Re: Marking up Temporal Logic

Robert Miner <rminer@geomtech.com> writes:

Date: Thu, 2 Sep 1999 11:22:56 -0500

> What operators do you need?  I am unfamiliar with first order temporal
> logic.  

Good question.

> As far as ways and means go, if you really meant a *Web document*, as
> opposed to the pseudo-paper PDF route that Fateman advocates, then
> your best strategy at the moment is still to make images of your
> equations.  

With the caveat that the choice of strategy depends on the intended
audience, I must say that I agree with Richard Fateman in the case of
the general U.S. and U.K. audience.  (Of course, this is a guess and
for the U.K. it's more of a guess.)

I assume that the average browser in the U.S. (really, only part of
the general audience) is three years old (now Sept 1996) and is
configured to spin out Acrobat for (font-bloated) PDF.

> Back in the TeX world, there is the venerable LaTeX2HTML program, but
> unless you have it installed and are familiar with it, it is
> definitely not a *quick* way of making a couple of equation images.
> another interesting alternative if you want to stick with TeX, but
> don't want to mess with images or a mess of software installation, is
> TeX2HTML, which uses a combination of HTML hacks and symbol fonts to
> fake equations in pure HTML.

(Isn't Tex2HTML the commercial version of Ian Hutchinson's "tth"?)

What will be the source markup is a very serious issue, unless one is
talking about a few casual document instances.

Humans put a great deal of time into creating marked up source.
Their archives grow.  If that source becomes obsolete five or ten
years later, it's a BIG problem.

With that in mind a good SGML language that admits down-translation to
many formats (including *good* editable LaTeX or TeX or source) is the
best theoretical approach.  It's not here yet, but it will arrive some
day despite the naysayers who claim that it cannot be done.

The second best approach is LaTeX or TeX source.  I do not expect
either to become obsolete in my lifetime.  Hence, my agreement with
Richard Fateman for today.

The most interesting approach for translating TeX to HTML is that used
by Eitan Gurari with "tex4ht": use TeX, the Program, which knows TeX
the language best, with "tex4ht" macros, which the document author
need not know, to generate a dvi with specials that are consciously
designed to pass information to a special dvi processor for
translating the dvi file into another language such as "HTML",
"XHTML", "XHTML-with-MathML", ... .  The dvi file that is created
this way is essentially equivalent to a file in an XML language that
is suited mainly for down-translation to a single format.  (But
that format *could* be one that admits multiple down translations.)

(On this point: has anyone done the interesting exercise of using this
approach to make a dvi that is suited mainly for translation back to
TeX [or LaTeX] with reasonable equivalence to the original?)

I see no reason why the Gurari special-loaded-dvi approach is limited
to outputs that are SGML languages.

Its disadvantage is its overburdening of TeX, the Program.

An SGML language (formally "SGML application") is a template for
processing.  It gives you your choice of programming language instead
of the language of TeX, the Program (as a programming language in
contrast with the TeX document markup language).

                                 -- Bill

Received on Thursday, 2 September 1999 13:23:57 UTC