Re: Chat with Ivan on RDFa 1.1 Authoring features

On Wed, 2010-03-24 at 16:47 -0400, Ivan Herman wrote:
> What I said, to be more precise: if our goal is (and I think it is) to
> simplify RDFa and bring it closer to HTML(5) authors, then we should
> keep away from RDFS. We would shoot ourselves in the foot in the sense
> that it would make RDFa more complex conceptually, something we do not
> want to do... (sorry Toby:-) 

As I've said, I'm not suggesting that RDFa includes RDFS/OWL. (I had
previously suggested this though.)

My suggestion is that RDFa does not need to contain a profile feature -
all it needs to do is provide a way of setting a default prefix.

This would enable Google, say, to recommend publishing data like this:

 <p vocab="http://rdf.data-vocabulary.org/#" typeof="Person">
   <a property="name" rel="url" href="/">Toby Inkster</a>
 </p>

i.e. as they currently do, but without prefixes.

The default prefix (indeed, non-default prefixes too) *may* be
dereferenced to find an RDF schema, which might result in the consumer
being able to infer additional triples. But that's a consideration
orthogonal to RDFa 1.1. (After all, RDFS and OWL reasoning like that can
already be conducted on existing RDFa 1.0 documents if the consumer
chooses to do so.)

No, this suggestion is not as powerful as RDFa profiles, but the use of
profiles introduces fragility into RDFa documents, and I'm not convinced
that the use cases suggested so far justify taking the risk with them.

Whatsmore, I think that my suggestion leaves us room for adding profiles
in a future version of RDFa without changing the syntax. (You'd just
tell RDFa parsers to start performing profile processing on the URIs
given in @vocab.) Jumping straight to profiles doesn't allow us a
natural path back from them if they turn out to be a mistake.

-- 
Toby A Inkster
<mailto:mail@tobyinkster.co.uk>
<http://tobyinkster.co.uk>

Received on Wednesday, 24 March 2010 23:15:05 UTC