- From: Jos De_Roo <jos.deroo.jd@belgium.agfa.com>
- Date: Wed, 8 Jan 2003 00:42:14 +0100
- To: "Deborah McGuinness <dlm" <dlm@ksl.Stanford.EDU>
- Cc: Dan Connolly <connolly@w3.org>, Christopher Welty <welty@us.ibm.com>, www-webont-wg@w3.org, www-webont-wg-request@w3.org
right, that's indeed a very good idea (just that the current two testcases or so use owl:allDistinct as a name, but what's in a name, owl:allDifferent sounds good) -- , Jos De Roo, AGFA http://www.agfa.com/w3c/jdroo/ Deborah McGuinness <dlm@ksl.Stanford. To: Dan Connolly <connolly@w3.org> EDU> cc: Christopher Welty <welty@us.ibm.com>, www-webont-wg@w3.org Sent by: Subject: Guide Comment with respect to the Unique Names issue (5.18) www-webont-wg-requ est@w3.org 2003-01-08 12:19 AM I did a search for 5.18 in the subject line to see if there was a last message on the unique names assumption issue. I enclose the last message I found on the topic that does not make me think we have resolution. The reason i did this was that I was reviewing the guide wines ontology and I would like to make a recommendation for a cleanliness pass over the ontology. My experience in the past from this kind of document is that the ontology gets picked up and expanded upon in a number of tutorial settings. If the ontology author has not done a little bullet proofing in advance on the ontology, students end up expanding it in all sorts of unexpected and sometimes confusing ways. Thus, I would like for the ontology not to only be correct but to have a good foundation for extensibility. Right now it is open to some unintended consequences if it gets used by naive users. One of those areas is impacted by the unique names assumption and differentIndividual statements. Given that we have a number of functional properties: hasSugar, hasFlavor, hasBody, etc. and given that a user could inadvertently state that for example, the hasFlavor property for a wine is both strong and moderate, that would result in the unintential incorrect (from a conceptual modeling perspective) inference that Strong = Moderate. We as the designers of the ontology can keep this inference from happening without signaling a contradiction by stating that strong is different from moderate. The ontology of course has a number of wine properties all of which have sets of values. Within each sets of values, we should have some way of stating that: full, medium, and light are different. red, rose, and white are different. strong, moderate, and delicate are different. dry, offDry, and sweet are different, etc. with our pairwise statements, this of course gets ugly fast. Still I think the pairwise statements should be added to the ontology from a pedagogical perspective of what it takes to write an ontology that is expected to be used. I would prefer though to suggest that we use allDifferent if we can agree to add that or some other construct. so this message has two points: 1 - it points out a place in our guide where we would greatly benefit from some notion of allDifferent 2 - it strongly suggests that the guide ontology include some way of stating difference between all of the individuals that fill functional properties. If that must be done in a pairwise method, then i would include those statements but if we have included some kind of allDifferent, we can use that in the ontology. d Dan Connolly wrote: > On Thu, 2003-01-02 at 20:15, Christopher Welty wrote: > > Dan, > > > > I think what Mike meant was explain the LEGAL syntax. It will be no > > problem to describe what it *does*, you tell us how to say it. > > I can't think of a nice looking syntax; that's why I'm > lukewarm on AllDifferent. But you can > see an example that has the right meaning in Jos's message... > > As I said... > > It seems Jos already made the point in his message > > of Sat, 21 Dec 2002 20:56:12 +0100 > > > > | I just don't see how rdf:parsetype="Collection" > > | could work in this case > > his message continues... > > ======= > excerpt from Jos's message > http://lists.w3.org/Archives/Public/www-webont-wg/2002Dec/0273.html > > I just don't see how rdf:parsetype="Collection" > could work in this case, but anyhow, e.g. > > <owl:AllDistinct rdf:nodeID='A0'> > <rdf:first rdf:resource='premises001#a'/> > <rdf:rest rdf:nodeID='A1'/> > </owl:AllDistinct> > <owl:AllDistinct rdf:nodeID='A1'> > <rdf:first rdf:resource='premises001#b'/> > <rdf:rest rdf:nodeID='A2'/> > </owl:AllDistinct> > <owl:AllDistinct rdf:nodeID='A2'> > <rdf:first rdf:resource='premises001#c'/> > <rdf:rest rdf:resource > ='http://www.w3.org/1999/02/22-rdf-syntax-ns#nil'/> > </owl:AllDistinct> > > should entail > > <rdf:Description rdf:about="premises001#a"> > <owl:differentFrom rdf:resource="premises001#c"/> > </rdf:Description> > ======= > > -- > Dan Connolly, W3C http://www.w3.org/People/Connolly/ -- Deborah L. McGuinness Knowledge Systems Laboratory Gates Computer Science Building, 2A Room 241 Stanford University, Stanford, CA 94305-9020 email: dlm@ksl.stanford.edu URL: http://ksl.stanford.edu/people/dlm/index.html (voice) 650 723 9770 (stanford fax) 650 725 5850 (computer fax) 801 705 0941
Received on Tuesday, 7 January 2003 18:43:34 UTC