Re: fresh blank nodes

I'm not sure what more formal would be here, to your taste.

Let's identify rule applications in terms of output triples in a document. An output triple is part of a rule's (R) application IFF the triple is part of the output graph due to the "firing" of R. Two applications of a rule are distinct iff they have (numerically) distinct sets of output triples.

The constraint implies that no two sets of output triples share bnode ids. It also implies that no two distinct rule applications (with Bnodes in the output) can be merely numerically distinct.

 Thus, yes, your example is incorrect.

Whether the parser should throw and error or warning is outside the scope of the specification. I'd certainly expect the validator to do so. The OWL API performs some repairs on input.

Cheers,
Bijan.

Received on Tuesday, 23 November 2010 09:28:31 UTC