Re: Exploring new vocabularies for HTML

On Mon, 31 Mar 2008 10:17:59 -0700, Bruce Miller <bruce.miller@nist.gov>  
wrote:
> So, Classic MathML, provided it didn't use namespace prefixes,
> I assume, would be valid to embed in HTML5?

I would expect the subset to have more restrictions than just not using  
namespace prefixes. Perhaps even limited to presentational MathML as has  
been suggested.


> That would be good, as it would solve the problem
> of synthesizing HTML5 + externally generated MathML
> outside of the browser context.
> [And presumably similarly for SVG]

Yeah, I would expect the same for SVG.


We've already done something similar for embedding XHTML in a HTML context  
to not require everyone to learn new tricks. You can write <br> as <br/>,  
write <html> as <html xmlns="http://www.w3.org/1999/xhtml">, etc. There is  
no "semantic" difference, and there are a ton of restrictions (<script  
src="x"/> would not work in HTML for instance, you have to use <script  
src="x"></script>), but a lot of authors appreciate that they can write  
this kind of syntax.


> It is doubly-good if browsers were required to
> export the MathML as XML, since then that could
> be dropped directly into HTML5 w/o being reserialized
> in its HTML5 form.

Such UI requirements are really out of scope I think. At best the  
specification could encourage it.


> Are you referring to the exporting of MathML as XML
> a not-necessarily-mandated "UI feature" ?
> That seems really bad to me.
> It forces all tools to support 2 "standards" instead of one.
> Some tools may eventually support both, but for a long time
> we'll have taken a large backwards step.

You'll always have a transition period. In that period you'll likely get  
tools on the Web that convert from one to the other and MathML products  
might start releasing beta products with experimental HTML syntax support.  
And browser plugins or maybe even browsers will support the "copy as XML"  
contextmenu option.


> Besides, if, as you say, MathML as XML would be allowed
> in HTML5, there'd be no need for a browser to
> export the HTML5 serialization.

That would depend on whether the end user wants to edit it further by hand  
and put it in some other HTML document or edit it in a MathML application.  
(View source will always show the unaltered markup, fwiw.)


-- 
Anne van Kesteren
<http://annevankesteren.nl/>
<http://www.opera.com/>

Received on Monday, 31 March 2008 17:36:46 UTC