Search Mailing List Archives

Limit search to: Subject & Body Subject Author
Sort by: Reverse Sort
Limit to: All This Week Last Week This Month Last Month
Select Date Range     through    

[protege-owl] DL Query: is it possible to get indirect individuals?

Thomas Schneider schneidt at
Thu Jul 8 01:02:48 PDT 2010

A non-text attachment was scrubbed...
Name: indiv_test.owl
Type: application/octet-stream
Size: 2480 bytes
Desc: not available
URL: <>
-------------- next part --------------

I claim that this is a bug in Protégé's GUI since I've created a  
minimal example where the unintended behaviour occurs independently of  
the selected reasoner, and the missing answer occurs at another place:

* The attached ontology contains 2 classes and 2 individuals in the  
following 3 axioms:

   Super subClassOf Sub
   super instanceOf Super
   sub instanceOf super

* If I classify with Hermit 1.2.4, FaCT++ 1.4.0 or Pellet 2.1.0 and  
execute the DL query "Super", then only "super" is shown as an  
instance, not "sub". This is clearly a missing answer.

* When I change to the description of the individual "sub", then only  
the type "Sub" is shown, not "Super". I understand that only selected  
entailments are shown in the Description view. However, when I change  
to the description of "Super", then "sub" is shown as an inferred  
member. Therefore the above bug in the DL query tab is not the  
reasoner's fault -- it rather seems that Protégé doesn't ask the right  
question when processing a DL query, or forgets answers.



On 7 Jul 2010, at 22:38, Thomas Russ wrote:

> On Jul 7, 2010, at 2:04 AM, Mathilde Deffontaines wrote:
>> Hi
>> I am wondering if it is possible to get indirect instances of a  
>> class:
>> So is my ontology
>> Class Poste (individuals p1, p2, p3)
>>   SubClass LittlePoste (individuals p4, p5)
>> System infers that p4 and p5 have type Poste, now when I make DL  
>> query:  "Poste" it gives me p1, p2 and p3 but I would prefer to get  
>> p1, p2, p3, p4 & p5
> I would expect this to happen, although you may need to run a  
> reasoner and classify individuals before hand.  I haven't really  
> worked with DL query, so I'm not up on the details of what you need  
> to do to get the inferred results.
>> and for the DL query: "Poste and not LittlePoste", then only it  
>> would answer p1, p2 & p3.
> It would be unlikely to give you that answer.  I would expect that  
> it wouldn't give you any individuals at all.  That is because the  
> open world semantics means that it is possible that the individuals  
> of class Poste might ALSO be individuals of class LittlePoste, but  
> we just don't know it yet.
> The system will have to prove that (not LittlePoste) is true of the  
> individuals, so it needs to find some reason why individuals p1, p2  
> and p3 CANNOT also be in LittlePoste.  Usually this means that you  
> have to explicitly associate them with a negated class membership or  
> else some other restriction that is negated by the class LittlePoste.
> _______________________________________________
> protege-owl mailing list
> protege-owl at
> Instructions for unsubscribing:

|  Dr Thomas Schneider                    schneider (at)  |
|  School of Computer Science  |
|  Kilburn Building, Room 2.114                 phone +44 161 2756136  |
|  University of Manchester                                            |
|  Oxford Road                                             _///_       |
|  Manchester M13 9PL                                      (o~o)       |

Detchant (n.)
   The part of the hymn (usually a few notes at the end of the verse)
   where the tune goes so high or low that you suddenly have to change
   pitch to accommodate it.

                   Douglas Adams, John Lloyd: The Deeper Meaning of Liff

-------------- next part --------------
A non-text attachment was scrubbed...
Name: PGP.sig
Type: application/pgp-signature
Size: 203 bytes
Desc: This is a digitally signed message part
URL: <>

More information about the protege-owl mailing list