[uf-discuss] How about adding aRecipe, an RDFa serialization of hRecipe?

[Tom Morris]

On Thu, Oct 9, 2008 at 11:31 AM, Thomas Loertsch <loertsch.thomas at guj.de> wrote:
> That's where things come together for me: technically I need the
> extensibility of RDFa, vocabulary-wise I need the community process
> established by microformats.org. So why not develop a vocabulary and two
> serializations in parallel? An RDFa-serialization isn't much work to do, the
> hard part is the vocabulary and that's the same for both serializations,
> RDFa and microformat. Basically it would just mean that we give the
> RDFa-serialization a nice home on the web, side by side with the
> microformat-serialization, improving interoperabilty and reuse.
>

I'm not sure that an RDFa serialization is something that needs
endorsement or hosting by microformats.org. Once hRecipe is
formalised, RDF/RDFa-based work that uses the hRecipe vocabulary needs
no input from microformats.org. The syntax for the RDFa is derived
simply from the RDF/OWL model - there's nothing there that needs
deciding on. And, well, if you have a microformat, it seems pointless
to publish RDFa - just create a GRDDL profile describing the
transformation of hRecipe into RDF/XML and use that (HTML 5 @profile
concerns aside).

You do raise an important point though: the RDF world needs a
microformats.org-style development community. That was the intention I
had behind setting up GetSemantic, and I think the creation of such a
community is the intention behind VoCamp. At VoCampOxford, we
discussed the problem of the lack of a microformats-style venue for
discussion and creation of RDF vocabularies, and this is an active
problem that I think needs solving. We can't just tack this on to
microformats - it needs to be separate. Such a group would, of course,
tell people to use microformats where appropriate.

Here's how you (or whoever else is interested) should probably
proceed: when hRecipe reaches a point of maturity, sit down and
properly RDFize it - work out the classes and the properties involved,
and figure out how they map both ways from the RDF to the microformat.
Draft that up as an OWL ontology, put a draft version up in Notation3
format on the web and solicit comments. You don't have to specify an
RDFa mapping, because that's implicit - if you understand RDFa, and
you understand the vocabulary, the RDFa syntax that you would use
becomes obvious.

Read http://vocamp.org/wiki/Best_Practices
And if you've got questions, ask on irc.freenode.net #swig

Sorry to regular mf-discuss readers for that extended diversion into
RDF-land. Back to your regular scheduled programming.

-- 
Tom Morris
http://tommorris.org/