
something that is not yet here) i would welcome (in fact be delighted)
in being able to base the LaTeX kernel on a system that contains the
e-tex primitives.

but if omega and e-tex as of now would be equally available to the
public and i would have to make a decision between those two (and the
public would follow that decision) then i would (probably with a few
tears in my eyes) tend towards omega because it would solve more for
the end users and for the TeX community as a whole --- in opening up
the asian market and solving all (oops, most of) the language problems

now, omega and e-tex aren't in that position (omega even less than
e-tex probably) so i don't have to choose and i can't. and in fact i
don't want to because either decision would throw away the important
achievements of the other project.

therefore my plea to combine those stable parts *now* ---- meaning within one year
being productive

alternative right now seems to be staying with vanilla TeX and
implement further work on that level as well --- yes i do welcome
seeing a reimplementation of the current kernel with e-tex primitives
and i would support and encourage such a project a lot (similarly
development of add-on package) but it would be something that would
need to be done by people outside the ltx3 project as i couldn't put
our limited resources into that.

and this idea of best typeset with Omega-TeX, e-TeX,
e-TeX-version.1.3, ... really is a horror to me. Not for dedicated
applications but if it would become the standard situation.


 > I agree completely with Bernard's sentiments; at the moment, I feel it
 > is best that e-TeX, Omega and TeX2pdf evolve separately, each under the
 > control of a small, tightly-knit group; trying to pull them together
 > at this stage would simply result in friction, with each team pulling
 > in a different direction.  Let's wait for them each to become mature,
 > and then look towards a synthesis.

IMHO they are mature as they can get (as far as the projects are concerned)
and again, yes if i would imagine you and John shut in a room on bread
and water until you reach a common understanding, i'm sure both of you
would rather starve then to give an inch :-)

but this is why i don't see either project being an umbrella for the
other, your interests are just so different.


 > [from Uwe]
 >
 > >> I think, there are not many users right *now*. *But* there is the
 > >> TeX-Live-CD version2, on which will be an e-TeX (or the announcement
 > >> is a lie...) and the new version of teTeX (0.9) will have an
 > >> e-TeX. There are quite a large number of people following *and*
 > >> updating with this distribution.
 >
 > >> So I expect the number of people actually having e-TeX will be
 > >> significantly rising in the *near future*.
 >
 > I agree; the takeup of e-TeX was not helped by there not being an
 > emTeX version, but Eberhard has his own work and life, and can't
 > be expected to bring out new versions just to help the e-TeX project.

no he can't but even if he would this is not enough. having an etex
within tetex is a step but the next step need to be taken as well,
such as having the proper formats coming with the setup. for example,
the new texlife cd has everything on it, but when i want to use latex
with e-tex i have to compile it myself and there is no direct support
for doing so.


 [  .... continued in next mail ]

ATOM RSS1 RSS2