W3C home > Mailing lists > Public > public-owl-wg@w3.org > December 2007

Re: ISSUE-83 (Vipul): Property Chain Axiom: P1 o P2 => P2 o P1

From: Jim Hendler <hendler@cs.rpi.edu>
Date: Thu, 6 Dec 2007 08:47:06 -0500
Message-Id: <99567A3B-DA24-4B50-8810-87BCC4F08147@cs.rpi.edu>
Cc: Ian Horrocks <ian.horrocks@comlab.ox.ac.uk>, Uli Sattler <sattler@cs.man.ac.uk>, <conrad.bock@nist.gov>, "'Kashyap, Vipul'" <VKASHYAP1@PARTNERS.ORG>, "'Jeremy Carroll'" <jjc@hpl.hp.com>, "'OWL Working Group WG'" <public-owl-wg@w3.org>
To: Alan Ruttenberg <alanruttenberg@gmail.com>
yes, that is a flaw in the OWL 1.0 docs I was hoping this working  
group would not repeat.
  -JH


On Dec 5, 2007, at 8:34 PM, Alan Ruttenberg wrote:

> Hi Jim,
>
> Not all of OWL1.0 full is specified in the abstract syntax, just  
> the part that is OWL-DL, no? For example, you can't use a bnode as  
> a class name in the abstract syntax, but you can in OWL full, and  
> you can't create a subclass of rdfs:List in the abstract syntax but  
> you can in OWL Full.
>
> Isn't this the same situation?
>
> -Alan
>
> On Dec 4, 2007, at 8:50 PM, Jim Hendler wrote:
>
>> I have no trouble with that, except that implies that the  
>> functional syntax cannot be the normative definition of the  
>> language given our charter.  This is fine with me (i.e. as in OWL  
>> the model theory is where the normative behaviors lie, and the  
>> language is fully defined therein) but I go back to my first  
>> point, we are not rejecting or postponing issue 83. We are  
>> accepting it only in OWL 1.1 Full, and whichever document is the  
>> normative definition of OWL 1.1 will need to include it as a  
>> possibility (i.e. define whether this particular pattern of  
>> triples has any additional semantics)
>>  -JH
>> p.s. I guess the functional syntax could be normative for OWL DL,  
>> but it does mean we need to define where we put those  
>> specifications of additional semantics qua Ian's response...
>>
>>
>> On Dec 4, 2007, at 2:06 PM, Ian Horrocks wrote:
>>
>>> OWL Full permits *many* things that are not covered in the  
>>> functional syntax. The functional syntax provides a "high-level  
>>> abstract syntax for both OWL DL and OWL Lite" [1]. Such a  
>>> specification is not necessary for OWL Full, as *any* RDF graph  
>>> is an OWL Full ontology. Additional semantics may of course be  
>>> given to certain (patterns of) triples as specified in the OWL  
>>> Full semantics.
>>>
>>> Ian
>>>
>>> [1] http://www.w3.org/TR/owl-semantics/
>>>
>>>
>>>
>>> On 4 Dec 2007, at 18:37, Jim Hendler wrote:
>>>
>>>> Not sure I understand this answer.  If the functional syntax  
>>>> won't cover the things that can be done in OWL 1.1 Full, then  
>>>> how could it be the normative definitions (note that I asked  
>>>> this same question with respect to inverseFunctional Datatypes  
>>>> and didn't get an answer there either).  If functional syntax is  
>>>> "functional syntax for OWL DL 1.1." and there's some sort of  
>>>> addendum for those things in the RDF that aren't in the  
>>>> functional syntax, I can live with that - but the charter does  
>>>> mention maintaining OWL Full, so I find having a normative  
>>>> definition that doesn't include it confuses me.
>>>>  -JH
>>>> p.s. This is obviously more general a question than to this  
>>>> issue - but it does come up here.
>>>>
>>>> On Dec 4, 2007, at 12:43 PM, Ian Horrocks wrote:
>>>>
>>>>> It can already be expressed in the existing OWL Full in the  
>>>>> sense that the relevant triples can be included in an OWL Full  
>>>>> ontology. It cannot, however, be expressed in the functional  
>>>>> syntax (which is also the case for inverseFunctional Datatypes  
>>>>> in both 1.0 and 1.1). Extending the functional syntax would be  
>>>>> undesirable for the reasons stated.
>>>>>
>>>>> Ian
>>>>>
>>>>>
>>>>> On 4 Dec 2007, at 17:22, Jim Hendler wrote:
>>>>>
>>>>>>>
>>>>>> My understanding from Jeremy's email in this thread is that in  
>>>>>> the OWL Full version of 1.1 this can be expressed - if that is  
>>>>>> the case, why is this postponed rather than accepted, but only  
>>>>>> for Full (like we do for  inverseFunctional Datatypes)
>>>>>>
>>>>>>
>>>>>>
>>>>>> On Dec 4, 2007, at 4:47 AM, Ian Horrocks wrote:
>>>>>>
>>>>>>>
>>>>>>> To summarise: This is not allowed in the existing syntax, it  
>>>>>>> would lead to undecidability if it were allowed (even for  
>>>>>>> very restricted language subsets), it is not supported by  
>>>>>>> implementations and seems unlikely to be supported in the  
>>>>>>> foreseeable future. I therefore propose to postpone it on  
>>>>>>> these grounds.
>>>>>>>
>>>>>>> Ian
>>>>>>>
>>>>>>>
>>>>>>> On 3 Dec 2007, at 18:59, Uli Sattler wrote:
>>>>>>>
>>>>>>>>
>>>>>>>> Hi Conrad -
>>>>>>>>
>>>>>>>> On 3 Dec 2007, at 15:38, Conrad Bock wrote:
>>>>>>>>
>>>>>>>>>
>>>>>>>>> Uli,
>>>>>>>>>
>>>>>>>>>>  makes reasoning undecidable (even R o S => T in general,  
>>>>>>>>>> ie, without
>>>>>>>>>>  the restrictions imposed by OWL 1.1 because it allows you  
>>>>>>>>>> to reduce
>>>>>>>>>>  the intersection problem of contex-free languages to  
>>>>>>>>>> satisfiability
>>>>>>>>>>  of concepts) ...
>>>>>>>>>
>>>>>>>>> Do you know of anyone working on restrictions that would  
>>>>>>>>> make chains in
>>>>>>>>> the "super" position (right hand side) decidable?
>>>>>>>>>
>>>>>>>>
>>>>>>>> I don't know of anybody currently working on it, but we know  
>>>>>>>> that it (having R o S => T o U)  makes reasoning undecidable  
>>>>>>>> in the logic that
>>>>>>>>
>>>>>>>> - has only IntersectionOf an SomeValuesFrom restrictions  
>>>>>>>> (see Baader, DL 2003, http://sunsite.informatik.rwth- 
>>>>>>>> aachen.de/Publications/CEUR-WS/Vol-81/baader.pdf)
>>>>>>>>
>>>>>>>> - has only intersectionOf and AllValuesFrom Manfred Schmidt- 
>>>>>>>> Schau . Subsumption in KL-ONE is undecidable. In Ron J.
>>>>>>>> Brachman, Hector J. Levesque, and Ray Reiter, editors, Proc.  
>>>>>>>> of the 1st Int.
>>>>>>>> Conf. on the Principles of Knowledge Representation and  
>>>>>>>> Reasoning (KR'89),
>>>>>>>> pages 421-431. Morgan Kaufmann, Los Altos, 1989.
>>>>>>>>
>>>>>>>> Cheers, Uli
>>>>>>>>
>>>>>>>>> Conrad
>>>>>>>>>
>>>>>>>>>>  On 30 Nov 2007, at 14:53, Kashyap, Vipul wrote:
>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> Would be interested in on the ramifications on the  
>>>>>>>>>>> complexity of
>>>>>>>>>>> reasoning.
>>>>>>>>>>>
>>>>>>>>>>> ---Vipul
>>>>>>>>>>>
>>>>>>>>>>>> -----Original Message-----
>>>>>>>>>>>> From: public-owl-wg-request@w3.org
>>>>>>>>>>>> [mailto:public-owl-wg-request@w3.org] On Behalf Of  
>>>>>>>>>>>> Jeremy Carroll
>>>>>>>>>>>> Sent: Friday, November 30, 2007 8:53 AM
>>>>>>>>>>>> To: OWL Working Group WG
>>>>>>>>>>>> Subject: Re: ISSUE-83 (Vipul): Property Chain Axiom: P1  
>>>>>>>>>>>> o P2
>>>>>>>>>>>> => P2 o P1
>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>> OWL Working Group Issue Tracker wrote:
>>>>>>>>>>>>>
>>>>>>>>>>>>> ISSUE-83 (Vipul): Property Chain Axiom: P1 o P2 => P2 o P1
>>>>>>>>>>>>>
>>>>>>>>>>>>> http://www.w3.org/2007/OWL/tracker/issues/
>>>>>>>>>>>>>
>>>>>>>>>>>>> Raised by: Vipul Kashyap
>>>>>>>>>>>>> On product:
>>>>>>>>>>>>>
>>>>>>>>>>>>> I would like support for the property chain axiom.
>>>>>>>>>>>>> The use case is based on Alan Rector's example in the
>>>>>>>>>>  DL Handbook
>>>>>>>>>>>>>
>>>>>>>>>>>>> Skin of the finger is part of the skin of the hand.
>>>>>>>>>>>>> covers o part --> part o covers
>>>>>>>>>>>>>
>>>>>>>>>>>>> Thanks,
>>>>>>>>>>>>>
>>>>>>>>>>>>> ---Vipul
>>>>>>>>>>>>>
>>>>>>>>>>>>>
>>>>>>>>>>>>>
>>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>> Interestingly, the constructs we already have, put this
>>>>>>>>>>  into the OWL
>>>>>>>>>>>> Full version of the language ...
>>>>>>>>>>>>
>>>>>>>>>>>> Jeremy
>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> The information transmitted in this electronic  
>>>>>>>>>>> communication is
>>>>>>>>>>> intended only
>>>>>>>>>>> for the person or entity to whom it is addressed and may  
>>>>>>>>>>> contain
>>>>>>>>>>> confidential
>>>>>>>>>>> and/or privileged material. Any review, retransmission,
>>>>>>>>>>> dissemination or other
>>>>>>>>>>> use of or taking of any action in reliance upon this
>>>>>>>>>>  information by
>>>>>>>>>>> persons or
>>>>>>>>>>> entities other than the intended recipient is prohibited.  
>>>>>>>>>>> If you
>>>>>>>>>>> received this
>>>>>>>>>>> information in error, please contact the Compliance  
>>>>>>>>>>> HelpLine at
>>>>>>>>>>> 800-856-1983 and
>>>>>>>>>>> properly dispose of this information.
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>
>>>>>> "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
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>
>>>>
>>>> "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
>>>>
>>>>
>>>>
>>>>
>>>
>>
>> "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
>>
>>
>>
>>
>

"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 Thursday, 6 December 2007 13:49:18 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 14:13:29 GMT