Re: RDF and OWL rules

Geoff Chappell wrote:
> 
> I agree that's probably a better goal (at least initially). So what's
> the best way to express the scope of the effort - an axiomatization of
> rdf(s) and owl in LP-style rules? 

Perhaps "partial axiomatization" would be closer, since pure LP rules can't
completely capture even OWL-lite. 

It'd be nice to turn it round the other way, and treat it as a definition (and
axiomatization) of that subset of OWL entailments that are within scope of
rules-style reasoners. One of the dangers with OWL is that we'll end up with
lots of implementations which offer incomplete support for different subsets of
OWL-full leading to interoperability problems. A community agreed subset which
is practically implementable using modest technology would be great. The problem
here is that I'm talking as a "scruff" - you can't really subset semantics this
way.

> Tests that help to define the boundaries of the owl subset achievable
> via an axiomatization would be particularly useful.

Exactly.

Dave