brief implementation report, OWL Full, "surnia"

I've been developing a fairly naive OWL Full implementation based on
simply creating first-order axioms for OWL, turning RDF into triples
like
    rdf(uri1, uri2, uri3)
and feeding the results to Otter.  There's a python wrapper for Otter
to do the translation, feed it OWL test cases, etc.   Web page is
    
   http://www.w3.org/2003/08/surnia/

When last I looked at it (about two weeks ago), it was passing 76% of
the approved Positive Entailment Tests, which are by far the easiest
tests for it.  It's hard to tell whether its difficulties are due to
flaws in the axioms, my lack of giving Otter appropriate guidance
through the search space, or the fundamental limitations of this
approach.  

     -- sandro

Received on Tuesday, 26 August 2003 13:46:19 UTC