Re: raised in comment: owl:class still needed? Does this effect Test LC?

From: "Jos De_Roo" <jos.deroo@agfa.com>
Subject: Re: raised in comment: owl:class still needed? Does this effect Test LC?
Date: Sat, 24 May 2003 23:09:49 +0200

> 
> Last Thursday Jim Hendler wrote:
> >At 8:23 AM -0400 5/22/03, Peter F. Patel-Schneider wrote:
> >>From: Jim Hendler <hendler@cs.umd.edu>
> >>Subject: raised in comment: owl:class still needed? Does this effect Test
> LC?
> >>Date: Wed, 14 May 2003 11:19:38 -0400
> >>
> >>>
> >>>  One part of one of our public comments from RDF Core asks:
> >>>
> >>>  RDFCore: Comments on OWL Reference
> >>>
> http://lists.w3.org/Archives/Public/public-webont-comments/2003May/0004.html
> >>>
> >>>     #owlref-rdfcore-owl-class-denotation
> >>>     It has been suggested to
> >>>     RDFCore that owl:Class is not needed.  RDFCore requests the
> >>>     creation of test cases to clearly illustrate the differences
> >>>     between owl:Class and rdfs:Class.
> >>>
> >>>  I'd like to see such a test (or tests) in our LC Test document, as it
> >>>  is likely that we will get this same or similar comment again.  If
> >>>  such a test cannot be generated, then I believe we need to reopen
> >>>  issue 5.20 as it was determined at the Bristol f2f:
> >>>
> >>>    re 5.20 Should OWL provide synonyms for RDF and RDFS objects? no,
> >>>  owl should not have synonyms; owl:Class is not a synonym.
> >>>
> >>>  (this is part of a long thread and the resolution included this and
> >>>  other statements, but I believe the above is where the WG officially
> >>>  agreed owl:class was not a synonym)
> >>>
> >>>  and appropriately change our documents.
> >>>
> >>>  Peter/Ian (or anyone else) - can one of you remind the WG the
> >>>  difference and design a test case for it?
> >>
> >>[copied out of another message]
> >>
> >>The RDFS-compatible semantics for OWL DL heavily depends on owl:Class
> (more
> >>precisely, on IOC, the class extension of owl:Class).  If the distinction
> >>between owl:Class and rdfs:Class was removed the semantics would be quite
> >>different.
> >>
> >>Test cases are rather hard to come by, as OWL DL is designed so as to
> >>prevent one from interacting with classes that are not OWL classes.
> >>
> >>However, if one looks at RDF graphs that are not in OWL DL one can see
> the
> >>difference.  For example,
> >>
> >>ex:a rdf:type rdfs:Class .
> >>ex:ia rdf:type ex:a .
> >>
> >>currently does not OWL DL entail
> >>
> >>ex:ia rdf:type _:i .
> >>_:i owl:intersectionOf _:l1 .
> >>_:l1 rdf:type rdf:List .
> >>_:l1 rdf:first ex:a .
> >>_:l1 rdf:rest rdf:nil .
> >>
> >>but it would if owl:Class was replaced with rdfs:Class in the semantics.
> >>
> >>peter
> >
> >Looks good.
> 
> I was thinking that too, but the comment was
> 
> >>>     RDFCore that owl:Class is not needed.  RDFCore requests the
> >>>     creation of test cases to clearly illustrate the differences
> >>>     between owl:Class and rdfs:Class.
> 
> and although that test case illustrates the differences
> it is not showing that owl:Class is needed.

Well, in a certain sense none of owl:Class, owl:DatatypeProperty,
owl:ObjectProperty, and probably quite a few other bits of OWL vocabulary
are not *needed*.  However, it is *desirable* to have them around.

> >             What about another one that somehow reflects that
> >rdfs:class is a member of rdfs:class, but owl:class is not a member
> >of owl:class?
> 
> Maybe owl:Class rdf:type owl:Class is not a legal DL document
> just as rdfs:Class rdf:type rdfs:Class isn't, but that doesn't
> mean that it's not legal OWL (and a true statement); again it
> is not showing that owl:Class is needed.
> 
> --
> Jos De Roo, AGFA http://www.agfa.com/w3c/jdroo/

peter

Received on Saturday, 24 May 2003 17:48:59 UTC