Re: MathML-in-HTML5

On 12/10/2006 2:25 AM, Ian Hickson wrote:
> 
> Given how little the namespace-prefix syntax has taken off, and the number 
> of people who are confused by namespace syntax of any kind, and the large 
> amount of content on the Web that uses namespace-like syntax in a way that 
> would break renderings if it was interpreted according to namespace-like 
> parsing rules, however, I would strongly recommend against any sort of 
> namespace-based parsing extensions to HTML. As I mentioned before, I've 
> worked with an implementor in the past who tried to do this, and they 
> ended up wasting months of intense work trying to be compatible with the 
> large amounts of legacy content before eventually backing out the whole 
> thing and ignoring namespaces in text/html content.

I can technically restrict the implementation to only support the plain 
<math>...</math>, while authors will effectively use <math 
xmlns="...">...</math> for IE+MathPlayer. (That is, Mozilla wouldn't 
actually require the xmlns attribute for the MathML rendering to kick 
in. Authors only put the attribute there for interoperability.) The 
issues now become:
- can  IE+MathPlayer be made to at least support <math 
xmlns="...">...</math>? (from David's recent post, it seems possible in 
IE7.x?)
- what is the implication for HTML5? Just one of those random attributes 
that people put in their HTML document? (which wll be fine with me, BTW).
- what about generators that emit prefixed tags?
---
RBS

Received on Wednesday, 11 October 2006 16:44:56 UTC