Presumably what is intended here is that this creates a named class that is equivalent to a restriction. But as you said, the owl api, sees that this class is named and then seems to ignore the restriction predicates.
I am not sure if the owl api is wrong in doing this. It seems like it probably should not behave this way. But I will say this construct seems suspect to me and will probably give many tools a problem. Interestingly Protege 3 does the opposite. It sees the restrictions but seems to forget the name.
As far as I know, this isn't valid OWL. Nodes that are typed as
restrictions should be anonymous nodes. You need to specify that it
is equivalent to, or a subclass of (or a disjoint with) the named OWL
How did you create this RDF?
On 28 Jul 2008, at 19:23, Timothy Redmond wrote:
> Cool! A KAoS ontology. I worked with KAoS a long time ago...
> I don't know what happened to your message but here is an answer.
> This is an owl api issue. I will post a message to the owl api
> developers group and see what they say.
> However you are doing something a little strange here and this is
> what is giving you trouble. The restrictions in the ontology you
> sent are named:
> <owl:Restriction rdf:about="urn:KAoS#AddPolicyAction-Subclass-
> <owl:ObjectProperty rdf:about="http://ontology.ihmc.us/PolicyAction.owl#involvedPolicy > "/>
> <owl:allValuesFrom rdf:resource="http://ontology.ihmc.us/Policy.owl#PosObligationPolicy > "/>
> Presumably what is intended here is that this creates a named class
> that is equivalent to a restriction. But as you said, the owl api,
> sees that this class is named and then seems to ignore the
> restriction predicates.
> I am not sure if the owl api is wrong in doing this. It seems
> like it probably should not behave this way. But I will say this
> construct seems suspect to me and will probably give many tools a
> problem. Interestingly Protege 3 does the opposite. It sees the
> restrictions but seems to forget the name.
> p4-feedback mailing list
> [hidden email] > https://mailman.stanford.edu/mailman/listinfo/p4-feedback
> I agree, I think it is not a valid OWL as well but two rather
> facts (as already pointed by Timothy) made me share the issue with
> 1) when parsing in P4, build 64, no error is generated, the
> is parsed and rendered WITHOUT the restrictions
> 2) when parsing in P3.4 beta, no error is generated, the ontology
> parsed and rendered WITH the restrictions
Well, I suppose both tools try to do some patching. We originally had
the parser in a strict mode in the OWL API (and hence P4) but this
seemed more problematic for many users. I'll add some functionality
into the OWL API so that these errors can be retrieved after parsing.
> The OWL piece was generated by KPAT v.2.0.
o.k. I don't know anything about this tool, but it may be worth
discussing this problem with the appropriate people.