[whatwg] Mathematics in HTML5

[juanrgonzaleza at canonicalscience.com]

Ian Hickson wrote:
> On Mon, 19 Jun 2006 juanrgonzaleza at canonicalscience.com wrote:
>>
>> 1) It has been proven that via Standard CSS 2.1 not designed for math
>> one can render math better than browsers with native support (as
>> Firefox  1.0) and infinitely better than MSIE, Safari, and Opera
>> (rendering  natively zero mathml).
>
> I don't think so. In my experience Mozilla renders maths in a way much
> superior to the CSS approach. I'm also not convinced that the CSS
> approach  is good enough for people to use instead of GIFs. People want
> things to be  pretty, that's why people abuse HTML for layout purposes

Well, as said I will add some snapshots in canonical science today and
people could decide that approach they prefer.

Moreover, I have asked in many times what technical limitation of CSS does
that rendering of mathematics cannot be true. Here one would separate
George specific CSS techniques from CSS rendering.

Or in other words, what part of rendering Mathematics on the web could not
be implemented via a CSS layout engine?

> so much. There's  no point providing a solution that people won't use --
> so before we add it  to the spec, we have to be sure it will be used.

Of course, I agree!

But adding it to the draft of the spec could be a good step for receiving
feedback from community. If they reject that part of the draf then
mathematical markup would be eliminated from the final spec. If many
voices claim for a minimal cheap mathematical markup without rely on
abominable MathML then success would be guarantee.

In this list I have heard more “pro” voices. Maybe statistic can be
applied to the rest of community.

>> This philosophy was not applied to the rest of HTML5 spec, no?
>
> Yes. Canvas was only added once Apple proved it was implementable and
> popular, and had documented it.

Popular in the web?

I think that it has been proven that CSS approach to Mathematics is
implementable and is well documented (note that is not need for special
parsing modes, mstyle extensions or DOM extensions). Is popular?

Well, I entered [math CSS] in Google and returns several pages with
different techniques, approaches. Apparently people is interested in
rendering of mathematics via CSS. I know most of people think that MathML
is the only approach to mathematical rendering. Therefore, interest could
be greater when that online myth broken.

I have also been trained that Mozilla developers have add CSS mathematical
extensions to the browser, e.g. -moz-math-rowline CSS rule, but I do not
know anymore.

> Drag and drop was added based on both IE
>  and Safari implementing a compatible API that is used on the Web, and
> that  is widely documented.

Contrary to MathML, the CSS approach reuse available well-documented “drag
and drop”.

> Most of the new elements -- <footer>, <nav>,
>  <header>, <menu>, etc -- were based on a survey of over one BILLION
> documents to see what authors were using/needing.

Typping [math online] on Google returns some 10^6 of webpages.

Typping [mathematics online –math] returns some 10^6 more.

Typping [need math -online] returns some 10^6 more.

Now add scientific and engineering communities. I think one obtain
something close to 10^7 documents. Whereas I am not claiming that <math>
was more popular than <menu> I still think that <math> is a primary need
on the web.

> The event and personal
> information sections are waiting on feedback and experience from the
> Microformats community.

HTML5-Math could wait on feedback and experience from the CSS-Math community.

> The parsing model is based on extensive
> reverse-engineering of the four most widely used browsers; the browsers
> used in the study themselves were picked based on extensive research.

HTML5-Math is based in standard approaches availables in any modern browser.

> control. The list goes on.

But i) HTML5-Math reuse available code and standards before reinvent the
wheel (MathML way) ii) HTML5-Math is based in ISO-12083 (broadly
researched and implemented in SGML world) with minimal changes for
adapting it to web. iii) Your own approach with special parsing model is
not based in extensive research (at least at my best current knowledge).

> Yet, in the real world, we have to worry about such things.

I said not the contrary (in fact one of my points against MathML was that
it is very expensive), just that you calculation appears “strange” to me
because the CSS approach is more cheap.

At least other guy reacted in a similar way to you looking-strange
calculation of costs.

>
> My intent is that they would be the same <mrow>, but I admit I haven't
> explained my proposal fully. This is mostly because I am not convinced
> that MathML itself is "good enough" either.

Additional reason for the research in alternatives. I could understand
that people want delay this for a more comlete discussion and maybe
HTML-Math could be implemented in a hypotetical HTML6 (this could be
discussed) or in an alternative XML approach –somewhat as OpenMath is
already a better option to content MathML 2.0-.

Any case you could reuse the ideas, proposal, and inertia began here.
Apparently, there exists a strong need for online math has been not solved
by current w3c approach.

> Using MathML has the advantage that it already exists, and so requires
> no  mathematical expertise on the part of WHATWG. If we want to invent
> our own  language it will take many months if not years, with scientists
> from all  kinds of disciplines being involved to get a good language,
> etc.

Exists only in a browser. Moreover, also XHTML1 exists but you are
impulsing a HTML5. XSL-FO exists but is not very popular at this lists,
etc. TeX and ISO-12083 also exists for decades but you are not reuding
why? Because “it already exists” is not all the equation.

Note we do not begin from zero, we began from an international standard
ISO-12083 and adapt it to modern web. Moreover, I do not find consistent
the claim for the need of years of debate whereas implementing a language
(MathML) is not popular. Sorry to say this but I think that you are
oversizing the problem, since if my thoughts are heard we would focus just
in structure of equations.

> Because I don't think WHATWG is an appropriate venue for a task of such
> huge scope, I would prefer if that happened elsewhere.

Well, I already suspected this by the lack of thecnical and specific
arguments of messages posted at the list. You are the editor and you can
take your own decisions regarding the spec. I cannot change that.

>
>> > > In other words math proposal is rejected, mathematics in HTML is
>> blocked one more time.
>> >
>> > I have suggested a process by which you could prove your proposal
>> would work. That is hardly a rejection.
>>
>> Some people can read between lines.
>
> There is nothing to read between my lines. I am being as honest and
> candid  as possible. There is no conspiracy here. I have given you the
> exact  reasoning I have used, I have suggested how you can move forward.
> I am  being quite sincere.

Ok, but I was able to read betweeen lines. For example, your initial
silence when the thread on math was launched indicated to me that maybe
you were waiting it automatically closed, but did not and obtained mayor
audience obligating to you to add messages. Well, maybe I was completely
wrong (I make mistakes each day) but your recent

"Because I don't think WHATWG is an appropriate venue for a task of such
huge scope, I would prefer if that happened elsewhere."

Has confirmed one of my suspicions. I will not be so pesimistic as George
was some days ago. I see a great future to CSS (or XSL-FO or SVG)
approaches to math rendering. Will it will take more time? Sure. Many
people initially go to MathML and after will abandon (my own case for
example), sure.

I am completely convinced that p-MathML has no more future than <font> and
<center> even if is being artificially spreaded and maintained on the web
thanks to evil $.

Will people in this list waiting mathematical markup on HTML5 use other
things? sure also, if they want solve problems and current tools are of no
utility for them, since at the end (at the very end), the world is really
simple: works or do not.


Juan R.

Center for CANONICAL |SCIENCE)