| [higgins-dev] Re: LDAP "person" sample ontology based on Higgins ontology |
All, Attached is an updated sample of the LDAP "person" ontology that Mark has been commenting on. This one will actually load in Protégé, wahoo! I added the additional definitions for the OID equivalence references as Mark suggested and it still wouldn't load. Then I noticed that the Properties were loading just fine, but not the Classes. So I looked at the Class definitions and noticed the equivalence references had not had the '#' prefixed on them. That fixed it so it would load. Then I looked at the Property equivalence references and noticed they did NOT have the '#' prefixed on them either. I looked at them in Protégé and realized that, for Property definitions, it assumes I'm referencing some global definition. When I added the '#' prefix, it used the "bandit" ones I had defined as per Mark's suggestion in my ontology file. So, this brings up the question, which should we be using? I've been following Sebastian Dietzold's lead on these where he did NOT prefix the '#' on either class or property equivalences which I assumed meant he was referencing an already globally defined URN that OWL parsers would allow. However, if that was true, I wouldn't have to define the urn:oid locally as Mark suggested. Is there some difference between Class definitions and Property definitions that I'm missing? Based on what I've seen, I could eliminate all the extra ObjectProperty definitions and be fine as long as the property equivalences don't use the '#'. But if I do that, what am I actually referencing? And, why can't I do that for OWL classes? More aid for an OWL newbie needed! :) BTW, thanks again Mark. I did download a couple of OWL plugins for Protégé including the Manchester plugin but it doesn't appear on my tools. I got the WonderWeb validator to show up but it doesn't handle the local "import" statement for the higgins ontology correctly so that didn't help either. Just figured it out the old fashioned way. Tom Doman Novell Inc. >>> "Tom Doman" <TDoman@xxxxxxxxxx> 9/21/2006 4:48 PM >>> Again, thanks Mark, you are correct, I left off the "Of". I think you're right about the object properties for the references equivalents because Protégé gave a different stack with it's latest failure that seems to indicate it's looking for referenced properties that aren't defined. I'll try that next. Tom Doman Novell Inc. >>> Mark Wahl <Mark.Wahl@xxxxxxxxxxxxxxxxxxxx> 9/20/2006 7:43 PM >>> Tom Doman wrote: > Good catch! Thanks! Unfortunately, Protégé still isn't happy with it. I also noticed that you are using rdfs:subProperty. Do you perhaps mean rdfs:subPropertyOf? (I don't remember seeing the former in http://www.w3.org/TR/rdf-schema/). With these two changes, your file appears to be OWL-Full. Another observation- Your object properties for your data model attrs have owl:equivalentProperty rdf:resource="urn:oid:...". It might be good for this OWL file to also have owl:ObjectProperty rdf:about="urn:oid:..." definitions for each of these attributes, otherwise parsers will run into difficulty as they won't know the RDF "type" of the resource indicated by the URN. Similarly, you might also want to have owl:Class definitions for each of the object class OIDs. (The Manchester parser gives these as reasons why your file is not OWL-DL or OWL-Lite). Mark Wahl Informed Control Inc. _______________________________________________ higgins-dev mailing list higgins-dev@xxxxxxxxxxx https://dev.eclipse.org/mailman/listinfo/higgins-dev
Attachment:
HigginsLDAP0.1.3.owl
Description: Binary data