- From: Hart, Lewis <lhart@grci.com>
- Date: Tue, 23 Oct 2001 12:23:21 -0400
- To: "Peter F. Patel-Schneider" <pfps@research.bell-labs.com>
- Cc: www-rdf-logic@w3.org
Sorry for the confusion. #3 below is exactly what I am trying do to. State a local to the subject restriction on the object of a property. In my made-up, non-DAML+OIL compliant example I desire to state "If the property SuspectedAssoication is the subject of an assertion of the property EvidenceOf, then the object of that assertion must be an instance of the property ObservedAssociation." This is easily done if the subject is a DAML+OIL Class using a DAML+OIL Restriction. I now know that this can not be done within DAML+OIL if the subject is a DAML+OIL Property. Thanks - Lewis ___________________________________________ Lewis L Hart GRC International lhart@grci.com 1900 Gallows Rd. Voice (703)506-5938 Vienna, Va 22182 Fax (703)556-4261 -----Original Message----- From: Peter F. Patel-Schneider [mailto:pfps@research.bell-labs.com] Sent: Tuesday, October 23, 2001 9:55 AM To: lhart@grci.com Cc: www-rdf-logic@w3.org Subject: RE: Properties of Properties Question I'm still a bit confused. There are a number of things that you could be asking for: 1/ The ability to associate information with assertions. (In RDF terms this would probably be using a statment as the subject of another statement, which can't be done in RDF.) What you would be trying to do below would be to restrict the kinds of information that can be associated with a particular assertion. This can't be done in DAML+OIL because assertions cannot have associated information. 2/ The ability to associate information with a property. This can easily be done in DAML+OIL, because properties are resources. What you would be trying to do below would be to associate information with SuspectedAssociation, treated as a resource. However, this associated information will not restrict the kinds of assertions that can be made using SuspectedAssociation. 3/ The ability to restrict the kinds of assertions that be made using a property. Certain of these restrictions can be made in DAML+OIL, both globally and locally. What you would be trying to do below would be to limit the subject or object of SuspectedAssociation assertions to given classes. 4/ The ability to perform extra-DAML+OIL reasoning on assertions, but to limit the kinds of reasoning done on some assertions. You are, of course, on your own here. You could probably use method 2/ above to associate information with properties and use this information in your extra-DAML+OIL reasoner. I suspect that you want to do 4/, but I'm not sure. peter From: "Hart, Lewis" <lhart@grci.com> Subject: RE: Properties of Properties Question Date: Tue, 23 Oct 2001 09:25:34 -0400 > Using hypothetical 'RestrictionOfProperty' and 'toProperty' constructs, a > concrete example would be: > > <daml:ObjectProperty rdf:ID='SuspectedAssociation'> > <daml:subPropertyOf> > <daml:RestrictionOfProperty> > <daml:onProperty rdf:resource='#EvidenceOf'/> > <daml:toProperty rdf:resource='#ObservatedAssociation'/> > </daml:RestrictionOfProperty> > </daml:subPropertyOf> > </daml:ObjectProperty> > > Allowing you to build evidence of suspected associations (e.g. > conspiringWith, friendOf) based on observable associations (e.g. worksWith, > meetsWith) but disallowing suspected associations as direct evidence. This > example could also be stated differently, restricting the evidence to be a > Class 'Observation': > > <daml:ObjectProperty rdf:ID='SuspectedAssociation'> > <daml:subPropertyOf> > <daml:RestrictionOfProperty> > <daml:onProperty rdf:resource='#EvidenceOf'/> > <daml:toClass rdf:resource='#Observation'/> > </daml:RestrictionOfProperty> > </daml:subPropertyOf> > </daml:ObjectProperty> > > The crux of the problem is the lack of a 'RestrictionOfProperty' analogous > to daml:Restriction. It only seems reasonable to me that properties of > properties should be able to have restrictions in the same way properties of > classes can. Is there a reason that this should not be the case, other than > an implementation limitation of DAML+OIL? > > - Lewis > __________________________________________ > Lewis L Hart > GRC International lhart@grci.com > 1900 Gallows Rd. Voice (703)506-5938 > Vienna, Va 22182 Fax (703)556-4261 > > > > -----Original Message----- > From: Peter F. Patel-Schneider [mailto:pfps@research.bell-labs.com] > Sent: Monday, October 22, 2001 4:32 PM > To: lhart@grci.com > Cc: www-rdf-logic@w3.org > Subject: Re: Properties of Properties Question > > > From: "Hart, Lewis" <lhart@grci.com> > Subject: Property's of Properties Question > Date: Mon, 22 Oct 2001 16:07:47 -0400 > > > A simple question, with hopefully a simple answer. Is it possible to place > a > > restriction on the property of a property in DAML[+OIL]? > > No. At least I don't think so. Certainly not in general. > > > What I would like to say, for example is: > > > > <daml:ObjectProperty rdf:ID='Foo'> > > <daml:subPropertyOf> > > <daml:Restriction> > > <daml:onProperty rdf:resource='#Bar'/> > > <daml:toProperty rdf:resource='#Baz'/> > > </daml:Restriction> > > </daml:subPropertyOf> > > </daml:ObjectProperty rdf:ID='Foo'> > > > > Except for the fact that daml:Restriction is a daml:Class and > > daml:toProperty doesn't exist, this would be fine. > > If you have to use except, it isn't DAML+OIL. :-) > daml:Restriction forms a daml:Class, which is not really suitable as the > superproperty of a property. > > All that said, RDF allows entities to be both classes and properties, so > what you are doing above is to > 1/ create an unnamed class with a malformed restriction, which means that > there will be no semantic constraints for the class > 2/ assert that Foo is a subproperty of the property that has the same > identity as the class > > So you have said roughly the same thing as > > <daml:ObjectProperty rdf:ID='Foo'> > <daml:subPropertyOf> > <rdf:Property /> > </daml:subPropertyOf> > </daml:ObjectProperty rdf:ID='Foo'> > > as far as DAML+OIL is concerned. > > > Thanks - Lewis > > I don't think that you can do what you want. However, if you give a > concrete example, I would be able to tell for sure. > > Peter F. Patel-Schneider > Class of '72, GRCI
Received on Tuesday, 23 October 2001 12:23:50 UTC