At 23:33 -0800 3/24/04, Uschold, Michael F wrote: Jim makes a good point, and it is most valid in those cases when alternate approaches will work well. But that is not always the case, and there are many gray areas. When there ARE clear arguments for or against a given modeling choice, then I believe it IS the role of this group to identify commonly arising BAD ways to model things and recommend to avoid them, as well as to recommend GOOD ways to do certain kinds of things. We should avoid taking positions UNLESS there are clear arguments one way or the other, and as Jim says, indicate the consequences of decisions, so users can choose what will work best in their particular circumstances. MIke while I don't disagree, let me be clear -- it is rarely the case that one thing is right and one is wrong in the capital letter sense that Mike uses GOOD and BAD -- I don't mind educating in the few cases there may be, but I'm gonna be awfully hard to convince that somehow the way we built closed-world, generall small ontologies in the traditional AI systems is going to be the noly GOOD way to build them on the Semantic Web -- in fact, in teaching SW to my students, I'm learning new lessons all the time that counter how I've been teaching things in my AI courses for the last 28 years -JH -- Professor James Hendler http://www.cs.umd.edu/users/hendler Director, Semantic Web and Agent Technologies 301-405-2696 Maryland Information and Network Dynamics Lab. 301-405-6707 (Fax) Univ of Maryland, College Park, MD 20742 240-277-3388 (Cell)
Received on Thursday, 25 March 2004 08:40:03 UTC