Re: ISSUE-8 (dataproperty chains): REPORTED: add chains ending with data properties)

On 7 Nov 2007, at 15:23, Alan Ruttenberg wrote:

> Uli,
>
> Do the same decidability issues arise if only the last property on  
> the chain is a datatype property?
>

not really, but semantically, it doesn't make sense to have datatype  
properties anywhere else but in the last place of a chain: a datatype  
value can only ever occur as the "filler" of a property (or "have in  
incoming property edge"), but never be "subject" of a property.

Cheers, Uli

> -Alan
>
> On Nov 7, 2007, at 8:32 AM, Jim Hendler wrote:
>
>>
>> So we can allow this in OWL 1.1 Full, but not in OWL 1.1. DL since  
>> it is only related to decidability which is the primary  
>> differentiator between DL and Full.  So I propose that we include  
>> this construct in 1.1 but make it clear that using it will take  
>> you to Full.
>>  Since this is on agenda for discussion at a meeting I cannot  
>> attend, I state for the record that RPI would oppose any closure  
>> of this issue that would not allow a property chain to end in a  
>> datatype property in the RDF realization
>>  -JH
>> p.s. I realize now that my primary problem with the structural  
>> document relates to this DL v. Full issue, and will take that up  
>> in another thread.
>>
>>
>> On Nov 7, 2007, at 5:12 AM, Uli Sattler wrote:
>>
>>>
>>> Dear all,
>>>
>>> a few days ago, I sent this email below as an answer to Owl Dev  
>>> only, overlooking that I should have sent it to owl-wg as  
>>> well...so here it is with a bit of delay, cheers, Uli
>>>
>>>
>>> On 5 Nov 2007, at 15:13, Uli Sattler wrote:
>>>>
>>>> Hi Michael,
>>>>
>>>> there are reasons why these sub-property chains are only made up  
>>>> of object properties:  decidability in OWL (DL and 1.1) relies  
>>>> on the fact that "datatype consistency" can be checked for each  
>>>> object separately, without referring to other objects and the  
>>>> values of their datatype properties. If we would need to do  
>>>> this, we would more likely be in trouble, and would need to
>>>>
>>>> - be much more careful about what datatypes and datatype  
>>>> predicates to allow without loosing decidability and
>>>> - use more complex reasoning mechanisms that have, to the best  
>>>> of my knowledge, only been described on paper and never been  
>>>> implemented or tested.
>>>>
>>>> So, I can see your use case, but I don't think we know enough  
>>>> about this yet.
>>>>
>>>> If you want to know more, check out
>>>>
>>>> Carsten Lutz and Maja Milicic. A Tableau Algorithm for  
>>>> Description Logics with Concrete Domains and General TBoxes.  
>>>> Journal of Automated Reasoning. To appear.
>>>> http://lat.inf.tu-dresden.de/~clu/papers/archive/jar06.pdf
>>>>
>>>> Carsten Lutz. Description Logics with Concrete Domains - A  
>>>> Survey. In Philippe Balbiani, Nobu-Yuki Suzuki, Frank Wolter,  
>>>> and Michael Zakharyaschev, editors, Advances in Modal Logics  
>>>> Volume 4. King's College Publications, 2003.
>>>> http://lat.inf.tu-dresden.de/~clu/papers/archive/aiml4.ps.gz
>>>>
>>>> Cheers, Uli
>>>>
>>>>
>>>>
>>>>
>>>> On 2 Oct 2007, at 13:26, Michael Schneider wrote:
>>>>
>>>>
>>>>>
>>>>> Hi!
>>>>>
>>>>> It just stroke me that there seem to be only Sub/Object/ 
>>>>> PropertyChains in
>>>>> the current OWL-1.1 draft [1]. Does anyone know if there is a  
>>>>> problem with
>>>>> also having sub property chains of the form
>>>>>
>>>>>   SubDataPropertyOf(
>>>>>       SubDataPropertyChain(R1 ... Rn-1 Dn)
>>>>>       D )
>>>>>
>>>>> where Dn and D are DataPropertyS (having compatible datatypes  
>>>>> as their
>>>>> ranges), while R1 ... Rn-1 are ObjectPropertyS?
>>>>>
>>>>> With such a SubDataPropertyChain, one could for instance  
>>>>> translate rules
>>>>> like:
>>>>>
>>>>>   ?x hasFather ?y AND ?y hasFamilyName ?fn
>>>>>   ==> ?x hasFamilyName ?fn
>>>>>
>>>>> with ?fn being an xsd:string, into an equivalent OWL axiom
>>>>>
>>>>>   SubDataPropertyOf(
>>>>>       SubDataPropertyChain(hasFather hasFamilyName)
>>>>>       hasFamilyName )
>>>>>
>>>>> In this case, the super property whould equal the final chain  
>>>>> property (both
>>>>> 'hasFamilyName').
>>>>>
>>>>> An example for a more general rule type (the analogon of the  
>>>>> 'uncle' rule)
>>>>> would be:
>>>>>
>>>>>   ?g containsUser ?u AND ?u hasUserID ?i
>>>>>   ==> ?g containsUserWithID ?i
>>>>>
>>>>> where ?g would stand for some user group. Here, the DataPropertyS
>>>>> 'hasUserID' and 'containsUserWithID' differ from each other,  
>>>>> because they
>>>>> are intended to have a different meaning.
>>>>>
>>>>> Any ideas, if this feature has a chance to enter the family of  
>>>>> OWL-1.1 (or
>>>>> 1.2 :)) axioms? Or did I overlook some fundamental issue here?
>>>>>
>>>>> Cheers,
>>>>> Michael
>>>>>
>>>>> [1] OWL-1.1 Semantics
>>>>>     http://www.webont.org/owl/1.1/semantics.html#2
>>>>>
>>>>> --
>>>>> 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
>>>>>
>>>>>
>>>
>>
>> "If we knew what we were doing, it wouldn't be called research,  
>> would it?." - Albert Einstein
>>
>> Prof James Hendler				http://www.cs.rpi.edu/~hendler
>> Tetherless World Constellation Chair
>> Computer Science Dept
>> Rensselaer Polytechnic Institute, Troy NY 12180
>>
>>
>>
>>
>>
>

Received on Wednesday, 7 November 2007 16:49:42 UTC