Re: [closed] Re: Error in current WG test cases?

Thanks Jan, and thanks to Pat for his detailed comments.

I now understand that these test cases are correct wrt to the LC (and 
editor's draft) semantics document so this response is satisfactory.

Whilst we could, fairly easily, modify Jena to pass these specific test 
cases our implementation report remains unchanged for now.

[Like several RDF/RDFS implementations we are effectively implementing 
query over a closure rather than a direct entailment test. Including the 
conclusions vocabulary (as well as the premise vocabulary) in the closure 
is indeed required by the semantics document and technically 
straightforward but it leads to behaviour which seems likely to appear 
counter-intuitive to application developers. We are thus currently cautious 
about implementing this aspect.]

Dave

Jan Grant wrote:

> Dave, thank-you for your comment (and I apologise that this wasn't
> responded to earlier).
> 
> On Fri, 7 Nov 2003, Dave Reynolds wrote:
> 
> 
>>In a message on 25th July [1] I noted that the tests:
>>   rdfms-seq-representation/Manifest.rdf#test002
>>   rdfms-seq-representation/Manifest.rdf#test004
>>appeared to be incorrect.
>>
>>This comment does not seem to have been addressed and these tests remain in
>>the current wg approved tests (wg20031010).
> 
> 
> You comment generated quite a bit of discussion amongst the RDFCore WG.
> Those test cases are correct: according to the LC document,
> 
> [[
> S rdfs-entails E when every rdfs-interpretation of the vocabulary of S
> union E which satisfies every member of S also satisfies E.
> ]]
> 
> In particular, in these two test cases, interpretations of the empty
> graph are only considered when they contain the vocabulary (and
> therefore related axiomatic triples and semantic conditions) of the
> respective conclusions.
> 
> The test cases' correctness is unchanged with respect to the latest,
> post-last call editor's draft of the semantics document.
> 
> I hope this adequately addresses your comments. If you would be so
> kind, please respond, CC:ing the mailing list, indicating if this
> response is satisfactory.
> 
> Cheers,
> jan
>

Received on Monday, 17 November 2003 12:04:11 UTC