[foaf-protocols] WebID spec: supported RDF formats
Bruno.Harbulot at manchester.ac.uk
Fri Jul 16 18:04:01 CEST 2010
(By the way, please see the latest draft across repositories. The
current repositories are:
Right now, the latest one is in retobg's master branch.)
This one is about the SHOULD/MUST requirements for the RDF formats.
The initial draft was saying this:
> The XHTML+RDFa 1.1 [XHTML-RDFA] serialization format must be
> supported by the mechanism, e.g. a Web Service, providing the WebID
> Profile document. Alternate RDF serialization formats, such as N3
> [N3], Turtle [TURTLE], or RDF/XML [RDF-SYNTAX-GRAMMAR] may be
> supported by the mechanism providing the WebID Profile document.
On which Toby said in
> I'd also like for RDF/XML to be another format that implementations
> are *required* to support.
Then Reto said in
> - I'm against requiring RDFA as this means I cannot expose my triple
> store as a collection of Web-Ids without adding some templates. If
> we really think a format mus be guaranteed the I think it should be
And Kingsley said "No!" to that.
After this, I suggested this wording:
> A Verification Agent must be able to process documents in RDF/XML
> [RDF-SYNTAX-GRAMMAR] and XHTML+RDFa [XHTML-RDFA]. A server responding
> to a WebID Profile request must be able to return a representation in
> RDF/XML (using media type application/rdf+xml) or XHTML+RDFa (using
> either media type text/html or media type application/xhtml+xml). In
> addition, either parties may support any other RDF format via HTTP
> content-type negotiation.
... which got a mixed reception.
The rationale behind this was that this is a way to guarantee that any
WebID will be processable by the Verification Agents, and that it's
easier to mandate things from the Verification Agent than from the
publisher (especially, regarding how the HTML+RDFa is produced).
Then, as far as I understand there were different angles:
- Some of us think a WebID must be able to return an HTML/XHTML
- Some of us think that format specifications are not important and that
things will eventually work, perhaps using GRDDL in the worst case.
- Some of us (me) think that we must ensure that an implementation of a
WebID host that only follows the MUSTs and an implementation of a
Verification Agent that only follows the MUSTs must be able to talk to
each other, and that this requires some MUSTs w.r.t. to formats.
I'm still supporting the wording I suggested: a few formats are
mandatory on the Verification Agent side and the WebID provider MUST
support at least one of those mandatory formats.
I'd be OK to relax the second requirement to a SHOULD, with an
accompanying sentence saying more or less "if you don't publish in one
of the formats that the V.A. must support, it's your fault when it
It seems that some of us are not in favour of a mandatory base-line that
guarantees that things will work. I think this is a strong requirement
for the specification (it's a protocol spec, not a best-practice spec).
I'm also against mandating HTML, a WebID profile should be able to use
only RDF/XML if it wants too.
More information about the foaf-protocols