Re: my action on conformance

Steven Pemberton wrote:
> 
> On Thu, 20 Sep 2007 01:13:38 +0200, Ben Adida <ben@adida.net> wrote:
[...]
>> Note that this does *not* mean that RDFa will generate triples for the
>> old Dublin Core notation, just that if a tool like Mark's Sidewinder
>> chooses to generate triples for the legacy Dublin Core approach, we
>> won't say that it no longer complies with RDFa.

In the GRDDL spec, the central concept is a "GRDDL result". It's
an RDF graph. If Sidewinder spits out more triples than the spec
calls for, it just can't call them a GRDDL result, and it
can't justify them from the GRDDL spec. Another way to look
at it is: an agent can produce more triples, but only if the user
asked for something more than GRDDL results.

David Booth made a proposal to define something like "complete GRDDL
results" which was the merge of all the GRDDL results, and no more,
but it was pretty late in our process and the WG didn't agree that
it was part of the minimum required to declare victory.

We did define a class of software product called a GRDDL-aware Agent.
http://www.w3.org/TR/grddl/#sec_agt

If you wanted to cite GRDDL normatively, you could just re-use
the conformance definitions there to define RDFa result:

   If <the-official-rdfa-transformation> is a GRDDL transformation
   of ?DOC and that transformation applied to ?DOC gives ?GRAPH,
   then ?GRAPH is an RDFa result of ?DOC.

and for agent:

   An RDFa agent is a software module that computes RDFa results.


I suppose you could make analagous definitions without citing
GRDDL normatively, by copy/paste/adapt.

> Still, I don't think RDFa should necessarily be the sole source of 
> triples for a document. Think microformats and RDFa in the same document.
> 
> But I think our test set should attempt to spot dirty triples.
> 
> Steven


-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/

Received on Thursday, 20 September 2007 15:32:40 UTC