Re: www-html-d Digest V96 #224

Thomas Breuel (tmb@best.com)
Wed, 17 Jul 1996 22:07:39 -0700


Date: Wed, 17 Jul 1996 22:07:39 -0700
From: Thomas Breuel <tmb@best.com>
Message-Id: <199607180507.WAA29822@shellx.best.com>
To: www-html@w3.org
Subject: Re:  www-html-d Digest V96 #224

> Folks don't seem to be suitably impressed with Ping's work.  Here's a
> guy who, rather than bemoaning the lack of HTML math, rolled up his
> sleeves and built something. Not only that: he did the 3x more work
> that it takes to make it usable by others.
> 
> I can tell you that, as a result, this is having real influence on the
> design of HTML math.
> 
> Is anybody using it? I would expect to see questions about it,
> and I don't see them.
> 
> I've looked at MINSE, and so has Dave Raggett. I think the design is
> really clean and elegant. Ping: what did Dave say about it? I forget.

There are some serious problems with basing HTML math on semantic markup:

     -- defining the structures and their renderings used in each 
	branch of mathematics, applied math, physics, chemistry,
	computer science, economics, etc. is a huge undertaking,
	and authors frequently make up new notation on the fly
	(MINSE lacks even constructs necessary to typeset 
	the formula from the cover of my freshman year mathematics
	textbook)

     -- a lot of mathematical typesetting is done by people 
	(secretaries, professional typesetters, part-time student
	help) who just don't know much about the structure of the
	math they are typesetting; requiring structural markup
	will force authors to do the typesetting themselves

     -- large amounts of existing, on-line math is not in a
	structural representation and cannot be converted
	automatically

     -- even worse, hundreds of years of mathematics need
	to be converted to on-line form; with a layout language,
	staff can do this with moderate training; structural markup
	will require much more mathematical sophistication

     -- it's yet another notation people have to learn, a notation
	with orders of magnitude more symbols

     -- the premise that you need structural rendering in order
	to represent mathematical notation in a text to speech system
	is likely to be wrong; TeX-like notation may not literally what
	you would use to describe a formula in natural language, but it
	is a reasonably abstract linear rendering of mathematics that
	you can understand even without typesetting it; this is not at
	all analogous to in-line images; in fact, there is no a-priori
	reason to believe that TeX-like notation is worse than 
	automatically generated natural language for reading out a 
	formula (and authors can supply ALT= markup if they choose)

I would like to see a standard, supported way of typesetting
mathematics on the web.  Experiments with novel and unproven ways of
structural markup are going to put standard mathematical markup for the
web in jeopardy, because there is a strong risk that browser vendors
won't implement it and that users won't or even can't use it.  The HTML
3.0 math proposal, based on LaTeX, integrated with the notational
conventions of HTML, seemed like a usable starting point; maybe a
little bit more support for semantic markup, including a simple macro
facility for defining semantically meaningful macros, can be added to
it gradually over the years.  Unlike MINSE, I actually can typeset even
the formula on the cover of my freshman year math textbook with it.

Thomas.