- From: Ray Fergerson <fergerson@SMI.Stanford.EDU>
- Date: Tue, 29 Feb 2000 14:41:08 -0800
- To: Tim Berners-Lee <timbl@w3.org>
- CC: www-rdf-interest@w3.org, Stefan Decker <stefan@DB.Stanford.EDU>
Tim, Protege has a root class :THING and a subclass of :THING called rdfs:Resource. We considered making rdfs:Resource an alias of :THING but we could not because rdfs:Resource has properties (such as "isDefinedBy" and "seeAlso"). Thing has no properties (other than "name" which some would call a property). Thus we had to have two separate classes. Now all RDF users should probably only subclass rdfs:Resource but they are not prevented from subclassing :THING directly. If they do so then this information needs to be preserved. This is the point that Stefan is trying to make. Ray Tim Berners-Lee wrote: > > From: Stefan Decker <stefan@db.stanford.edu> > > >if an application does not know protege:thing, it can ignore it. > > Surely, if it does *not* know protege:thing it *cannot* ignore it without > throwing away information. > If it *does* know protege:thingthen it *can* ignore the statement as it > knows it contains no information. Every class is a subclass of Thing. > > >If a user defines a resource to be a subclass of protege:Thing, protege can > >not ignore this and we have to save that. > > No, you can just hang all classes not defined as subclasses of anything else > off the "thing" hook in the UI. > > >However, i think there is a small missunderstanding: you probably meant > >that #Calendar is a subclass of rdfs:Resource. > > In RDF, every class is a subclass of rdfs:Resource so surely that is just as > informationless too. > (I do wish RDF had used "thing" instead of "resource" which has a meaning in > URI already).
Received on Tuesday, 29 February 2000 17:40:03 UTC