Re: XHTML 2.0 User Agent Conformance

David Woolley <david@djwhome.demon.co.uk> writes:

>>   Why shouldn't UAs have knowledge of the XHTML DTD?  If a DOCTYPE is 
>> specified, then isn't it a requirement of XML that the document 
>
> Firstly I believe that they should,

Of course, they should know the vocabulary of XHTML, but it's not a
general requirement on web browsers for an arbitary XML document type.
Knowing the vocabulary is different from either (1) validating or (2)
reading external document subsets, both of which are optional for user
agents.  There is a question, however, about how many XHTML namespaces
an XHTML user agent should know.  It is fortunate, for example, that
Mozilla knows the vocabulary of MathML.

>                                     but there is a lobby, particularly
> on the www-css list, that takes the view that browsers should purely
> apply CSS to XML.  My impression is that the person to whom I wrote the
> long response about the semantic web, etc., belongs to that lobby.

No.  The XML I write is mostly in the gellmu article document type,
which is reasonably semantic.  I've done nothing more than dally
poorly with CSS styling for it.  For presentation it goes through
careful translation to PDF (via LaTeX) and HTML 4.01.  Translation to
XHTML 1.1 + MathML 2.0 is under development.

                                    -- Bill

Received on Saturday, 1 November 2003 11:12:31 UTC