Protege Project › Protege-OWL 3.x Support

Reasoning problem with hasValue restrictions with respect to imported ontologies

Classic

List

Threaded

6 messages Options

brian dorsey

Reasoning problem with hasValue restrictions with respect to imported ontologies

Hi all,

I have a question regarding reasoning over an imported ontology, in particular with the "hasValue" restriction.

I am using Protege 3.4.4 build 579

I notice that a locally defined class using the hasValue restriction does not infer an individual from an imported class as expected.

For example:

Consider the following example,

Ontology A (ontA)
++++++++++++++

Class Definition:
ClassA SUBSET EXISTS hasProp ClassC AND FORALL hasProp ClassC

Assertion:
ClassA(a)
hasProp(a,c)

ClassATest EQUIV hasProp HAS c

When the reasoner is executed, individual "a" of ClassA is inferred to be a member of ClassATest.

This is as expected.

Ontology B (ontB)
++++++++++++++

ClassTestB EQUIV hasProp HAS ontA:c

When the reasoner is executed, I was expecting individual "ontA:a" from ontology A (imported into ontology B) to be inferred as a member of ClassTestB.
This did not happen. Note, ClassTestA and ClassTestB are defined as necessary and sufficient classes (EQUIV).

Is what I am trying to do define in OWL-DL incorrect or is this not currently supported in Protege?

Note, if I state, for example, in ontology B that class ontA:ClassA SUBSET ClassB (another class of Ontology B) and execute the reasoner then individual "ontA:a" is also a member of ClassB as expected. It just seems to not work using the "hasValue" restriction with respect to imported ontologies.

Comments welcome,

regards,
Brian

_______________________________________________
protege-owl mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-owl

Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03

brian dorsey

Re: Reasoning problem with hasValue restrictions with respect to imported ontologies

Hi all,

I see what I have done wrong.

Instead of the following class definition:

ClassTestB EQUIV hasProp HAS c

I should have defined the following:

ClassTestB EQUIV ontA:hasProp HAS c

where the "hasProp" in the local ontology B is replaced by "ontA:hasProp" from ontology A.

I notice from within Protege, one cannot add a super property to ontA:hasProp or an equivalence relation to state my local property is equivalent to the imported property.

Perhaps my original ClassTestB definition would have worked if ontA:hasProp SUBPROP hasProp where hasProp is referring to a local property in Ontology B.

While I cannot define a super property relation on imported properties, I can for imported classes!
Is this a deliberate feature in Protege with regard to super-properties? Note, on importing, I chose the option of import with read-only.

Brian.

_______________________________________________
protege-owl mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-owl

Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03

Timothy Redmond

Re: Reasoning problem with hasValue restrictions with respect to imported ontologies

While I cannot define a super property relation on imported properties, I can for imported classes!
Is this a deliberate feature in Protege with regard to super-properties? Note, on importing, I chose the option of import with read-only.

I didn't replicate this - I was able to drag a property from the imported ontology to go underneath a property from the importing ontology. It may depend on how you try to change the superclass. it is true that Protege 3 has constraints rules about what it believes can be done to a property declared in an imported ontology. These rules are part of legacy of where Protege 3 came from and you won't have these issues with Protege 4.

-Timothy

On 02/15/2011 03:58 AM, brian dorsey wrote:

Hi all,

I see what I have done wrong.

Instead of the following class definition:

ClassTestB EQUIV hasProp HAS c

I should have defined the following:

ClassTestB EQUIV ontA:hasProp HAS c

where the "hasProp" in the local ontology B is replaced by "ontA:hasProp" from ontology A.

I notice from within Protege, one cannot add a super property to ontA:hasProp or an equivalence relation to state my local property is equivalent to the imported property.

Perhaps my original ClassTestB definition would have worked if ontA:hasProp SUBPROP hasProp where hasProp is referring to a local property in Ontology B.

While I cannot define a super property relation on imported properties, I can for imported classes!
Is this a deliberate feature in Protege with regard to super-properties? Note, on importing, I chose the option of import with read-only.

Brian.
_______________________________________________
protege-owl mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-owl

Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03

brian dorsey

Re: Reasoning problem with hasValue restrictions with respect to imported ontologies

Timothy,

On Tue, Feb 15, 2011 at 5:07 PM, Timothy Redmond <[hidden email]> wrote:

While I cannot define a super property relation on imported properties, I can for imported classes!
Is this a deliberate feature in Protege with regard to super-properties? Note, on importing, I chose the option of import with read-only.

I didn't replicate this - I was able to drag a property from the imported ontology to go underneath a property from the importing ontology. It may depend on how you try to change the superclass. it is true that Protege 3 has constraints rules about what it believes can be done to a property declared in an imported ontology. These rules are part of legacy of where Protege 3 came from and you won't have these issues with Protege 4.

I figured out what I have to do in Protege 3. One must make the imported ontology "active" for a period when one wants to say some properties of the imported ontology are subproperties of the ontology doing the importing.

Thanks for your response,
Brian.

Thomas Russ

Re: Reasoning problem with hasValue restrictions with respect to imported ontologies

On Feb 16, 2011, at 12:58 AM, brian dorsey wrote:

> Timothy,
>
> On Tue, Feb 15, 2011 at 5:07 PM, Timothy Redmond <[hidden email]
> > wrote:
>
>>
>> While I cannot define a super property relation on imported
>> properties, I can for imported classes!
>> Is this a deliberate feature in Protege with regard to super-
>> properties? Note, on importing, I chose the option of import with
>> read-only.
>
> I didn't replicate this - I was able to drag a property from the
> imported ontology to go underneath a property from the importing
> ontology. It may depend on how you try to change the superclass. it
> is true that Protege 3 has constraints rules about what it believes
> can be done to a property declared in an imported ontology. These
> rules are part of legacy of where Protege 3 came from and you won't
> have these issues with Protege 4.
>
>
> I figured out what I have to do in Protege 3. One must make the
> imported ontology "active" for a period when one wants to say some
> properties of the imported ontology are subproperties of the
> ontology doing the importing.

The problem with doing that, though, is that the axiom is then placed
in the imported ontology. This might not be where you want that
information to live.

_______________________________________________
protege-owl mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-owl

Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03

brian dorsey

Re: Reasoning problem with hasValue restrictions with respect to imported ontologies

Thomas,

The problem with doing that, though, is that the axiom is then placed in the imported ontology. This might not be where you want that information to live.

I see what you mean. I notice if I open the imported ontology on its own, I see the properties I made sub-properties of the importing ontology have now been removed from this ontology. And instead are only seen in the ontology that done the importing.

I was not able to click and drag properties as Timothy was able to do in the version of Protege I am using. The imported ontology has its properties faded out in appearance. So in order to click and drag I have to "Set Active Ontology" in the Meta tab to be the imported ontology and then revert back to the importing ontology afterwards.

Perhaps I should just refer to the property names in the imported ontology's name space when defining class definitions.

For example, ClassX ontB:hasSomeProp ontB:ClassY rather than

ClassX hasSomeProp ClassY where the following holds:
ontB:ClassX = ClassX,
ontB:Classy = ClassY,
ontB:hasSomeProp = hasSomeProp