RE: Possibly wrong use of RL-theorem in the Conformance spec

>-----Original Message-----
>From: Ian Horrocks [mailto:ian.horrocks@comlab.ox.ac.uk]
>Sent: Tuesday, July 14, 2009 7:49 AM
>To: Michael Schneider
>Cc: W3C OWL Working Group
>Subject: Re: Possibly wrong use of RL-theorem in the Conformance spec
>
>The idea is to identify circumstances under which the RL rules are
>complete, i.e., where they can be correctly used to determine non-
>entailment. 

I always understood theorem TR1 to show a relationship between what was
formerly known OWL 2 RL-DL and OWL 2 RL-Full, i.e. where these two languages
coincide.

>Is it not the case that a direct semantics non-entailment
>where the two ontologies satisfy the conditions in Theorem PR1 is
>also an RDF-based semantics non-entailment?

I have to admit that I have never before thought about this question
(probably due to my original understanding of the theorem as stated above). 

Well, hard to tell. The RDF-based semantics definitely provides, in some
situations, additional entailments compared to the Direct Semantics. The
question is whether this is also the case under the restrictions given by
TR1. It would need some careful investigation of the whole semantics to find
this out (or some sudden inspiration for its refutation ;-). The result
would be some sort of "little sister" of the OWL 2 correspondence theorem. 

Regrettably, I won't find the time to check this, at least for the remainder
of the CR phase. I have even still to work on the correspondence theorem
itself, where I already know of several bugs that need to be fixed in the
next two weeks (but *this* was always on my schedule at least :). 

>Ian

Best,
Michael

>On 3 Jul 2009, at 20:29, Michael Schneider wrote:
>
>> Hi!
>>
>> In Section 2.3.1 of the Conformance spec, appended to the
>> definition of an "OWL 2 RL entailment checker", there is the
>> following note:
>>
>> [[
>> Note that it follows from Theorem PR1 of Profiles [OWL 2 Profiles]
>> that it is always safe for an OWL 2 RL entailment checker
>> using the RDF-Based Semantics
>> to return False if:
>> ... several conditions ...
>> ]]
>>
>> But the actual theorem PR1 does only make a assertion about the
>> relationship between the RL ruleset and the *Direct Semantics*.
>> Nothing is said about the RDF-Based Semantics by that theorem.
>>
>> So I believe that in the above note, the term "RDF-Based Semantics"
>> has to be replaced by the term "Direct Semantics".
>>
>> Michael

--
Dipl.-Inform. Michael Schneider
Research Scientist, Dept. Information Process Engineering (IPE)
Tel  : +49-721-9654-726
Fax  : +49-721-9654-727
Email: michael.schneider@fzi.de
WWW  : http://www.fzi.de/michael.schneider
=======================================================================
FZI Forschungszentrum Informatik an der Universität Karlsruhe
Haid-und-Neu-Str. 10-14, D-76131 Karlsruhe
Tel.: +49-721-9654-0, Fax: +49-721-9654-959
Stiftung des bürgerlichen Rechts, Az 14-0563.1, RP Karlsruhe
Vorstand: Prof. Dr.-Ing. Rüdiger Dillmann, Dipl. Wi.-Ing. Michael Flor,
Prof. Dr. Dr. h.c. Wolffried Stucky, Prof. Dr. Rudi Studer
Vorsitzender des Kuratoriums: Ministerialdirigent Günther Leßnerkraus
=======================================================================