RE: A possible way of going forward with OWL-R unification (ISSUE-131)

Bijan Parsia wrote:

>On 16 Jul 2008, at 18:01, Ivan Herman wrote:
>
>> Peter F. Patel-Schneider wrote:
>[snip]
>>> 4.4
>>> 	The rules from Section 4.3 can be applied to arbitrary RDF
>>> 	graphs, in which case the produced consequences are sound but
>>> 	not necessarily complete.
>>
>> I have already objected to this type of description elsewhere
>>
>> HTTP://www.w3.org/mid/487A187C.4070509@w3.org
>>
>> this type of slightly derogatory description
>
>How is it derogatory? 

It doesn't sound derogatory to me, but...

>It's an accurate description and I think it's a
>more useful conceptualization for users. Certainly better than
>"semantic subset" which, frankly, I often don't understand :)

... would be confusing to me if I was a potential implementer. The message
would probably be to me: Even if I implement the whole set of official OWL R
triple rules, I cannot safely assume to have a complete OWL R reasoner at
the end. This may technically really be the case, but it's certainly hard to
communicate.  

>> is certainly not what vendors would put as part of their product
>> announcement
>
>http://jena.sourceforge.net/inference/#overview
>
>"""RDFS rule reasoner
>Implements a configurable subset of the RDFS entailments.
>OWL, OWL Mini, OWL Micro Reasoners
>A set of useful but incomplete implementation of the OWL/Lite subset
>of the OWL/Full language.
>DAML micro reasoner"""

I think this misses the point. In the case of the Jena reasoner, I, as a
user of that reasoner, understand this reference to OWL/Lite more as a means
for getting some coarse idea about the semantic expressivity of the Jena
reasoner. Jena/OWL was never intended to be an *exact* implementation of
OWL/Lite-Full, and that's pretty clear from the feature list (see the same
document): For example, Jena supports class disjointness, and unionOf to
some degree, two features which are not in OWL/Lite.

However, if a vendor is willing to implement OWL-R exactly, then such a
statement as the above one about soundness but not guaranteed completeness
is not at all useful: A simple RDFS implementation (or even less) will have
the same properties.

So let's better drop this statement (as Peter proposed)! This doesn't change
the situation from a technical pov, but as long as no one asks... ;-)
 
>> let alone the fact that they would not even have a clear name and
>> standard to refer to. I regard that as a major problem.
>
>I'm confused. It seems like there is.

I didn't get this argument either.

@Ivan, can you please elaborate on this?

>Cheers,
>Bijan.

Cheers,
Michael

--
Dipl.-Inform. Michael Schneider
FZI Forschungszentrum Informatik Karlsruhe
Abtl. Information Process Engineering (IPE)
Tel  : +49-721-9654-726
Fax  : +49-721-9654-727
Email: Michael.Schneider@fzi.de
Web  : http://www.fzi.de/ipe/eng/mitarbeiter.php?id=555

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 Regierungspräsidium Karlsruhe
Vorstand: Rüdiger Dillmann, Michael Flor, Jivka Ovtcharova, Rudi Studer
Vorsitzender des Kuratoriums: Ministerialdirigent Günther Leßnerkraus