Re: Telecon draft minutes: 2002-05-03

On Mon, 6 May 2002, Patrick Stickler wrote:

> > Item 11: Issue rdfs-isDefinedBy-semantics
> >
> > http://www.w3.org/2000/03/rdf-tracking/#rdfs-isDefinedBy-semantics
> > http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2002Apr/0478.html
> >
> > Aaron's proposed resolution:
> >
> > - isDefinedBy is a Property
> > - it has no further semantics in the RDF Model Theory
> >
> > - further subclasses of it may be created by others
> > - it is out of scope for RDF Core to create these subclasses
> >
> > No dissent.
> > ACCEPTED: tihs proposal was accepted.

If I had been there, I would have dissented. I wasn't, and didn't, but for
the record I have two problems with the detail (if not the intent) of this
resolution.

(i) we don't specialise properties with sub-class relations; the correct
term is sub-property. these are different relationships in rdf.

(ii) it is not out of scope for RDF Core to create these sub-properties,
we have merely decided to focus our energies elsewhere. The phrase 'out of
scope' shouldn't be over-used for everything we decide not to do. The
charter asks us to complete work on RDFS; this could quite plausibly
include refinements of rdfs:isDefinedBy based on a better understanding of
(for eg) the relationship between RDF Schema and XML Schema, or of the
needs of WebOnt (eg. mechanisms such as daml:imports).

I don't want this re-opened, but wanted these points on record.


A couple more casual comments below. Nothing major.

Dan

> > The question posed was: should an RDF parser reject the following as
> > illegal?
> >
> > <rdf:Bag rdf:about="http://example.org/foo">
> > <rdf:_1 rdf:resource="http://example.org/a" />
> > <rdf:_1 rdf:resource="http://example.org/b" />
> > </rdf:Bag>
> >
> > It was decided (no dissent) -
> > RESOLUTION: the WG resolves that this is syntactically legal RDF

Phew :)

One reading: If URIrefs are names, http://example.org/a and
http://example.org/b could be names for the same thing.


> > AOB#2: isDefinedBy - futhter discussion
> >
> > The difference between seeAlso and isDefinedBy was discussed.
> > FrankM: we're not looking at reopening this, we're soliciting advice as
> > to what to say in the primer.

One option I discussed face to face with brian: renaming rdfs:isDefinedBy
to be called rdfs:ns or rdfs:namespace. We wondered whether doing so might
make its purpose more transparent. Not sure.

Received on Tuesday, 7 May 2002 15:44:07 UTC