> For the record: I disagree with that direction. Trying to integrate SVG
> and MathML into HTML seems to be inferior to having a generic solution.
There are two ways to get a generic solution: top down and bottom up. By
the former, I mean trying to specify a syntax for a generic solution, and
then seeing if it applies to a number of specific use cases. By the
former, I mean taking a number of use cases, addressing them, and then
seeing what (if anything) can be generalized.
SVG has mixed case attributes, an element named title, and makes use of
xlink. MathML has foreign object. Both continue to evolve, and any HTML5
solution should be prepared for such evolution. Anything that completely
(or even mostly) satisfies the above constraints likely can be generalized.
If focusing first on such a set of requirements in a bottoms-up fashion
enables progress to be made faster, then I'm all for it.
Microsoft's implementation is also reasonably general (albeit with a number
of restrictions). Since compatibility with IE to the extent it makes sense
is a goal, this may also help.
- Sam Ruby