What OWL doesn't do (LC response editing)

During the WWW conference, I spent some time talking to Ken Laskey, 
who raised a Last Call comment that basically said he wasn't 
convinced we'd built the right thing [1].  Ken is not only the AC 
representative for a large company, but also a member of the W3C 
Advisory Board, and I wanted to discuss his concerns to be absolutely 
sure we understood them.  Summarizing, he was concerned that our over 
slavishness (my words, not his) to the DL model had caused us to 
leave out some important things -- probabilities, defaults, 
procedural attachment and other things he felt would be important to 
the overall success of a web ontology langauge in the long run.
  After talking with Ken, and discussing what we'd been chartered to 
do, how we had proceeded, what was in DAML+OIL which we were based 
on, and the capabilities for extension in OWL Full, he was a bit more 
comfortable.  We decided that a good thing for the WG to do, would be 
to add a short section somewhere in the documents, explaining a 
little more clearly what the scope of OWL is, what sort of things are 
left out of Owl and why (lack of community consensus, difficult 
research issues, non-monotonicity, etc.).  In short, Ken would really 
like to see a section somewhere (I think maybe overview or guide) 
that explains a little better what OWL does, and more importantly 
doesn't, do.
   Anyone feel like taking a stab at this?
p.s. Might also be a good place to set the expectation that extending 
the OWL vocabulary, even in fairly small ways, is likely to take one 
out of DL -- that was another thing that came as a surprise to many 
people I talked to.


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)

Received on Wednesday, 28 May 2003 13:29:03 UTC