Marcel Oliver writes:
> Sorry, I think the discussion is going in circles. What's the point
> then to improve LaTeX if I should author in XML? I don't give a damm what
so that the typesetting is nicer
> typesetting system I use, as long as it is widely available, at least
> as powerful as LaTeX is, and it doesn't cost me any more effort to
> author documents (after a reasonable initial learning period).
how about separating typesetting from editing and manipulation? its
not the only paradigm
> So should we improve LaTeX? Just as a backend to an XML based system
> it wouldn't be worth doing anything more than removing a few quirks from
> the TeX engine. So what are we doing here???
do you think LaTeX does floats and multiple columns perfectly? I
dont. I think it can improve a lot as a batch formatter
> Why can the AMS get my files on paper without retyping, and a huge
> big-budget operation like Elsevier can't?
cos the AMS use TeX to typeset. why dont Elsevier's typesetters?
why dont we watch Betamax videos?
> But paper is the only thing you can do better than I can. I can run
> a web-site with no problem and publish my stuff in whatever format I
indeed. if this is the world model you (the academics) want, go for
it. i have no shares in publishers....
> > nightmare, since almost anything can happen. maybe it takes 5 minutes,
> > maybe it takes 5 hours. you cannot run a business on those lines, IMO.
> Why don't you make the converter free software? Then you can go and
> tell everyone to check first if a file is convertable before submitting
> it. You should get a much better quota than 50% this way.
1. if people cannot read Lamport and get the basic markup right, you
think they'll bother to test my converter?
2. at present, our converter is slightly complex to install, and we
cannot afford to `support' it externally
3. we *are* a business. we dont want to hand solutions out to other
companies _quite_ so easily
but i take the point
that said, the principles of the conversion are explained
publicly, and you can (these days) write a better one using TeX4ht
or Omega (IMHO. next years project.)
> Or is this one of the cases where "big" companies prefer to use
> undocumented and proprietary formats just for their own monetary
in this case, no. argument 2. is the real one - we could not possibly
respond to thousands of authors saying "remind me how to install a
Perl extension, what is ngsmls, i dont have a copy of flex" etc etc
anyone who really wants this animal is very welcome to ask. i dont
guarentee getting permission, but i'll ask.
> Also, I don't buy your argument about the impossibility of converting
> LaTeX into other formats. A clean LaTeX file (i.e. no TeX which is
> not explicitly part of LaTeX) and a conservative choice of packages
> (which is what this discussion is all about) should allow very close
> to 100% conversion.
yes. our program will do/does do a 100% conversion. one seldom sees such
files outside the testbed, however.
> production process of your organization. To me it seems that (at least
> at some point in the past) either TeX or LaTeX has been the back-end
> (I may be wrong here, but some things looked very TeX-like).
some processes and typesettters have used TeX, yes
> the point of going via XML as an intermediate format? Why not let
> people submit in clean LaTeX for those who prefer to write clean LaTeX,
> and typesetting it directly, and let others submit their RTF or XML or
> whatever, and feed it to whatever backend?
because we want a stable archival form, which has to be XML/SGML.
for instance, we absolutely must get HTML out there for all articles, and the
production flow from archival format
must be _smooth_ and _uniform_. we are talking about
100s of 1000s of articles. we cannot afford to have 5-10% of the
articles done with a rogue system like latex2html.
> And, repeating myself here, I believe that a publisher cannot "add
> value" to papers in my field (applied analysis) by storing or distributing
> publications in any format which is much more explicitly marked up
> (I guess that's what XML is supposed to do) than LaTeX.
oh, i agree. good LaTeX is very good. no argument.
> So the final question: Why do small publishing operations (professional
> societies and university presses) who are closer to the scientific
> community and whose journal subscriptions are usually cheaper, use
> LaTeX more readily than the large multinational presses? Can we
because they serve the academic communities who use LaTeX? because
they dont maintain long-term archives? re-use data for large Web
databases like Science Direct?
> conclude that LaTeX is doing the job pretty well after all?
LaTeX is fine, great. like the C language is fine. but do we believe
that large software engineering projects should use C, or a language
like Java or Smalltalk?
> technical discussion. I think it is in everybody's interest that
> whatever future typesetting system is acceptable both to authors
> and publishers. But if publishers are neither very forthcoming in what
let me reiterate, publishers do not typeset! typesetters do
that. publishers are *agents*, like bankers, shifting stuff back and
forth and taking a cut.
> they need, nor really listening to the author's needs, then at some
> point one just has to say "screw them and let them find somebody
> cheap in the third world to retype all the manuscipts".
and then what will you do with them? put them in the monastery
i feel a bit awkward here. i do get paid by a large, very commercial,
publisher, and i feel obliged to explain their viewpoint. equally, i
can see that a future world might have no place for us, if everything
exists only on tiny individual web servers. who knows?
i do have one last (trite) remark - why are you (the academic authors
reading this) spending so much time on arguing about typesetting and
publishing? why don't you spend my tax money on doing research into
what do i care. if Marcel's institution dumps big publishers and sets
up its own business, they'll have to employ people to do the work, so
i hope i'll get a job until i retire. then Marcel will be the evil
money-making employer (you watch).