Re: FW: TopQuadrant response to OWL2 LC

Michael Schneider wrote:
>> -----Original Message-----
>> From: Ivan Herman [mailto:ivan@w3.org]
>> Sent: Wednesday, January 28, 2009 3:58 PM
>> To: Michael Schneider
>> Cc: W3C OWL Working Group
>> Subject: Re: FW: TopQuadrant response to OWL2 LC
>>
>> I agree that the current encoding in RDF looks fairly funny. 
> 
> IMO it's worse than just looking funny. The current two-triple encoding might make people believe that the left-hand-side property (a bNode in the RDF mapping, but it is, of course, allowed to be a URI in OWL Full) is a reference to the (sub-) property chain. And then people might go and use this reference elsewhere in the ontology expecting certain entailments from it. But the LHS property is actually only the root node of the syntactic construct (just as the bNodes of other multi-triple RDF encodings), and has no particular semantic relevance for the language construct. We don't have "real" property chains in OWL 2 Full, and in particular, we don't have a means to refer to them. 
> 

Agreed.

>> Ie, having
>> what you propose does look indeed better. My only cosmetic comment is
>> the usage of the term 'subsume' which, although we use it in our
>> conversations and logicians know what it means, it does not appear among
>> the usual RDF terms. 
> 
> Fair enough.
> 
>> May I propose, instead:
>>
>> q owl:superProperty (SEQ p_1....)
>>
>> which seems to be closer to the term subProperty?
>>
>> Of course, the question arises whether
>>
>> q owl:superProperty p
>>
>> is also valid where 'p' is a property and not a list...
> 
> I'm not so much for the "subsumes" here. But I would really like to see three aspects satisfied by such a property name:
> 
> 1) It should somehow talk about the "chain".
> 
> 2) It should somehow talk about the "sub".
> 
> 3) It should be clear what the super property and what the sub property chain is.
> 
> This all gets reasonably clear from my proposal, I think, but not from yours: 1) is completely missing, and 3) is at least not perfectly clear. 
> 
> However, I admit that point 3) is certainly not the most important point, because sub-property-chain triples will be asymmetric, having a property name on the LHS and a list of property names on the RHS. So confusion is actually not really possible.
> 

O.k., I take your points. But I am afraid of the confusion created by
the LHS and the RHS, more exactly the assymetry with rdfs:subPropertyOf

What was wrong with having something like

(SEQ p_1 ... p_n) owl:subPropertyChainOf q .

Though I am afraid I know: it looks quite awful in RDF/XML because the
parseType="collection" cannot be used:-( Sigh... It is o.k. in turtle...

If this is indeed the case, then I would weaken your case:-) and go for

q owl:superOfPropertyChain (SEQ p_1 ... p_n)

ain't nice, I know... but avoids confusions with LHS and RHS...

Ivan




> So, with this in mind, what about:
> 
>   q owl:subPropertyChain (SEQ p_1 ... p_n)
> 
> Would be the most obvious, I guess.
> 
>> Ivan
> 
> Cheers,
> Michael
> 
>> Michael Schneider wrote:
>>>  [send to WG list only for internal discussion]
>>>
>>> >From the "Other Comments" section:
>>>
>>> [[
>>> Property chain inclusion axioms
>>>     These appear to be quite widely useful. We have some concerns with
>> the use of blank nodes in the subPropertyOf triple corresponding to a
>> RIA. These are likely to cause problems for RDF implementations which
>> expect all predicates to be URI nodes. We think that drilling down and
>> fixing all instances where RDF software makes this assumption is costly
>> and unlikely to happen and to introduce incompatibility between OWL2 and
>> RDF. We believe introducing a new property in the RDF mapping for RIA
>> and avoiding the use of subPropertyOf is probably a better trade off
>> here.
>>> ]]
>>>
>>> I would be fine with this proposal, also from the OWL 2 Full
>> perspective, because the current encoding introduces a certain form of
>> "semantic side effect" (not too bad, otherwise I would have spoken up
>> earlier, but it's at least an annoyance). And it would simplify the
>> semantic conditions in OWL 2 Full. Also, I would say that we should, in
>> general go with single property solutions whenever possible.
>>> So, in the case that we decide to follow this proposal, here is, for a
>> start, my concrete suggestion for a new RDF encoding of sub property
>> chains:
>>>   q owl:subsumesPropertyChain (SEQ p_1 ... p_n)
>>>
>>> for given property names q and p_1 ... p_n .
>>>
>>> Michael
> 
> --
> Dipl.-Inform. Michael Schneider
> Research Scientist, Dept. Information Process Engineering (IPE)
> Tel  : +49-721-9654-726
> Fax  : +49-721-9654-727
> Email: schneid@fzi.de
> WWW  : 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
> Stiftung Az: 14-0563.1 Regierungspräsidium Karlsruhe
> Vorstand: Prof. Dr.-Ing. Rüdiger Dillmann, Dipl. Wi.-Ing. Michael Flor,
> Prof. Dr. rer. nat. Dr. h.c. Wolffried Stucky, Prof. Dr. rer. nat. Rudi Studer
> Vorsitzender des Kuratoriums: Ministerialdirigent Günther Leßnerkraus
> 
> ==============================================================================
> 

-- 

Ivan Herman, W3C Semantic Web Activity Lead
Home: http://www.w3.org/People/Ivan/
mobile: +31-641044153
PGP Key: http://www.ivan-herman.net/pgpkey.html
FOAF: http://www.ivan-herman.net/foaf.rdf