- From: Jim Hendler <hendler@cs.umd.edu>
- Date: Wed, 11 Sep 2002 17:37:04 -0400
- To: Jeff Heflin <heflin@cse.lehigh.edu>, Tim Finin <finin@cs.umbc.edu>
- Cc: www-webont-wg@w3.org
At 4:13 PM -0400 9/11/02, Jeff Heflin wrote: >Tim, > >I believe that you have done it correctly. My personal opinion is that >when you refer to classes and properties in a foreign ontology without >importing it then technically you should not include the axioms from the >foreign ontology when reasoning. Of course, this is almost always an >error, or the result of someone like Dan who doesn't believe in imports. >So we may decide to be lenient on this one and say that there is an >implicit imports of any ontology that includes a class or property that >is referenced. I'm not wild about that idea, but could live with it. >Either way, we need to make the expected behavior clear in our >semantics. > >Jeff ---- chair neutrality put into drydock somewhere ---- I have a LOT of problems with that and couldn't live with it. In my work, we often point to a specific class definition in another document without wanting to import all the axioms. In fact, we've developed the convention of using import when we do want the whole thing, and just pointing when we don't. In this sense the Xincludes would work for us just fine. As an example, the OpenCyc ontology contains a lot of useful classes (we tend to use real objects like dog, cat, etc. - Dan C. uses some of the more abstract things like time, I believe). However, we do not want to import all of CYC to use one piece -- we use the following operational semantics when we point to a specific property or class we treat it as if we did a lexical import of just that URI. WHen we use imports, we treat it as the lexical import of the entire document the ontology appears in. We realize neither of these is completely correct, but it works well enough for our purposes. The specific tool we are playing with is a markup tool called SMORE [1] which allows users to mark up documents against an ontology they create by picking and choosing pieces of existing ontologies and/or extending them -- thus it fits in our ontology sharing and linking use cases. [1] http://www.mindswap.org/~aditkal/editor.html (note: the downloadable beta of SMORE has some RDF bugs in it - a new version will be out soon that fixes the RDF dump and simplifies the interface) --- cloak of chair neutrality re-donned ---------- -- Professor James Hendler hendler@cs.umd.edu 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 240-731-3822 (Cell) http://www.cs.umd.edu/users/hendler
Received on Wednesday, 11 September 2002 17:37:08 UTC