From: tim finin ([email protected])
Date: 04/08/02
Adam -- I think this is quite good, but there is still a potential problem in that it doesn't answer the question of why DAML+OIL and not just RDFS. So, it would be good to add to point out that the first example is one that works in RDFS (which also has semantics and licenses inferences) but DAML+OIL has (and OWL will have) additional features powers that make it more expressive. I like the example of using the unique property attribute to infer that two individuals are the same if they have the same email address. Another example to build on might be that DAML+OIL allows one to say that MEN and WOMEN are disjoint classes. Knowing that John is a MAN lets us conclude that John is not a woman, so we know, perhaps, that we need not ask him certain questions in a medical interview. If we do this then we should also point out that DAML+OIL is intentionally limited in its expressive power. We should give an example or two of the kinds of knowledge that one can not express in DAML+OIL. An example might be "A person is necessarily younger than his parent". And we might point out the benefits of limiting the expressive power (e.g., making the reasoning more tractable) and that the decision about exactly how expressive to make it is a question open to debate and experimentation. Tim
This archive was generated by hypermail 2.1.4 : 03/26/03 EST