[whatwg] Spec should give guidance on compound document integration points

[Henri Sivonen]

On Feb 25, 2007, at 19:52, Robert Sayre wrote:

> On 2/24/07, Henri Sivonen <hsivonen at iki.fi> wrote:
>>
>> The obvious candidates for compound document mixing are SVG and
>> MathML.
>
> I agree that these are relevant, as is XBL.

But XBL isn't supposed to be embedded in an XHTML5 host document, is it?

>> Also, there are indications that people will want to embed
>> RDF metadata in documents even though the syntax of RDF is designed
>> for external metadata and isn't really all that good for embedded
>> metadata. I think WA 1.0 should give some guidance on these matters.
>
> RDF is irrelevant. I don't dispute that there are use cases for this,
> but at some point you have to deliver on them. Years and years and
> years later, it is clear that RDF has failed to do so.

I don't disagree.

But if I don't make a hole for RDF in the conformance checker, people  
who insist on embedding RDF in XHTML5 will do something very bogus  
like putting RDF stuff in comments. (Check out Trackback  
autodiscovery for fashionable regexp-scraped RDF in comments.)

> It is not worth spending WG time on, IMHO.

 From my selfish point of view, it is nice to have spec text to point  
to in cases like this, so I can reply to support emails by citing the  
spec. From outside my selfish point of view, I argue that this is  
like an XML 1.1 serialization for HTML5. It is not a good idea, but  
some people will try it anyway, so the spec might as well give  
guidance so that people do it right.

I hope the following paragraph is copy-paste-ready enough not to  
waste too much time:

In XHTML5 (but not in HTML5), the RDF element from the namespace  
http://www.w3.org/1999/02/22-rdf-syntax-ns# may appear where metadata  
elements are allowed. The structure and the meaning of the subtree  
rooted at the RDF element from the namespace http://www.w3.org/ 
1999/02/22-rdf-syntax-ns# is outside the scope of this specification.  
This specification does not require XHTML5 user agents to support  
such subtrees in any particular way other than leaving them in the  
DOM. However, when such a subtree is used for communicating metadata  
about the host document to an RDF-aware agent, the empty string  
should be used as the value of the about attribute from the namespace  
http://www.w3.org/1999/02/22-rdf-syntax-ns# to refer to the host  
document.

-- 
Henri Sivonen
hsivonen at iki.fi
http://hsivonen.iki.fi/