Re: MathML-in-HTML5

>>   <OBJECT ID=...
>>    CLASSID="ExtremelyLongIndecipherableRandomString"></OBJECT>
>>   <?IMPORT NAMESPACE="..." IMPLEMENTATION="#..."?>

> This construction is baroque, grotesque, and insecure.  It is simply
> unwise to allow an arbitary content provider to have this level of
> control over a user agent.

> Moreover, it presents challenges for public sector content providers
> who by law cannot allow the impression of being unduly influenced by
> particular market participants.

If you serve XHTML+MathML content as application/xhtml+xml you don't need to include junk quoted above in each document (it has to be added explicitly only if you serve page as text/html). 
And if you use content type that is not supposed to be used for XHTML + MathML documents, then you should not complain about UA specific code, as institutionalizing spec violations is a UA specific practice "unduly influenced by particular market participants" :P

> I hope we can persuade the purveyors of user agents offering such
> constructions to join with others in providing a better approach that
> is both secure and inter-operably meaningful.

Today application/xhtml+xml is the most interoperable solution that all MathML aware UAs recognize.



-- 
_______________________________________________
Surf the Web in a faster, safer and easier way:
Download Opera 9 at http://www.opera.com

Powered by Outblaze

Received on Thursday, 12 October 2006 07:58:09 UTC