- From: Aditya A Kalyanpur <adityak@wam.umd.edu>
- Date: Thu, 15 Apr 2004 10:37:27 -0400 (EDT)
- To: Natasha Noy <noy@SMI.Stanford.EDU>
- Cc: public-swbp-wg@w3.org
I agree - its a very clear and well presented document! a great start:) Regarding the approaches covered, I have an additional approach (a variant of approach 2) that seems reasonably intuitive to me. So in the current document, we define 'LionImage' as an Instance of Class 'AnimalImage' - why not make it a subclass of 'AnimalImage' instead? (i.e. define a set of all images that have Lion as subject) and instead of having a separate subject hierarchy (approach 2, option 2), have an image hierarchy where the subjects are restricted accordingly..so the class definitions would be: <owl:Class rdf:ID="AnimalImage"> <rdfs:subClassOf> <owl:Restriction> <owl:onProperty> <owl:ObjectProperty rdf:about="dc:subject"/> <owl:allValuesFrom rdf:resource="#Animal"/> </owl:onProperty> </owl:Restriction> </rdfs:subClassOf> </owl:Class> <owl:Class rdf:ID="LionImage"> <rdfs:subClassOf rdf:resource="#AnimalImage"/> <rdfs:subClassOf> <owl:Restriction> <owl:onProperty> <owl:ObjectProperty rdf:about="dc:subject"/> <owl:allValuesFrom rdf:resource="#Lion"/> </owl:onProperty> </owl:Restriction> </rdfs:subClassOf> </owl:Class> Thus, one could create instances of 'LionImage' referring to specific photos and have their 'dc:subject' point to specific 'Lion' individuals, which seems like a natural thing to do. Also, one could directly retrieve 'only images of Lions' this way (i.e. instances of class 'LionImage'), rather than search through all AnimalImage instances instead Also, I'm not entirely clear on the third point of consideration for approach 2 - why would creating an instance of class lion be 'inconsistent with the interpretation'? If the 'Lion' class is completely defined by what is stated in the document, then maybe that holds, however, given that the complete ontology would have other properties of 'Lion' and other classes that point to it, why would I need a separate Lion class to represent a specific lion at a zoo? Finally, maybe we should mention somewhere in the document that using 'allValuesFrom' restriction on property 'dc:subject' in the cases shown ensures that an 'AnimalImage' can only have its subject as 'Animal' and nothing else..hmm, makes me wonder if someValues is a better option - since it covers photos where we need to describe something more than just the animals in it (such as surroundings etc) - which would typically be the case? Thanks, Aditya On Thu, 15 Apr 2004, Jim Hendler wrote: > > At 9:27 -0700 4/14/04, Natasha Noy wrote: > >Folks, > > > >At the last telecon, I took an action to produce a draft note on > >dealing with the issue of classes as property values. My first pass > >at it is at > > > >http://lists.w3.org/Archives/Public/www-archive/2004Apr/att-0061/ClassesAsValues.html > > > >This is a strawman, so please feel free to and do poke holes in it! > >I think the goal here is really two-fold: (1) to give some > >suggestions on how to deal with the specific issue (classes as > >values) and (2) to figure out what the general format and tone of > >such notes and patterns should be. (The second one being arguably > >more important at this point) > > > >I have shamelessly lifted most of the ideas from the discussion on > >the mailing list a couple of weeks ago. So, thanks to everyone who > >participated in that discussion. I would also like to thank Oscar > >for his comments on an earlier draft. > > > >Natasha > > > I must admit that up until I saw this document, I was seriously > thinking of suggesting that OEP was a bad idea and that we should not > have the task force. Natasha's document shows me I was wrong, that > it may be possible for us to create some useful information for > Semantic Web users based on our experiences in our non-semantic web > -- if all of the things written up are done as carefully, > thoughtfully, and comprehensively as this - and are balanced in > presenting the solutions as this is - then maybe there's hope. > Natasha -- great job! > -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, 15 April 2004 10:37:34 UTC