Inconsistent classes

classic Classic list List threaded Threaded
13 messages Options
Reply | Threaded
Open this post in threaded view
|

Inconsistent classes

Maria Koutraki
Hello all,

i want to ask what means when a class is inconsistent after the check consistency using pellet reasoner.

The problem is that i have 4 ontologies, i check the consistency by using pellet independently and the ontologies are consistent.
When i import those ontologies to another ontology then all the classes of the 4 ontologies are inconsistent. 

Do you have any idea why i have this problem?

Thank you.
--
Mary

_______________________________________________
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
Reply | Threaded
Open this post in threaded view
|

Re: Inconsistent classes

Joshua TAYLOR
On Fri, Sep 14, 2012 at 6:36 AM, Mary Koutraki <[hidden email]> wrote:
> i want to ask what means when a class is inconsistent after the check
> consistency using pellet reasoner.
>
> The problem is that i have 4 ontologies, i check the consistency by using
> pellet independently and the ontologies are consistent.
> When i import those ontologies to another ontology then all the classes of
> the 4 ontologies are inconsistent.
>
> Do you have any idea why i have this problem?

If a class is inconsistent, it means that it is not possible for there
to be any instances of it. This could happen, for instance, if

InconsistentClass subClassOf likes min 3
InconsistentClass subClassOf likes max 2

since any member of InconsistentClass would have to like at least 3
things, but can like at most 2 things.  So there can be no members of
this class.  (It's interesting to note that this isn't an
inconsistency in the traditional logical sense (unless some individual
were asserted or inferred to be a member of this class).  However,
it's typically a modeling error if some class can't have instances.)

It could be very easy to have ontologies that are independently
consistent but whose combination is inconsistent.  In some cases you
can view the explanations generated by reasoners in Protege.  Maybe
that will help to explain the problem in your case?

//JT
--
Joshua Taylor, http://www.cs.rpi.edu/~tayloj/
_______________________________________________
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
Reply | Threaded
Open this post in threaded view
|

Re: Inconsistent classes

Maria Koutraki
When i run the reasoner i receive this message: 

Errors at synchronization: edu.standford.smi.protegex.owl.model.impl.DefaultRDFSNamedClass cannot be cast to  edu.standford.smi.protegex.owl.model.OWLClass. 

If it helps. 

Thank you again.

On Fri, Sep 14, 2012 at 4:19 PM, Joshua TAYLOR <[hidden email]> wrote:
On Fri, Sep 14, 2012 at 6:36 AM, Mary Koutraki <[hidden email]> wrote:
> i want to ask what means when a class is inconsistent after the check
> consistency using pellet reasoner.
>
> The problem is that i have 4 ontologies, i check the consistency by using
> pellet independently and the ontologies are consistent.
> When i import those ontologies to another ontology then all the classes of
> the 4 ontologies are inconsistent.
>
> Do you have any idea why i have this problem?

If a class is inconsistent, it means that it is not possible for there
to be any instances of it. This could happen, for instance, if

InconsistentClass subClassOf likes min 3
InconsistentClass subClassOf likes max 2

since any member of InconsistentClass would have to like at least 3
things, but can like at most 2 things.  So there can be no members of
this class.  (It's interesting to note that this isn't an
inconsistency in the traditional logical sense (unless some individual
were asserted or inferred to be a member of this class).  However,
it's typically a modeling error if some class can't have instances.)

It could be very easy to have ontologies that are independently
consistent but whose combination is inconsistent.  In some cases you
can view the explanations generated by reasoners in Protege.  Maybe
that will help to explain the problem in your case?

//JT
--
Joshua Taylor, http://www.cs.rpi.edu/~tayloj/
_______________________________________________
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



--
Mary Koutraki

_______________________________________________
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
Reply | Threaded
Open this post in threaded view
|

Re: Inconsistent classes

Joshua TAYLOR
On Fri, Sep 14, 2012 at 9:39 AM, Mary Koutraki <[hidden email]> wrote:
> When i run the reasoner i receive this message:
>
> Errors at synchronization:
> edu.standford.smi.protegex.owl.model.impl.DefaultRDFSNamedClass cannot be
> cast to  edu.standford.smi.protegex.owl.model.OWLClass.

I'm not a Protege developer, so I don't know what's causing that, but
this raises two more questions:

1) Is that the entire output of the reasoner?  You mentioned the
Pellet reasoner;  you might get more descriptive results if you try
the same queries using the Pellet command line tools (avoiding the
Protege intermediary).  If you can post your ontologies, someone might
be able to do the same.
2) Based on this output, what indication is there that the new
ontology that imports the other four is inconsistent?

//JT

--
Joshua Taylor, http://www.cs.rpi.edu/~tayloj/
_______________________________________________
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
Reply | Threaded
Open this post in threaded view
|

Re: Inconsistent classes

Robert A. Elliott, Sr.
In reply to this post by Maria Koutraki
Mary,
I have run across this problem before. Here is what caused my ontology to become inconsistent.

In some of my data-type property definitions, I removed an entry from the "allowed values" list after I had already created an individual using that property. 

Running the consistency check placed the ontology in an inconsistent state.

Placing the entry back into the "allowed values" list will clear this up. 

Alternatively, you can delete the individual(s) that contain values no longer in "allowed values" list for the offending data-type property(s).

HTH,
Robert

On Fri, Sep 14, 2012 at 5:36 AM, Mary Koutraki <[hidden email]> wrote:
Hello all,

i want to ask what means when a class is inconsistent after the check consistency using pellet reasoner.

The problem is that i have 4 ontologies, i check the consistency by using pellet independently and the ontologies are consistent.
When i import those ontologies to another ontology then all the classes of the 4 ontologies are inconsistent. 

Do you have any idea why i have this problem?

Thank you.
--
Mary

_______________________________________________
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




--
Robert A. Elliott, Sr.
Ph.D Candidate
Department of Computer Science & Engineering
Mississippi State University
[hidden email]

_______________________________________________
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
Reply | Threaded
Open this post in threaded view
|

Re: Inconsistent classes

Timothy Redmond
In reply to this post by Maria Koutraki

I would recommend that if you are interested in inference then you should use Protege 4.  The reasoning capabilities of Protege 4 are far superior to that in Protege 3.

The exception that you are getting suggests the possibility that your ontology is OWL full, which is also non-optimal for OWL reasoners.  You could try an OWL validator.  Use [1] if you need the ontology to be OWL 1 or use [2] if it is ok for the ontology to be in OWL 2.

But it could also be that you have an OWL 2 ontology that you are reading with an OWL 1 tool.  It would be hard to know without seeing the ontology.

-Timothy


[1] http://www.mygrid.org.uk/OWL/Validator
[2] http://owl.cs.manchester.ac.uk/validator/


On 9/14/12 6:39 AM, Mary Koutraki wrote:
When i run the reasoner i receive this message: 

Errors at synchronization: edu.standford.smi.protegex.owl.model.impl.DefaultRDFSNamedClass cannot be cast to  edu.standford.smi.protegex.owl.model.OWLClass. 

If it helps. 

Thank you again.

On Fri, Sep 14, 2012 at 4:19 PM, Joshua TAYLOR <[hidden email]> wrote:
On Fri, Sep 14, 2012 at 6:36 AM, Mary Koutraki <[hidden email]> wrote:
> i want to ask what means when a class is inconsistent after the check
> consistency using pellet reasoner.
>
> The problem is that i have 4 ontologies, i check the consistency by using
> pellet independently and the ontologies are consistent.
> When i import those ontologies to another ontology then all the classes of
> the 4 ontologies are inconsistent.
>
> Do you have any idea why i have this problem?

If a class is inconsistent, it means that it is not possible for there
to be any instances of it. This could happen, for instance, if

InconsistentClass subClassOf likes min 3
InconsistentClass subClassOf likes max 2

since any member of InconsistentClass would have to like at least 3
things, but can like at most 2 things.  So there can be no members of
this class.  (It's interesting to note that this isn't an
inconsistency in the traditional logical sense (unless some individual
were asserted or inferred to be a member of this class).  However,
it's typically a modeling error if some class can't have instances.)

It could be very easy to have ontologies that are independently
consistent but whose combination is inconsistent.  In some cases you
can view the explanations generated by reasoners in Protege.  Maybe
that will help to explain the problem in your case?

//JT
--
Joshua Taylor, http://www.cs.rpi.edu/~tayloj/
_______________________________________________
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



--
Mary Koutraki


_______________________________________________
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


_______________________________________________
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
Reply | Threaded
Open this post in threaded view
|

Re: Inconsistent classes

Alan Ruttenberg-2
In reply to this post by Joshua TAYLOR


On Fri, Sep 14, 2012 at 3:19 PM, Joshua TAYLOR <[hidden email]> wrote:
On Fri, Sep 14, 2012 at 6:36 AM, Mary Koutraki <[hidden email]> wrote:
> i want to ask what means when a class is inconsistent after the check
> consistency using pellet reasoner.
>
> The problem is that i have 4 ontologies, i check the consistency by using
> pellet independently and the ontologies are consistent.
> When i import those ontologies to another ontology then all the classes of
> the 4 ontologies are inconsistent.
>
> Do you have any idea why i have this problem?

If a class is inconsistent, it means that it is not possible for there
to be any instances of it. This could happen, for instance, if

This is called an unsatisfiable class. An inconsistency is a contradiction, such as when you say some instance is of a certain type and also the complement of that type. In protege such classes are colored red. For inconsistencies, older versions of protege 4 throw an exception. In the latest version of protege 4 a window pops up telling you of the inconsistency and offering to help debug it.

-Alan



InconsistentClass subClassOf likes min 3
InconsistentClass subClassOf likes max 2

since any member of InconsistentClass would have to like at least 3
things, but can like at most 2 things.  So there can be no members of
this class.  (It's interesting to note that this isn't an
inconsistency in the traditional logical sense (unless some individual
were asserted or inferred to be a member of this class).  However,
it's typically a modeling error if some class can't have instances.)

It could be very easy to have ontologies that are independently
consistent but whose combination is inconsistent.  In some cases you
can view the explanations generated by reasoners in Protege.  Maybe
that will help to explain the problem in your case?

//JT
--
Joshua Taylor, http://www.cs.rpi.edu/~tayloj/
_______________________________________________
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


_______________________________________________
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
Reply | Threaded
Open this post in threaded view
|

Re: Inconsistent classes

Joshua TAYLOR
On Sat, Sep 15, 2012 at 7:07 AM, Alan Ruttenberg
<[hidden email]> wrote:

>
>
> On Fri, Sep 14, 2012 at 3:19 PM, Joshua TAYLOR <[hidden email]>
> wrote:
>>
>> On Fri, Sep 14, 2012 at 6:36 AM, Mary Koutraki <[hidden email]>
>> wrote:
>> > i want to ask what means when a class is inconsistent after the check
>> > consistency using pellet reasoner.
>> >
>> > The problem is that i have 4 ontologies, i check the consistency by
>> > using
>> > pellet independently and the ontologies are consistent.
>> > When i import those ontologies to another ontology then all the classes
>> > of
>> > the 4 ontologies are inconsistent.
>> >
>> > Do you have any idea why i have this problem?
>>
>> If a class is inconsistent, it means that it is not possible for there
>> to be any instances of it. This could happen, for instance, if
>
>
> This is called an unsatisfiable class. An inconsistency is a contradiction,
> such as when you say some instance is of a certain type and also the
> complement of that type. In protege such classes are colored red. For
> inconsistencies, older versions of protege 4 throw an exception. In the
> latest version of protege 4 a window pops up telling you of the
> inconsistency and offering to help debug it.

Thanks for the correction here.  I was aware of the difference between
inconsistency (which is a property of an ontology) and
unsatisfiability (which is a property of a class) (and even noted it
farther down in the message), but for some reason had it in my mind
that the DL community (and by extension, OWL folks) used the terms in
a somewhat non-traditional way.

I think I'd been thinking of the usage of the terms in traditional
logics.  (Un)satisfiable and (in)consistent are both properties of a
formula or a set of formulae;  the former meaning that there's no
model of the formula (or set), the latter meaning that an
inconsistency can be derived.  But there's no (to my knowledge)
traditional way to say that a unary predicate P has no individual x
such that P(x).  In Description Logics, (un)satisfiable is a property
of a class/concept, and means exactly that;  there is (no) individual
x such that x is a P.

Sorry for the mis-information!

Also, a note about the Protege behavior.  I'm running 4.2.0 (b 256)
and just create an ontology with a class Foo, an object property
frobs, and the axioms

  Foo subClassOf frobs max 2
  Foo subClassOf frobs min 3

and ran Pellet.  I get a result that owl:Nothing is an equivalent
class, and the explanation offered by the "?" button shows the two
problematic axioms, but I neither got an exception nor was prompted
with a debug window.

>> InconsistentClass subClassOf likes min 3
>> InconsistentClass subClassOf likes max 2
>>
>> since any member of InconsistentClass would have to like at least 3
>> things, but can like at most 2 things.  So there can be no members of
>> this class.  (It's interesting to note that this isn't an
>> inconsistency in the traditional logical sense (unless some individual
>> were asserted or inferred to be a member of this class).  However,
>> it's typically a modeling error if some class can't have instances.)
>>
>> It could be very easy to have ontologies that are independently
>> consistent but whose combination is inconsistent.  In some cases you
>> can view the explanations generated by reasoners in Protege.  Maybe
>> that will help to explain the problem in your case?
>>
>> //JT
>> --
>> Joshua Taylor, http://www.cs.rpi.edu/~tayloj/
>> _______________________________________________
>> 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
>
>



--
Joshua Taylor, http://www.cs.rpi.edu/~tayloj/
_______________________________________________
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
Reply | Threaded
Open this post in threaded view
|

Re: Inconsistent classes

Joshua TAYLOR
On Sat, Sep 15, 2012 at 9:55 AM, Alan Ruttenberg
<[hidden email]> wrote:

>
> On Sep 15, 2012, at 3:19 PM, Joshua TAYLOR <[hidden email]> wrote:
>
>> On Sat, Sep 15, 2012 at 7:07 AM, Alan Ruttenberg
>> <[hidden email]> wrote:
>>>
>>>
>>> On Fri, Sep 14, 2012 at 3:19 PM, Joshua TAYLOR <[hidden email]>
>>> wrote:
>>>>
>>>> On Fri, Sep 14, 2012 at 6:36 AM, Mary Koutraki <[hidden email]>
>>>> wrote:
>>>>> i want to ask what means when a class is inconsistent after the check
>>>>> consistency using pellet reasoner.
>>>>>
>>>>> The problem is that i have 4 ontologies, i check the consistency by
>>>>> using
>>>>> pellet independently and the ontologies are consistent.
>>>>> When i import those ontologies to another ontology then all the classes
>>>>> of
>>>>> the 4 ontologies are inconsistent.
>>>>>
>>>>> Do you have any idea why i have this problem?
>>>>
>>>> If a class is inconsistent, it means that it is not possible for there
>>>> to be any instances of it. This could happen, for instance, if
>>>
>>>
>>> This is called an unsatisfiable class. An inconsistency is a contradiction,
>>> such as when you say some instance is of a certain type and also the
>>> complement of that type. In protege such classes are colored red. For
>>> inconsistencies, older versions of protege 4 throw an exception. In the
>>> latest version of protege 4 a window pops up telling you of the
>>> inconsistency and offering to help debug it.
>>
>> Thanks for the correction here.  I was aware of the difference between
>> inconsistency (which is a property of an ontology) and
>> unsatisfiability (which is a property of a class) (and even noted it
>> farther down in the message), but for some reason had it in my mind
>> that the DL community (and by extension, OWL folks) used the terms in
>> a somewhat non-traditional way.
>>
>> I think I'd been thinking of the usage of the terms in traditional
>> logics.  (Un)satisfiable and (in)consistent are both properties of a
>> formula or a set of formulae;  the former meaning that there's no
>> model of the formula (or set), the latter meaning that an
>> inconsistency can be derived.  But there's no (to my knowledge)
>> traditional way to say that a unary predicate P has no individual x
>> such that P(x).  In Description Logics, (un)satisfiable is a property
>> of a class/concept, and means exactly that;  there is (no) individual
>> x such that x is a P.
>>
>> Sorry for the mis-information!
>>
>> Also, a note about the Protege behavior.  I'm running 4.2.0 (b 256)
>> and just create an ontology with a class Foo, an object property
>> frobs, and the axioms
>>
>>  Foo subClassOf frobs max 2
>>  Foo subClassOf frobs min 3
>>
>> and ran Pellet.  I get a result that owl:Nothing is an equivalent
>> class, and the explanation offered by the "?" button shows the two
>> problematic axioms, but I neither got an exception nor was prompted
>> with a debug window.
>
> To create an inconsistency add an individual and make its superclass Foo

D'oh!  Yes, of course.  Because, as we just discussed, class
unsatisfiability is *not the same* as inconsistency, and you made the
point that Protégé prompts on inconsistency, not unsatisfiability.  I
really should wait to respond to things until I've been awake for a
few hours.

... goes to test in Protege ...

Ah, here's an interesting point. I added the individual to the
unsatisfiable class.  When I run Pellet, I just get a dialog box
telling me that there's an inconsistency;  I don't get the fancy debug
window.  When I run Hermit, though, I get the nice "let's debug your
ontology..." window and the nice interface to see the problems.  It
has three tabs:

* ""Bad" individuals"
* Inconsistent Classes
* Hot spots

The individual I created is obviously a "bad" individual.  However, my
Foo is listed as an inconsistent class (though there is note "These
classes can be proved inconsistent but this may not be a reason why
the ontology is inconsistent.").  If there's no problematic
individual, then running Hermit doesn't cause the ontology debugging
window to appear (and probably rightfully so, since the ontology isn't
inconsistent).

If the distinction between "inconsistent" (applied to an ontology) and
"unsatisfiable" (applied to classes) is to be maintained, the ontology
debugging window should probably refer to these classes as
"unsatisfiable" rather than "inconsistent".

//JT

--
Joshua Taylor, http://www.cs.rpi.edu/~tayloj/
_______________________________________________
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
Reply | Threaded
Open this post in threaded view
|

Re: Inconsistent classes

Timothy Redmond

> Ah, here's an interesting point. I added the individual to the
> unsatisfiable class.  When I run Pellet, I just get a dialog box
> telling me that there's an inconsistency;  I don't get the fancy debug
> window.  When I run Hermit, though, I get the nice "let's debug your
> ontology..." window and the nice interface to see the problems.  It
> has three tabs:
>
> * ""Bad" individuals"
> * Inconsistent Classes
> * Hot spots

One of the good things that happened in Protege 4.2 was that we added
Matthew's explanation tool.  In particular the explanation of
inconsistent ontologies has become infinitely better.  The explanation
that you are looking is a major hack and was only included because at
the time there was nothing better.  It has since been completely removed
from the sources.  If you are working with ontologies that become
inconsistent then I would recommend Protege 4.2.  (Actually all
explanation is much better in 4.2.)

> If the distinction between "inconsistent" (applied to an ontology) and
> "unsatisfiable" (applied to classes) is to be maintained, the ontology
> debugging window should probably refer to these classes as
> "unsatisfiable" rather than "inconsistent".

Agreed.  Fortunately this code is no longer necessary.

-Timothy



On 09/15/2012 09:51 AM, Joshua TAYLOR wrote:

> On Sat, Sep 15, 2012 at 9:55 AM, Alan Ruttenberg
> <[hidden email]> wrote:
>> On Sep 15, 2012, at 3:19 PM, Joshua TAYLOR <[hidden email]> wrote:
>>
>>> On Sat, Sep 15, 2012 at 7:07 AM, Alan Ruttenberg
>>> <[hidden email]> wrote:
>>>>
>>>> On Fri, Sep 14, 2012 at 3:19 PM, Joshua TAYLOR <[hidden email]>
>>>> wrote:
>>>>> On Fri, Sep 14, 2012 at 6:36 AM, Mary Koutraki <[hidden email]>
>>>>> wrote:
>>>>>> i want to ask what means when a class is inconsistent after the check
>>>>>> consistency using pellet reasoner.
>>>>>>
>>>>>> The problem is that i have 4 ontologies, i check the consistency by
>>>>>> using
>>>>>> pellet independently and the ontologies are consistent.
>>>>>> When i import those ontologies to another ontology then all the classes
>>>>>> of
>>>>>> the 4 ontologies are inconsistent.
>>>>>>
>>>>>> Do you have any idea why i have this problem?
>>>>> If a class is inconsistent, it means that it is not possible for there
>>>>> to be any instances of it. This could happen, for instance, if
>>>>
>>>> This is called an unsatisfiable class. An inconsistency is a contradiction,
>>>> such as when you say some instance is of a certain type and also the
>>>> complement of that type. In protege such classes are colored red. For
>>>> inconsistencies, older versions of protege 4 throw an exception. In the
>>>> latest version of protege 4 a window pops up telling you of the
>>>> inconsistency and offering to help debug it.
>>> Thanks for the correction here.  I was aware of the difference between
>>> inconsistency (which is a property of an ontology) and
>>> unsatisfiability (which is a property of a class) (and even noted it
>>> farther down in the message), but for some reason had it in my mind
>>> that the DL community (and by extension, OWL folks) used the terms in
>>> a somewhat non-traditional way.
>>>
>>> I think I'd been thinking of the usage of the terms in traditional
>>> logics.  (Un)satisfiable and (in)consistent are both properties of a
>>> formula or a set of formulae;  the former meaning that there's no
>>> model of the formula (or set), the latter meaning that an
>>> inconsistency can be derived.  But there's no (to my knowledge)
>>> traditional way to say that a unary predicate P has no individual x
>>> such that P(x).  In Description Logics, (un)satisfiable is a property
>>> of a class/concept, and means exactly that;  there is (no) individual
>>> x such that x is a P.
>>>
>>> Sorry for the mis-information!
>>>
>>> Also, a note about the Protege behavior.  I'm running 4.2.0 (b 256)
>>> and just create an ontology with a class Foo, an object property
>>> frobs, and the axioms
>>>
>>>   Foo subClassOf frobs max 2
>>>   Foo subClassOf frobs min 3
>>>
>>> and ran Pellet.  I get a result that owl:Nothing is an equivalent
>>> class, and the explanation offered by the "?" button shows the two
>>> problematic axioms, but I neither got an exception nor was prompted
>>> with a debug window.
>> To create an inconsistency add an individual and make its superclass Foo
> D'oh!  Yes, of course.  Because, as we just discussed, class
> unsatisfiability is *not the same* as inconsistency, and you made the
> point that Protégé prompts on inconsistency, not unsatisfiability.  I
> really should wait to respond to things until I've been awake for a
> few hours.
>
> ... goes to test in Protege ...
>
> Ah, here's an interesting point. I added the individual to the
> unsatisfiable class.  When I run Pellet, I just get a dialog box
> telling me that there's an inconsistency;  I don't get the fancy debug
> window.  When I run Hermit, though, I get the nice "let's debug your
> ontology..." window and the nice interface to see the problems.  It
> has three tabs:
>
> * ""Bad" individuals"
> * Inconsistent Classes
> * Hot spots
>
> The individual I created is obviously a "bad" individual.  However, my
> Foo is listed as an inconsistent class (though there is note "These
> classes can be proved inconsistent but this may not be a reason why
> the ontology is inconsistent.").  If there's no problematic
> individual, then running Hermit doesn't cause the ontology debugging
> window to appear (and probably rightfully so, since the ontology isn't
> inconsistent).
>
> If the distinction between "inconsistent" (applied to an ontology) and
> "unsatisfiable" (applied to classes) is to be maintained, the ontology
> debugging window should probably refer to these classes as
> "unsatisfiable" rather than "inconsistent".
>
> //JT
>

_______________________________________________
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
Reply | Threaded
Open this post in threaded view
|

Re: Inconsistent classes

Joshua TAYLOR
On Tue, Sep 18, 2012 at 1:00 PM, Timothy Redmond <[hidden email]> wrote:

>
>> Ah, here's an interesting point. I added the individual to the
>> unsatisfiable class.  When I run Pellet, I just get a dialog box
>> telling me that there's an inconsistency;  I don't get the fancy debug
>> window.  When I run Hermit, though, I get the nice "let's debug your
>> ontology..." window and the nice interface to see the problems.  It
>> has three tabs:
>>
>> * ""Bad" individuals"
>> * Inconsistent Classes
>> * Hot spots
>
> One of the good things that happened in Protege 4.2 was that we added
> Matthew's explanation tool.  In particular the explanation of inconsistent
> ontologies has become infinitely better.  The explanation that you are
> looking is a major hack and was only included because at the time there was
> nothing better.  It has since been completely removed from the sources.  If
> you are working with ontologies that become inconsistent then I would
> recommend Protege 4.2.  (Actually all explanation is much better in 4.2.)

Actually, this was in Protege 4.2 (build 256)...

--
Joshua Taylor, http://www.cs.rpi.edu/~tayloj/
_______________________________________________
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
Reply | Threaded
Open this post in threaded view
|

Re: Inconsistent classes

Timothy Redmond

> Actually, this was in Protege 4.2 (build 256)...

Yes - Matthew's explanation tool hadn't made it in by that version of
Protege 4.2.  A later version of Protege 4.2 included both explanation
tools and the latest version includes Matthew's tool only.

-Timothy


On 09/18/2012 10:18 AM, Joshua TAYLOR wrote:

> On Tue, Sep 18, 2012 at 1:00 PM, Timothy Redmond <[hidden email]> wrote:
>>> Ah, here's an interesting point. I added the individual to the
>>> unsatisfiable class.  When I run Pellet, I just get a dialog box
>>> telling me that there's an inconsistency;  I don't get the fancy debug
>>> window.  When I run Hermit, though, I get the nice "let's debug your
>>> ontology..." window and the nice interface to see the problems.  It
>>> has three tabs:
>>>
>>> * ""Bad" individuals"
>>> * Inconsistent Classes
>>> * Hot spots
>> One of the good things that happened in Protege 4.2 was that we added
>> Matthew's explanation tool.  In particular the explanation of inconsistent
>> ontologies has become infinitely better.  The explanation that you are
>> looking is a major hack and was only included because at the time there was
>> nothing better.  It has since been completely removed from the sources.  If
>> you are working with ontologies that become inconsistent then I would
>> recommend Protege 4.2.  (Actually all explanation is much better in 4.2.)
> Actually, this was in Protege 4.2 (build 256)...
>

_______________________________________________
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
Reply | Threaded
Open this post in threaded view
|

Re: Inconsistent classes

Joshua TAYLOR
On Wed, Sep 19, 2012 at 11:28 AM, Timothy Redmond <[hidden email]> wrote:
>
>> Actually, this was in Protege 4.2 (build 256)...
>
>
> Yes - Matthew's explanation tool hadn't made it in by that version of
> Protege 4.2.  A later version of Protege 4.2 included both explanation tools
> and the latest version includes Matthew's tool only.

Well, then, it looks like I'll try to update soon!  Thanks for the
clarification!

--
Joshua Taylor, http://www.cs.rpi.edu/~tayloj/
_______________________________________________
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