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] Error with DIG interface

Esteban Zimanyi ezimanyi at ulb.ac.be
Wed Jan 30 09:28:09 PST 2008


Dear all

I have a problem with the enclosed ontology.
When computing the inferred types using
the RacerPro Version 1.9.0 2005-12-05!
it states that the ABox is incoherent, 
which is obviously incorrect.

Having asked the people from Racer
I received the answer below,
which states that the error comes
from Protege.

Is there a place where bugs found
in Protege are kept to be solved
in the future ?

Regards

Esteban Zimanyi

--------------------------------------------------------------------------
Prof. Esteban Zimanyi
Department of Computer & Decision Engineering  (CoDE) 
Faculty of Applied Sciences CP 165/15    tel: + 32.2.650.31.85
Universite Libre de Bruxelles            fax: + 32.2.650.47.13
Avenue F. D. Roosevelt 50                e-mail: ezimanyi at ulb.ac.be
B-1050 Brussels, Belgium                 Internet: http://code.ulb.ac.be/
--------------------------------------------------------------------------

----- Original Message ----- 
From: "Racer Systems Support" <support at racer-systems.com>
To: "Esteban Zimanyi" <ezimanyi at ulb.ac.be>
Sent: Tuesday, January 29, 2008 12:39 PM
Subject: Re: Incoherent ABox Error


On Tuesday 22 January 2008 21:24, you wrote:
> Dear Sirs

Dear Prof. Zimanyi, 

> I am using the RacerPro Version 1.9.0 2005-12-05!
>
> I have a problem with the enclosed ontology.
> When computing the inferred types it states that
> the ABox is incoherent.
>
> I think this is an error. Could you please help me
> to solve this issue ?

I have looked again at your OWL file and found the reason for the 
problem. Indeed, the inconsistency is not caused by Racer, but by 
Protege, sending an erroneous DIG KB to Racer. The following is sent 
by Protege via DIG: 


<tells uri="http://dl.kr.org/dig/kb-1849" 
xmlns="http://dl.kr.org/dig/2003/02/lang">
    <defconcept name="vmap1_enum"/>
    <equalc>
        <catom name="vmap1_enum"/>
        <and>
            <iset>
                <individual name="a"/>
                <individual name="b"/>
            </iset>
        </and>
    </equalc>
    <defconcept name="aml_enum"/>
    <equalc>
        <catom name="aml_enum"/>
        <and>
            <iset>
                <individual name="x"/>
                <individual name="y"/>
            </iset>
        </and>
    </equalc>
    <defindividual name="y"/>
    <instanceof>
        <individual name="y"/>
        <catom name="aml_enum"/>
    </instanceof>
    <defindividual name="a"/>
    <instanceof>
        <individual name="a"/>
        <catom name="vmap1_enum"/>
    </instanceof>
    <sameas>
        <individual name="a"/>
        <individual name="x"/>
    </sameas>
    <defindividual name="x"/>
    <instanceof>
        <individual name="x"/>
        <catom name="aml_enum"/>
    </instanceof>
    <defindividual name="b"/>
    <instanceof>
        <individual name="b"/>
        <catom name="vmap1_enum"/>
    </instanceof>
    <sameas>
        <individual name="b"/>
        <individual name="y"/>
    </sameas>
    <alldifferent>
        <individual name="a"/>
        <individual name="b"/>
        <individual name="x"/>
        <individual name="y"/>
    </alldifferent>
</tells>

So, this states that a,b,x,y are all different, which obviously 
contradicts 

  <sameas>
        <individual name="b"/>
        <individual name="y"/>
   </sameas>

etc. 

However, in your KB, the following is stated: 

  <owl:AllDifferent>
    <owl:distinctMembers rdf:parseType="Collection">
      <vmap1_enum rdf:about="#a"/>
      <vmap1_enum rdf:about="#b"/>
    </owl:distinctMembers>
  </owl:AllDifferent>

  <owl:AllDifferent>
    <owl:distinctMembers rdf:parseType="Collection">
      <aml_enum rdf:about="#x"/>
      <aml_enum rdf:about="#y"/>
    </owl:distinctMembers>
  </owl:AllDifferent>


Thus, 

 <alldifferent>
        <individual name="a"/>
        <individual name="b"/>  
 </alldifferent>

and 

 <alldifferent>
        <individual name="x"/>
        <individual name="y"/>  
 </alldifferent>

should have been sent instead... 

To bypass the problem, please use RacerPorter and owl-read-file. The 
KB is consistent then, but you will not get the inference that 
aml_enum and vmap1_enum are equivalent, since nominals are currently 
approximated in Racer. 


Best regards and thank you, 

Michael 

-- 
Racer Systems GmbH & Co. KG
Blumenau 50 | Tel.: +49 40/88306820
22089 Hamburg | Fax: +49 40/88306821
Germany | www.racer-systems.com


-------------- next part --------------
A non-text attachment was scrubbed...
Name: inequaltest.owl
Type: application/octet-stream
Size: 1538 bytes
Desc: not available
URL: <http://mailman.stanford.edu/pipermail/protege-owl/attachments/20080130/63a385ec/attachment.owl>


More information about the protege-owl mailing list