TeX is a de facto standard (was: Dynamic Font Size)

I think Jukka's responses (below) reflect a very
different vision of how HTML should function
than the one I have.

Is it really the case that human readability of
the source file is completely irrelevant???
(If so, why such statements in the document as
"Thus, the character entity reference ÷
is a more convenient form than ÷ for
obtaining the division sign (÷). "?)

TeX has no serious competitor for the presentation
of mathematics. Software like Lyx and Scientific
Word (which took forever to become available)
achieve good results and hide the TeX, but everyone
who writes mathematics ends up tinkering with
the underlying source files from time to time
and many people---myself included---find it most
efficient to work almost entirely at that level.
This is possible because the source is easily human
readable.

Is there no lesson for HTML in this history???

There are many documents containing modest
mathematics that I would like to offer in HTML.
I always produce HTML directly, since it gives
much better results than most (all?) software.
If I have to plan on typing Α every time
I want an Α  I will just give up the idea
of *ever* presenting documents with mathematical
content in HTML. I doubt I am alone.

One last comment: must we be stuck with the mistakes
made in naming the SGML character entity sets?
I emphasize again that there exists a standard usage
that cuts across a broad community: TeX.
I would think that there a clear benefits in acknowledging
this de facto standard to the greatest extent feasible.
(Note that these characters are explicitly
"intended for occasional Greek letters used
in technical and mathematical works".)

Let me close with a naive question:
if HTML must retain the SGML entities,
and I can see reasons for that,
what prevents HTML from having a TeX
based set of entities with identical numerical
equivalents?
(If this is allowed, I'll wager you'll get 90% of
the usage being the TeX based entities,
and *many* more technical documents will be
offered in HTML.)

--Alan G. Isaac



Jukka Korpela wrote:

> The entity names in HTML specifications have been taken
> from the character entity sets listed in appendix D to
> the SGML standard (ISO 8879-1986).
>
> It's actually not a big deal anyway. The entity names are
> just half-mnemonic, and as you point out, different names
> (or "names" - is something like "isin" or "ni" really a _name_
> or just an alphanumeric symbol?) for characters are in use in other
> contexts (e.g. TeX, PostScript). Moreover, Netscape 4 does not
> support things like ∈ but it supports ∈ under certain
> conditions (see http://www.hut.fi/u/jkorpela/HTML/chars.html ).
>
> So authors might just as well use numeric character references,
> either finding them out themselves, which means some work, see
> http://www.hut.fi/u/jkorpela/html/unicode.html
> or (hopefully, in the future) using software which lets them
> enter characters conveniently and takes the trouble of inserting
> the (correct!) numeric references into the HTML code. Thus,
> I don't think future HTML specifications should extend the repertoire
> of entity names for characters.

Received on Tuesday, 30 March 1999 11:24:31 UTC