Re: HTML 4 Profile for RDFa

Henri Sivonen wrote:
> On May 23, 2009, at 19:49, Philip Taylor wrote:
>
>> * Use some other prefix-binding mechanism (in both XHTML in HTML) 
>> like prefix="t=... T=..." instead of xmlns:t="..." (breaking current 
>> implementations and deployed content, but avoiding the mess of 
>> parsing differences between XHTML and HTML).
>>
>> I can't think of any other solutions, so something is going to break 
>> no matter what is chosen.
>
>
>  * Use a mechanism other than CURIEs to turn attribute values into 
> absolute IRIs.
>
> As it happens, the microdata to RDF conversion already has a mechanism 
> for mapping itemprop values to absolute IRIs without prefix-based 
> indirection. (And itemprop works nicely with Selectors, too, allowing 
> styling based on microdata, while Selectors don't support matching on 
> expanded CURIEs, so styling on RDFa would need to depend on the 
> prefixes that aren't supposed to be significant.) In fact, even RDFa 
> itself has a non-CURIE-based mapping from attribute values onto IRIs 
> for traditional rel tokens.
>
I wouldn't make an assumption that the microdata section will exist 
going into the future. Regardless, discussions of Microdata are 
irrelevant to discussions of RDFa, don't you think?

> http://rdfa.info/wiki/Rdfa-in-html-issues currently has 12 issues.

And how many of these are related to bugs in the processors, and how 
many are actually issues with RDFa? Do we file Firefox bugs with the 
HTML WG, now? Safari bugs? Opera bugs?

IE bugs? Oh, that could be fun.

Again, it's a wiki, people should do what they will. But I hope that 
people use the wiki to record issues with RDFa and HTML, not the 
individual processors.

>
> Of these 8 pertain to mapping attribute values to URIs. Of those, 7 
> would be addressed by adopting the same mapping to IRI that itemprop 
> uses. (The remaining 1 issue--Mapping algorithm producing an invalid 
> IRI--is not an issue to the extent RDF processing doesn't really 
> process properties as IRIs but processes them as opaque strings.)
What you're saying is that it's up to RDFa to break, because DOM 
implementations fail in specific, limited circumstances?

>
> 2 issues would be addressed if RDFa deferred to 
> http://www.whatwg.org/specs/web-apps/current-work/#the-lang-and-xml:lang-attributes 
>
>
> 2 issues would be addressed if RDFa deferred to 
> http://www.whatwg.org/specs/web-apps/current-work/#coercing-an-html-dom-into-an-infoset 
>
>
And how many of these are related to underlying problems with the DOM, 
not directly with RDFa?

Shelley

Received on Sunday, 24 May 2009 13:00:30 UTC