Re: RDFa, Embedded RDFa and GRDDL

On Tue, 2007-01-23 at 17:54 -0500, Harry Halpin wrote:
> I am having trouble with the part of our charter [1] that states we
> "tutorial materials and use cases sufficient to bootstrap adoption of 
> GRDDL", "as newer work such as RDFa  and Embedded RDF" due to the fact 
> that a RDFa->RDF/XML GRDDL transformation does not exist

Could you clarify? Did you forget our discussion of
Fabien'ss RDFa2RDFXML.xsl? The WG hasn't approved the
relevant test yet, but I'm not aware of any technical
reason why we should not:
  http://www.w3.org/2001/sw/grddl-wg/td/testlist1#rdfa1


>  and Embedded RDF 
> does not support XML Schema datatypes. I feel that in both these cases the 
> ball is clearly in the court of Embedded RDF and RDFa, not the GRDDL WG.

I pretty much agree; I think we've shown that GRDDL can be used for RDFa
and Embedded RDF.

> In the case of Embedded RDF, while it allows us to embed some fragment of 
> RDF, it does not currently work in the example in the primer due to its 
> lack of ability to support XML Schema Data Types  and I refuse 
> to ship broken examples to Last Call. Perhaps it could be used in the 
> second section of the primer, although it would complicate the example.
> If IanD does not get enough cycles free to respond to either adding 
> Embedded RDF data-types or changing the examples, we will have no choice 
> but to remove it from the primer and reference Embedded RDF in the 
> use-case document.

Hmm... why do we need datatypes in the primer? I guess I'll have
to take a closer look. But I agree that we need to get the primer
to be internally consistent before last call.

> Re RDFa, we have left  the spec open enough so that a GRDDL  result is 
> defined in terms of graphs, not RDF/XML, so a GRDDL result can be a RDFa document.

That's not using GRDDL for RDFa; that's using RDFa as an RDF syntax.

> However, we cannot demonstrate any test cases or put anything in the 
> primer as of yet concering RDFa except a fairly cursory notice that we 
> should in theory be able to work with RDFa unless:
> 
> 1) There is a GRDDL XSLT transformation that takes RDFa marked up-text and 
> produces RDF/XML (much like Embedded RDF currently does) that can be used 
> by our test-harness and primer,

Hm... it looks like you did forget Fabien's work and our discussion of
it (reminders above).

> or (more work):
> 
> 2) a RDFa aware parser that produces a graph that can be machine-readable, 
> and so used directly in our test harness.
> 
> and both options requires
> 
> 3) RDFa to have a stable syntax.

Why? For GRDDL WG purposes, I don't see any major problems if
RDFa syntax changes after our test suite is done. Our tests will
become less valuable as real-world examples, but not any less
valuable as GRDDL tests.

> I do not think option 1) is too much to ask from BenA or someone else in 
> the SWBD WG. However, it needs to happen *soon*, as if we do not get such 
> a transformation shortly we cannot reference it, as we would like, in the 
> use-cases or primer document. Otherwise we will not have RDFa appear in 
> the primer or test-cases and will only give it mention in the Use-Case 
> scenario.

-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/
D3C2 887B 0F92 6005 C541  0875 0F91 96DE 6E52 C29E

Received on Wednesday, 24 January 2007 00:40:10 UTC