[whatwg] Mathematics on HTML5

[Henri Sivonen]

On Jun 7, 2006, at 20:54, Ian Hickson wrote:

> On Wed, 7 Jun 2006, Mihai Sucan wrote:
>>
>> Yes, it's true authors don't generally jump on whatever comes new
>> (that's the reason MathML is not as widely used as LaTeX).
>
> I would say MathML is not widely used because MathML doesn't work  
> in HTML,
> personally.

I think it is an economic problem rather than a technical problem.

It follows that I don't think the slow adoption is *necessarily*  
evidence of technical flaws.

The part of population that is interested in mathematical typesetting  
is very small in proportion to the population as a whole. As I  
understood it at the time, that's why Netscape wasn't interested in  
devoting developer time to MathML. I don't know how decision making  
works at different browser companies, but I would hazard a guess that  
from the point of view of Opera, Apple and Microsoft, the business  
case for MathML is rather lousy compared to e.g. SVG, which is known  
to have spec problems as well, which conflicts with CSS and which  
also doesn't work in text/html.

So far the track record with math typesetting software suggests that  
the people who most want it and need it are at the math and physics  
departments at universities and from time to time some of those  
people actually implement the required software as Free Software but  
it takes years. In those cases, there is no business model of making  
money off the software itself. Rather, the people are not motivated  
by money and cover their opportunity cost to a sufficient degree by  
getting paid monthly by a university (or perhaps occasionally by  
public sector grants). However, the problem with this model is that  
it is still a big deal to get a university actually pay the salary if  
the hours go to writing software (as opposed to math or physics  
research), because a Free Software math typesetter is kind of like a  
public good. It would make sense for all the math and physics  
departments in the world to pool money to fund full-time teams  
working on MathML software, but from the point of view of any single  
university, there's a strong incentive to try to be the free rider  
and hope the others get the job done.

XSLT in IE and Firefox was enough to make Apple and Opera follow,  
even though it could be legitimately argued, that transformations  
from proprietary markup to HTML should happen at server side. Opera,  
Firefox and Safari doing SVG is enough to get Chris Wilson on the  
record saying that SVG is becoming a part of the "interoperable core"  
of Web standards. Apparently MathML in Gecko alone is not enough to  
push the others to consider it a part of the "interoperable core".

Hmm. Freaky economic problems are nowadays solved with Google money. :-P

> If we made MathML work in HTML, possibly with rules that make
> the syntax easier (by implying tags as I suggested earlier)

The implied stuff seems scary. I was hoping for no more tag inference  
beyond HTML 4 legacy.

FWIW, I completely agree with James Graham that automatic conversion  
from LaTeX is *the* top-of-the-list requirement for any kind of Web  
math. (It follows that it is futile to insist on semantics that you  
can't pull out of LaTeX as it is normally authored.) I gather that  
TeX4ht is the state of the art here. That already puts MathML ahead  
of anything else that WHAT WG could come up with. Running code rules.  
(Forget latex2html.)

If the WHAT WG really intends to address math, I think it would make  
sense to start by interviewing Roger Sidje, Jacques Distler, Eitan M.  
Gurari and Robert Miner to find out what they think are the problems  
that need to be solved (if any).

-- 
Henri Sivonen
hsivonen at iki.fi
http://hsivonen.iki.fi/