GRDDL and SWD

Dear participants of the GRDDL working group,

Harry Halpin asked [1] the Semantic Web Deployment WG to review the
GRDDL spec and try to  flag any potential issues w.r.t. SWD
deliverables. On behalf of the WG, I wrote a brief review which was
discussed in a telecon yesterday [2]. Please read it below. Also please
note that I just reviewed the spec (not the Primer nor the test cases).

Short version:

I don't see any conflict between GRDDL spec and SWD
Deliverables, although there is a minor remark about RDFa (see below).
We encourage you to go ahead and publish the documents.

Long version:

* GRDDL and SKOS: I cannot see any obvious connection between these
technologies (with the exception of SKOS being one of the candidate
formats to capture the output of GRDDL transformations). The GRDDL spec
makes no reference to SKOS, and I think that's just fine.

* GRDDL and Best Practice Recipes for Publishing RDF Vocabularies: the
first paragraph of section 3 of the GRDDL spec
suggests [3] the use of content negotiation to publish different
representations (XML Schema / RDF Schema) of a namespace document. It
also links the definition of "content negotiation" in the WebArch
REC [4], however it makes no reference to the Recipes. I see a
connection here, but I don't think it's strong enough to justify a
reference to the Recipes, considering that content negotiation is just a
side-question that appears in the context of an example. Therefore, I
suggest no change here.

* GRDDL and Vocabulary Management: there is also a connection between
the concern with avoiding the overload of webservers serving namespace
documents (which might be expressed as vocabularies in RDF Schema) and
the versioning of such documents, because the proposed solution is to
implement some kind of cache, and "GRDDL-aware agents should ensure that
this local memory is up to date [...]". However, an example later in the
document (section 7) illustrates the usage of HTTP headers to ensure
this, which is a simpler, widely-accepted and already available
solution. Therefore, I think it makes no sense to apply RDF versioning
to solve this problem.

* GRDDL and Semantic Integration: They're not related.

* GRDDL and RDFa: I kindly suggest to add a paragraph to emphatise the
complementarity of GRDDL and RDFa (like the one that already exists for
XProc). In this sense, I'm sure you are aware of this wiki page [5] by
Ben Adida. It would be nice if you could provide your readers with some
hints on when to use GRDDL and when to use RDFa.

Best regards,

[1] http://lists.w3.org/Archives/Public/public-swd-wg/2007Apr/0020.html
[2] http://www.w3.org/2007/04/24-swd-minutes.html
[3] http://www.w3.org/TR/grddl/#ns-bind
[4] http://www.w3.org/TR/webarch/#def-coneg
[5] http://www.w3.org/2006/07/SWD/wiki/RDFaGRDDL

-- 
Diego Berrueta
R&D Department  -  CTIC Foundation
E-mail: diego.berrueta@fundacionctic.org
Phone: +34 984 29 12 12
Parque Científico Tecnológico Gijón-Asturias-Spain
www.fundacionctic.org

Received on Wednesday, 25 April 2007 12:55:25 UTC