Re: MathML-in-HTML5

On Wed, 4 Oct 2006, Roger B. Sidje wrote:
> On 4/10/2006 10:38 AM, Ian Hickson wrote:
> > 
> > The proposal that I understand Roger intends to experiment with is making
> > any tag in a particular list of tags be added to the DOM as a node not in
> > the XHTML namespace but in the MathML namespace.
> The simpler <math xmlns="...">...</math> would make life much easier, as
> - it would avoid listing all the possible tags up-front
> - gives parity with IE+MathPlayer out-of-the-box.
> - future proof.
> Perhaps xmlns could be an exceptional attribute.

There are millions of pages that have bogus xmlns="" attributes on various 
tags. If we started doing things with those attributes, those pages would 
break. That's not an option. I haven't studied IE closely enough to know 
exactly what it's doing, but it isn't as simple as simply handling an 
xmlns="" attribute on any tag.

More importantly, xmlns="" isn't simple. Authors have huge issues with 
anything like namespaces. There really isn't a good reason to introduce 
them -- future proofing isn't an issue here (we can just change the spec), 
parity with IE+MathPlayer isn't an issue (it's easy to have a script act 
as a shim if we need such compatibility), and listing all the possible 
tags up-front is easy.

Ian Hickson               U+1047E                )\._.,--....,'``.    fL       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

Received on Wednesday, 4 October 2006 01:28:50 UTC