Re: Proposal for rdfms-not-id-and-resource-attr (that actual proposal)

This brings me on to an ambiguity in the grammar that needs sorting out.

On Mon, 11 Jun 2001, Aaron Swartz wrote:

> Now that we've wrapped up rdfms-empty-property-elements I suggest that it's
> time to finish rdfms-not-id-and-resource-attr. I hope I'm not stepping
> outside of my bounds here (as DaveB is really the issue owner) but I'd like
> to put forth a proposal.
>
> I'd suggest that the example in the issues list:
>
>     <rdf:Description>
>       <foo:bar rdf:ID="foobar" rdf:resource="http://foobar"/>
>     </rdf:Description>
>
> be interpreted as:
>
>     _:genid foo:bar <http://foobar> .
>     <#foobar> rdf:subject _:genid .
>     <#foobar> rdf:predicate foo:bar .
>     <#foobar> rdf:object <http://foobar> .

Skipping for a moment the use/mention arguments about reification (and
the base URI problem*), I got very close to this test case (although
looking back, I didn't include it) in my first message about
#rdfms-empty-property-elements:

http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2001Jun/0000.html

Which people seemed to be in agreement with. Modulo arguments about
expressing reification, I'm in agreement here (although you left out an
rdf:type arc).

HOWEVER. The reason behind my second message on the subject:
http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2001Jun/0059.html

was this: paragraphs 229-243 of M+S
http://ioctl.org/rdf/ms/rdfms#229

give another interpretation for rdf:id (in particular P232) if it
appears. In particular, there are two conflicting interpretations for

<blah:propname rdf:id="foo"></blah:propname>

(which people seemed happy to indicate reification - see my test case
here:)
http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2001Jun/att-0000/09-test5.rdf

and

<blah:propname rdf:id="foo" />

which currently matches a different clause in the grammar (the last
alternative of production [6.12] with zero occurrances of propAttr) and
which has a different interpretation (accoring to para 232).

Since these two cases are indistinguishable as XML, this is a problem.
Thus, (again, modulo the "how do we do reification" argument), M+S needs
a revision of paragraph 229-234 to bring them into accordance with
Aaron's, Brian's, (my) etc. understanding of rdf:id.

> DaveB, in his original proposal[1], decided that ID and resource could not
> be used because in an empty XML element, ID pointed to the object of the
> element. He later said[2] that the issue could be looked at again if
> empty-property-elements were decided differently. As they are now decided to
> eliminate the portion of the spec that cause the confusion, I suggest that
> the interpretation of ID be changed to fit with the test case above.

Otherwise, +1.

jan

* I don't want to make it seem that I'm making a big issue of this. I'd
just like to know how to couch my test case output; that's all.

-- 
jan grant, ILRT, University of Bristol. http://www.ilrt.bris.ac.uk/
Tel +44(0)117 9287163 Fax +44 (0)117 9287112 RFC822 jan.grant@bris.ac.uk
Generalisation is never appropriate.

Received on Tuesday, 12 June 2001 06:52:26 UTC