- From: Jeremy Carroll <jjc@hplb.hpl.hp.com>
- Date: Sat, 16 Mar 2002 15:14:09 -0000
- To: "Dan Connolly" <connolly@w3.org>
- Cc: <www-webont-wg@w3.org>
> If I have to mention, in my student/person ontology, > all of the possible constructions of classes that might > be relevant, including all the combinations of > student/person with all other classes in the web, > we're pretty much nowhere, aren't we? > No you don't have to mention them. If nobody mentions them these constructs do not appear. If you or somebody else mentions them then the constructs are present in all interpretations of the complete set of assertions and they have the meaning. So maybe you say: John a Person . John a Student . And I ask a query system WHO a owl:intersectionOf(Person Student) . The query system then adds the classes needed to express the query to the knowledge base e.g. by asserting _:2 owl:intersectionOf(Person Student ). then the rephrased query WHO a _:2 . can be answered by entailment . John a Person . John a Student . _:2 owl:intersectionOf(Person Student ). ==> John a _:2 . and your knowledge (described in one way with one set of constructions) has been shared with me asking a query using a different construction. Jeremy
Received on Saturday, 16 March 2002 10:15:25 UTC