[whatwg] Mathematics in HTML5

[juanrgonzaleza at canonicalscience.com]

James Graham wrote:
>
> That's a really particular use case which is hardly representative of
> the web as a whole. As sad as it is, 99.9% of authors have no use for
> maths (otherwise all these problems would have been solved long ago).
> Maths is certainly less of a core feature for most authors than vector
> graphics and WHATWG aren't trying to re-implement SVG despite the fact
> that it too has no obvious IE6 compatibility story, poor CSS integration
>  and various other problems.

I doubt of veracity of your percentage. Any reference?

And your arguments sound contradictory. Have you noted that w3c attempted
to develop mathematical markup already in the HTML3 draft, that was a lot
of time before SVG or any other vectorial graphics approach.

And do not forget the duality

SVG <===> Canvas

MathML <===> HTML5-Math

> Nowhere in the WHATWG document does it say that they're going to try and
>
>   fix everything.

Strictly speaking it does not say the contrary, and does not say one would
reuse MathML.

> You have to choose your battles and, personally, I
> agree with the idea that, if the proponents of CSS-based maths want to
> work in the structure of the WHATWG, they should demonstrate the
> feasibility of their approach using a microformat.
>
> Given the constraints
> under which they have chosen to operate it should be possible to do
> this  without any  difficulties. The microformat based approach has
> several  advantages too, e.g. instant implementation in existing HTML4
> UAs (a new  markup language would require changes to the parser). This
> should allow  the language to evolve as it encounters real-world needs
> so, if and when  it is formally standardized, it will be a better
> product than typically  results from an
> standardization-before-implementation approach.

Many times it was said that there is not need for microformats. If
microformats are so good why the need for HTML5?

[http://www.xml.com/lpt/a/2006/04/26/microformats-grddl-rdfa-nvdl.html]

[http://www.xml.com/lpt/a/2005/10/19/microformats-and-web-2.0.html]

Why a calendar control, address card, and all that stuff in HTML5? Why do
not just reusing available microformats on standard HTML4 or XHTML1
without new added elements?

You argue that language can contain initial errors so one would be forced
to evolutionate the language when was applied to real problems. Of course!
But this is true in any case! Precisely, XHTML2 introduces certain
backward incompatibility because errors done in the past, somewhat as XML
1.1 introduces also. MathML is backward incompatible with original HTML 3
Math draft. If the points about errors matters would not the full HTML5 be
implemented in an experimental fashion until was completely fine-tunned
after of years of usage.

What about <canvas>, why was so early implemented in browsers without
years of experimental usage on the web? I can see some people considering
lack of strong support for textual content in canvas one of the big flaws
of the spec.

About your comment on changes in parsers, well, I thought that anyone in
this list was warned that HTML5 will need of changes in parsers!! Any case
the changes introduced by George proposal are minimal when compared with
changes of a "text/html" version of p-MathML or of Ian’s proposal.


Juan R.

Center for CANONICAL |SCIENCE)