From: Jeff Heflin (heflin@cse.lehigh.edu)
Date: 11/01/01
Frank, I agree that you should be able to refer to classes before they are defined (as long as they are defined somewhere in the file), but I have strong misgivings about being able to refer to undefined classes. In my opinion, an ontology is supposed to provide two things: a standard vocabulary and formal definitions for that vocabulary. If you allow the use of undefined classes, you defeat the purpose of a standard vocabulary. This will also have very serious side effects in practical usage of the ontology. It means that any time somebody makes a typo, they've just created a new class. Is that really the way we want the Semantic Web to work? As Mike said in an earlier message, I think we need to strike a balance between database-style integrity constraints and inference opportunities. I think the way to address the problem is to consider the tradeoffs. In this case, the reference of undefined classes opens up a door for simple typos to result in misuse of the ontology. And what do we really gain from this? A shorthand way for creating classes that have no semantics? I think this is a case where the integrity constraints approach makes more sense. That is, I believe if you reference a class (or a property) in a particular namespace, that class (or property) MUST be defined in that namespace! Otherwise, your reference is in error. Note, that this does not mean that I come down on the side of "database-style integrity constraints" in every case. As I said before, I think a case-by-case cost/benefit analysis needs to be done. Jeff Frank van Harmelen wrote: > > Jeff wrote: > > > > 2) The Person class has a restriction stating the a person can have only > > > one FullTimeOccupation, but FullTimeOccupation is not defined in the > > > ontology. The FullTimeOccupation class should be added to the ontology. > > Peter answered: > > > It probably should be, but it is not strictly necessary, as its position in a > > restriction makes it an rdf:Class. > > We've discussed this a few times in the past. It was then considered a deliberate feature of the example to show that DAML+OIL allows the usage of (e.g.) classes before they are "defined", or even when they are not "defined" at all (particularly in a Web context). > > I still consider this a feature of the example, not a bug. > > Frank. > ----
This archive was generated by hypermail 2.1.4 : 04/02/02 EST