TEST, SEM: test cases for dark triples

At the f2f we agreed the following paragraph:

>If the WebOnt language is to
>   1/ use RDF syntax, and
>   2/ be an extension of RDF, and
>   3/ to be able to entail appropriate class memberships,
>then the WebOnt language needs to have a facility for having RDF triples
>that do not denote in the RDF model theory (a.k.a. dark triples) and URI
>references that do not denote in the RDF model theory (e.g., URI references
>that only exist in dark triples).

A crucial part is "entail appropriate class memberships".

I think it would be a fruitful and worthwhile activity for both focus areas
to produce examples that show such appropriate entailments.

Having done that it is then easier to justify the word *needs*.

So far, we have had a variety of student/employee entailments from Peter in
January; I suggest demonstrating those by two:
A:
   John isa student
   John isa employee
entailing
   John isa (intersectionOf [student employee])

and

B:
   John isa (intersectionOf [student employee])
entailing
   John isa (intersectionOf [employee student])


we could also have

C:
  (empty but with vocabulary including <foo>)
entailing
  <foo> isa (oneOf [<foo>])



The main reason I am suggesting such entailments is that they seemed to be
the essence of why the WG was not happy with the solipsistic system at the
f2f.

Are there other "appropriate class memberships" entailments that help
understand the WG's resolution?

(Another example, that is unfortunately dead, is the Patel-Schneider
paradox. This would have been an example of an appropriate non-entailment,
if we hadn't dropped qualified cardinality constraints.)

I am happy to formalize some such test cases to go into the repository.

Jeremy


PS: I am particularly interested in seeing an entailment for which dark
triples are needed, I do not believe that A,B, or C (or for that matter the
Patel-Schneider paradox) is one.

Received on Thursday, 18 April 2002 08:45:33 UTC