Re: TEST: Re: notes for 6/6 until 1:10 (oneOf/sameClassAs)

On Thu, 2002-06-06 at 17:31, Jos De_Roo wrote:
> 
> [...]
> 
> > 3d) Proposal to close issue 2.4 - Enumerated Classes (daml:oneOf)
> > issue:
[...]
> > Dan will reconsider a test case posted by Jos.
> 
> that is actually the one in
> http://lists.w3.org/Archives/Public/www-webont-wg/2002May/0276.html
> "TEST: sameClassAs testcase" [1][2][3]

i.e. from nothing, conclude:

[ owl:oneOf ( :a :a :b ) ] owl:sameClassAs [ owl:oneOf ( :b :a :a ) ] .


Well, my position on 5.10-DAML+OIL-semantics-is-too-weak
  http://lists.w3.org/Archives/Public/www-webont-wg/2002May/0235.html
is that OWL shouldn't entail the existence of any
classes from an empty premise.

i.e. there shouldn't be any axioms with existentials in
the conclusions. (there's a name for that fragment of FOL, no?
is that horn clauses? I often forget).

So I'm currently against rules such as:
  { :rule9o1 . ?L owl:item ?x } log:implies { ?x a [ owl:oneOf ?L ] } .
from
  http://www.agfa.com/w3c/euler/owl-rules


I worked out a similar test:

premise:

 :x owl:oneOf ( :a :a :b ).
 :y owl:oneOf ( :b :a :a ).

conclusion:

 :x owl:sameClassAs :y.

and I'm satisfied OWL should give us that much.

(I haven't gotten as far as checking the test in;
I had to kludge around a few problems to get it working.)


-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/