W3C home > Mailing lists > Public > public-rdfa-wg@w3.org > August 2010

Re: XHTML+RDFa DTD UA support (Was: Getting the XHTML+RDFa default profile to be interesting)

From: Leif Halvard Silli <xn--mlform-iua@xn--mlform-iua.no>
Date: Mon, 16 Aug 2010 18:49:06 +0200
To: Shane McCarron <shane@aptest.com>
Cc: RDFa WG <public-rdfa-wg@w3.org>
Message-ID: <20100816184906888167.bcd9d362@xn--mlform-iua.no>
Shane McCarron, Mon, 16 Aug 2010 11:20:59 -0500:
> Unfortunately, it is hard to be very granular about this.  Most user 
> agents, including Opera, will treat content as raw XML regardless of 
> the DOCTYPE if the media type is set to application/xhtml+xml 
> (according to Simon from Opera, anyway).

It sounds as if Simon eventually were generalizing too much. 

Whether a user agent supports named character entities or not, seems to 
me to be a good indicator for whether the user agent has implemented 
the DTD - or not.

My tests indicated [I did not test each and every named entity] that - 
simply put - Gecko, Opera and Webkit had implemented all the XHTML DTDs 
which are covered by HTML5. Thus they e.g. support named entities if 
you use the XHTML+MathML+SVG doctype. They had also implemented the 
XHTML 1.1 DTD.  But they had not implemented the XTHML+RDFa DTD.

Thus, despite that they are non-validating XML parsers, they are 
implementing - for the DTDs that they have chosen to support - the MAY 
option in XML 1.0: named entities of these DTDs are supported. 

(Btw: I suppose "raw XML" includes implementation of the XHTML 
namespaces - and all which implied by that.)

>  XHTML Modularization is 
> very clear about the naming scheme for XHTML family document types, 
> and at the time we defined that the browser makers were in the 
> working group and were in favor of using that as a trigger to put 
> user agents into XHTML Family mode.

Ah, I begin to understand the purpose of XHTML Modularization. I agree 
that if a DTD is known to be based on XHTML modularization, then it 
should be simple to support variants ... shouldn't be necessary to hard 
code support for new variants. Any place to read about XHTML Family 
mode? Is evidence that such modularity was ever supported?

>  Obviously this has been overcome 
> by events.  As a result, we feel it is in the best interests of the 
> XHTML-using community that they are aware there is a portability risk 
> when using application/xhtml+xml and XHTML family document types, and 
> that they are aware of the ways to minimize those risks.  That's the 
> purpose of the XHTML Media Types appendix.

Does it/Will it speak about standalone="yes"? In Toby's case, here 
weren't aware of the problem - because he used Opera and because Opera 
silently just don't parse such entities. By turning on standalone="no", 
one can at least easily verify that one do not use unsupported entities.

> Hope this helps.

Yes. Informative. The "events" has been hard to grasp. But now I begin 
to get a grip about them.

Received on Monday, 16 August 2010 16:49:39 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:19:48 UTC