October 2, 2001

Attendees:
Mike
Sandro
Peter
Pat
Harold
Jeff
Dan C.
Lynn

Announcements:
Sub property loop has been decided

Agenda:
minutes from September 25 - approved

any updates on submission (peter)
    
    there is a new version of the axiomatization that is better
    formated...not inputed to W3C yet (only formating corrections)

RDF(S) Model Theory (Peter, Pat)

       Pat: responding to updated model theory.  Why do you need
       untidy RDF graph
       
       Peter: don't really ... Pat had them so Peter had em.

       Peter: things are progressing rapidly

       Pat: Finally put out a workin draft (rdf wg) model theory for
       rdf / rdfs.  Peter found bugs in it (specifically for
       rdfs)... it is currently being revised and will be put out
       shortly which will be fixed (though no substantial changes).
       Peter has redone the DAML model theory to conform... A lot of
       discussion about literals, whether they should be resources or
       not...want to make some more changes to make the model theory
       more agnostic in this sense.

       Pat: want to keep it as agnostic on all the issues that are
       still being discussed in the working groups.  It should say
       whatever you decide I can run with and not here's my opinion
       and now use it.

       Peter: the trick in Pat's model theory made me think that we
       can use the same trick to extend the semantics for DAML+OIL to
       everything (RDFS as well?).  Everything works out and there
       seem to be some interesting benefits.  You can now create a
       list of specific things (a class list or a thing list).

       Peter: only remaining problem ... how do you treat literals.

       Pat: my model theory may work fine with DAML+OIL once I'm done
       with it

       Peter: real problem is about the meaning of the literal

       Peter: this will look much closer to the axiomatic semantics
       than last time.

       Peter: would be interesting to see how close the axiomatic and
       model theory semantics really are once we're done.

       Can we do a survey of how people are using things like data types.

       Mike:  probably

stripping rdfs:label/comment from DAML+OIL (Peter)

       Peter: my model theory doesn't include rdfs label and comment.
       Why do we need it.  But I think this is a decision that we can
       turn over to a working group to decide.