OWL vs Frames: Semantic relations of the same kind between differend classes

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

OWL vs Frames: Semantic relations of the same kind between differend classes

Андрей Михайлюк
Hi all.
In Protege Frames I could have two semantic relations of the same kind between differend classes using slots:
OntologyOS:
Slot "Vendor" with domain "Windows","MacOS" and allowed supperclasses "Company"
Class "Company" with subclasses "Microsoft" and "Apple"
Class "Windows" with slot "Vendor" restricted to allowed supperclasses "Microsoft"
Class "MacOS" with slot "Vendor" restricted to allowed supperclasses "Apple"

I thought it is valid ontology. But OWL cannot contain two objectProperties with the same name "Vendor" and differend domains/ranges. I mean, I cannot define two triples "Windows"-"Vendor"-"Microsoft" and "MacOS"-"Vendor"-"Apple". Is it true? When I create OntologyOS in CLIPS format - everythig is ok, when I convert it to OWL - everything is bad. Is it the OWL shortcoming or there is a way to define such information in OWL?

Thanks
_______________________________________________
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: OWL vs Frames: Semantic relations of the same kind between differend classes

Timothy Redmond

There are several issues mixed in this question.  I suspect that the short answer is that the problem you are seeing is with the frames to OWL translation and that OWL is plenty expressive enough to say what you want.

One of the issues in the question is the distinction between classes and individuals. When you suggest that the Vendor slot/property has domain Windows it sounds like Windows is going to be a class.  But then you talk about the triple Windows-Vendor-Microsoft, it sounds like Windows and Microsoft have either become individuals or the Vendor property is now a map between classes of classes.

So I will choose an interpretation and try to express what you are saying in OWL.  It seems reasonable that Widows and MacOS are classes - we can very reasonably put several different operating systems in both categories.  I would have been tempted to make Microsoft and Apple be individuals of type Company but I know this is a tricky area.  To be consistent with your example we can make them classes with the idea that perhaps they each only contain one individual.  And finally Vendor will be an object property that takes products and returns the vendor of that product.  For simplicity I will assume that the Vendor property is functional.

So one of the things that you might be wanting to say is

Any member of the Windows class is an operating system whose vendor is some company in the Microsoft class.

This can easily be expressed in OWL as

Class: <http://protege.org/ontologies/Vendors.owl#Windows>
    SubClassOf: 
        <http://protege.org/ontologies/Vendors.owl#Vendor> some <http://protege.org/ontologies/Vendors.owl#Microsoft>


-Timothy

On 05/15/2011 01:10 PM, Андрей Михайлюк wrote:
Hi all.
In Protege Frames I could have two semantic relations of the same kind between differend classes using slots:
OntologyOS:
Slot "Vendor" with domain "Windows","MacOS" and allowed supperclasses "Company"
Class "Company" with subclasses "Microsoft" and "Apple"
Class "Windows" with slot "Vendor" restricted to allowed supperclasses "Microsoft"
Class "MacOS" with slot "Vendor" restricted to allowed supperclasses "Apple"

I thought it is valid ontology. But OWL cannot contain two objectProperties with the same name "Vendor" and differend domains/ranges. I mean, I cannot define two triples "Windows"-"Vendor"-"Microsoft" and "MacOS"-"Vendor"-"Apple". Is it true? When I create OntologyOS in CLIPS format - everythig is ok, when I convert it to OWL - everything is bad. Is it the OWL shortcoming or there is a way to define such information in OWL?

Thanks
_______________________________________________
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

Vendors.owl (4K) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: OWL vs Frames: Semantic relations of the same kind between differend classes

Андрей Михайлюк
Thanks for the answer.

I supposed something like that, but I believed there is a simpler way.
Actualy, I am developing some ontology tool using OWL API. The one should have OWL as the ontology format but algorithms should be expressed in <Class-Relation> model - not the OWL. So I tried to find a correspondence between OWL and <Class-Relation> models. Now I see that I have to parse SubClassOf expressions to find out is this an inheritance between two classes or an expression of class-to-class relation.

Regards,
Andrii

--- Исходное сообщение ---
 От кого: "Timothy Redmond" <[hidden email]>
 Кому: [hidden email]
 Дата: 16 мая 2011, 08:26:46
 Тема: Re: [protege-owl] OWL vs Frames: Semantic relations of the same kind between differend classes
 


> There are several issues mixed in this question.  I suspect that the short answer is that the problem you are seeing is with the frames to OWL translation and that OWL is plenty expressive enough to say what you want.
>  
>  One of the issues in the question is the distinction between classes and individuals. When you suggest that the Vendor slot/property has domain Windows it sounds like Windows is going to be a class.  But then you talk about the triple Windows-Vendor-Microsoft, it sounds like Windows and Microsoft have either become individuals or the Vendor property is now a map between classes of classes.
>  
>  So I will choose an interpretation and try to express what you are saying in OWL.  It seems reasonable that Widows and MacOS are classes - we can very reasonably put several different operating systems in both categories.  I would have been tempted to make Microsoft and Apple be individuals of type Company but I know this is a tricky area.  To be consistent with your example we can make them classes with the idea that perhaps they each only contain one individual.  And finally Vendor will be an object property that takes products and returns the vendor of that product.  For simplicity I will assume that the Vendor property is functional.
>  
>  So one of the things that you might be wanting to say is
>  
>  >
> > Any member of the Windows class is an operating system whose vendor is some company in the Microsoft class.
>  
>  This can easily be expressed in OWL as
>  
>  Class: <http://protege.org/ontologies/Vendors.owl#Windows>
>  SubClassOf:
>  <http://protege.org/ontologies/Vendors.owl#Vendor> some <http://protege.org/ontologies/Vendors.owl#Microsoft>
>
>  
>  -Timothy
>  
>  On 05/15/2011 01:10 PM, Андрей Михайлюк wrote: >
> > Hi all.
> > In Protege Frames I could have two semantic relations of the same kind between differend classes using slots:
> > OntologyOS:
> > Slot "Vendor" with domain "Windows","MacOS" and allowed supperclasses "Company"
> > Class "Company" with subclasses "Microsoft" and "Apple"
> > Class "Windows" with slot "Vendor" restricted to allowed supperclasses "Microsoft"
> > Class "MacOS" with slot "Vendor" restricted to allowed supperclasses "Apple"
> >
> > I thought it is valid ontology. But OWL cannot contain two objectProperties with the same name "Vendor" and differend domains/ranges. I mean, I cannot define two triples "Windows"-"Vendor"-"Microsoft" and "MacOS"-"Vendor"-"Apple". Is it true? When I create OntologyOS in CLIPS format - everythig is ok, when I convert it to OWL - everything is bad. Is it the OWL shortcoming or there is a way to define such information in OWL?
> >
> > Thanks
> > _______________________________________________
> > 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
>
>
_______________________________________________
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: OWL vs Frames: Semantic relations of the same kind between differend classes

Thomas Russ

On May 16, 2011, at 10:33 AM, Андрей Михайлюк wrote:

> Thanks for the answer.
>
> I supposed something like that, but I believed there is a simpler way.
> Actualy, I am developing some ontology tool using OWL API. The one should have OWL as the ontology format but algorithms should be expressed in <Class-Relation> model - not the OWL. So I tried to find a correspondence between OWL and <Class-Relation> models. Now I see that I have to parse SubClassOf expressions to find out is this an inheritance between two classes or an expression of class-to-class relation.

Perhaps.

But rather than trying to parse the OWL, I would see if you can do better using the OWLAPI functions to find out the information that you want.  I would expect that to be easier and more robust than trying to do your own parsing.

You may also be able to find a simpler solution by designing your representation directly in OWL first.  As Timothy Redmond mentions, there are some choices that can be made with regard to what you do with individuals versus classes.  In general, I find that working at the individual level to be more straightforward.  This is also likely to map better into the Class-Relation model, since that would not really deal with restrictions that you can manipulate.  One such solution would be to make Microsoft and Apple be individuals of a Company class rather than subclasses.

You can then decide whether you want Windows to be a class (which you would want if you wanted to have different versions of Windows rather than a single one) or and individual.  Going with an all-class design if often popular in ontology design, and it does give you more flexibility, but for many applications you will find that working at the individual level easier and more intuitive.

Depending on which approach you take, the OWL representation of the ontology will be different.

Now to return to your initial question about how to represent class-level restrictions in Frames vs. OWL.  In OWL, properties have their own independent existence apart from classes.  This is actually also true of the Slots in the Protege Frames model, but is very different from typical object-oriented models.  So what you would need to do in the OWL model is to make sure that the class-specific restrictions on the properties are expressed using restrictions on the class definitions, while leaving the top-level domain and range constraints very general (some ontologists even recommend not specifying them at all), so as to not limit where you can add class-specific restrictions.  [I think that goes too far, but you want to keep the domain and range fairly general.]

So you would then translate:

>>> Slot "Vendor" with domain "Windows","MacOS" and allowed supperclasses "Company"
>>> Class "Company" with subclasses "Microsoft" and "Apple"
>>> Class "Windows" with slot "Vendor" restricted to allowed supperclasses "Microsoft"
>>> Class "MacOS" with slot "Vendor" restricted to allowed supperclasses "Apple"

into (class model:)

  Property vendor with domain "Product" and range "Company"
  Class OperatingSystem subclass of Product
  Class Company with subclasses Microsoft and Apple
  Class Windows, subclass of OperatingSystem with restriction
          vender some Microsoft
  ...

If you do an individual model instead, then you would use a different restriction.  If only the companies were individuals, then the restriction on the class Windows would be
      vender has Microsoft
which restrictions the filler to include Microsoft.  To make this the sole filler, the vender property would have to be functional, or you would have to add additional number or type constraints to the class Windows, such as
      vender exactly 1   OR   vender all {Microsoft}
in addition to the "vender has Microsoft" restriction.

>
> Regards,
> Andrii
>
> --- Исходное сообщение ---
> От кого: "Timothy Redmond" <[hidden email]>
> Кому: [hidden email]
> Дата: 16 мая 2011, 08:26:46
> Тема: Re: [protege-owl] OWL vs Frames: Semantic relations of the same kind between differend classes
>
>
>
>> There are several issues mixed in this question.  I suspect that the short answer is that the problem you are seeing is with the frames to OWL translation and that OWL is plenty expressive enough to say what you want.
>>
>> One of the issues in the question is the distinction between classes and individuals. When you suggest that the Vendor slot/property has domain Windows it sounds like Windows is going to be a class.  But then you talk about the triple Windows-Vendor-Microsoft, it sounds like Windows and Microsoft have either become individuals or the Vendor property is now a map between classes of classes.
>>
>> So I will choose an interpretation and try to express what you are saying in OWL.  It seems reasonable that Widows and MacOS are classes - we can very reasonably put several different operating systems in both categories.  I would have been tempted to make Microsoft and Apple be individuals of type Company but I know this is a tricky area.  To be consistent with your example we can make them classes with the idea that perhaps they each only contain one individual.  And finally Vendor will be an object property that takes products and returns the vendor of that product.  For simplicity I will assume that the Vendor property is functional.
>>
>> So one of the things that you might be wanting to say is
>>
>>>
>>> Any member of the Windows class is an operating system whose vendor is some company in the Microsoft class.
>>
>> This can easily be expressed in OWL as
>>
>> Class: <http://protege.org/ontologies/Vendors.owl#Windows>
>> SubClassOf:
>> <http://protege.org/ontologies/Vendors.owl#Vendor> some <http://protege.org/ontologies/Vendors.owl#Microsoft>
>>
>>
>> -Timothy
>>
>> On 05/15/2011 01:10 PM, Андрей Михайлюк wrote: >
>>> Hi all.
>>> In Protege Frames I could have two semantic relations of the same kind between differend classes using slots:
>>> OntologyOS:
>>> Slot "Vendor" with domain "Windows","MacOS" and allowed supperclasses "Company"
>>> Class "Company" with subclasses "Microsoft" and "Apple"
>>> Class "Windows" with slot "Vendor" restricted to allowed supperclasses "Microsoft"
>>> Class "MacOS" with slot "Vendor" restricted to allowed supperclasses "Apple"
>>>
>>> I thought it is valid ontology. But OWL cannot contain two objectProperties with the same name "Vendor" and differend domains/ranges. I mean, I cannot define two triples "Windows"-"Vendor"-"Microsoft" and "MacOS"-"Vendor"-"Apple". Is it true? When I create OntologyOS in CLIPS format - everythig is ok, when I convert it to OWL - everything is bad. Is it the OWL shortcoming or there is a way to define such information in OWL?
>>>
>>> Thanks
>>> _______________________________________________
>>> 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
>>
>>
> _______________________________________________
> 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: OWL vs Frames: Semantic relations of the same kind between differend classes

Timothy Redmond
In reply to this post by Андрей Михайлюк

So I tried to find a correspondence between OWL and <Class-Relation> models.

The first step in doing this is to figure out what the "<Class-Relation> models" really mean.    To make an analogy, if someone comes to me and asks

I love my pet.  Which is right - "I love my cat" or "I love my dog"?

the response would be to ask "what kind of pet do you have?"   Similarly with the OWL language, once you know what you are trying to say - provided it can be expressed in OWL - the OWL will flow without problems.

-Timothy


On 05/16/2011 10:33 AM, Андрей Михайлюк wrote:
Thanks for the answer.

I supposed something like that, but I believed there is a simpler way.
Actualy, I am developing some ontology tool using OWL API. The one should have OWL as the ontology format but algorithms should be expressed in <Class-Relation> model - not the OWL. So I tried to find a correspondence between OWL and <Class-Relation> models. Now I see that I have to parse SubClassOf expressions to find out is this an inheritance between two classes or an expression of class-to-class relation.

Regards,
Andrii

--- Исходное сообщение ---
 От кого: "Timothy Redmond" [hidden email] 
 Кому: [hidden email] 
 Дата: 16 мая 2011, 08:26:46 
 Тема: Re: [protege-owl] OWL vs Frames: Semantic relations of the same kind between differend classes 
 


There are several issues mixed in this question.  I suspect that the short answer is that the problem you are seeing is with the frames to OWL translation and that OWL is plenty expressive enough to say what you want.
 
 One of the issues in the question is the distinction between classes and individuals. When you suggest that the Vendor slot/property has domain Windows it sounds like Windows is going to be a class.  But then you talk about the triple Windows-Vendor-Microsoft, it sounds like Windows and Microsoft have either become individuals or the Vendor property is now a map between classes of classes.
 
 So I will choose an interpretation and try to express what you are saying in OWL.  It seems reasonable that Widows and MacOS are classes - we can very reasonably put several different operating systems in both categories.  I would have been tempted to make Microsoft and Apple be individuals of type Company but I know this is a tricky area.  To be consistent with your example we can make them classes with the idea that perhaps they each only contain one individual.  And finally Vendor will be an object property that takes products and returns the vendor of that product.  For simplicity I will assume that the Vendor property is functional.
 
 So one of the things that you might be wanting to say is
 
 > 
Any member of the Windows class is an operating system whose vendor is some company in the Microsoft class.
 
 This can easily be expressed in OWL as
 
 Class: <http://protege.org/ontologies/Vendors.owl#Windows>
 SubClassOf: 
 <http://protege.org/ontologies/Vendors.owl#Vendor> some <http://protege.org/ontologies/Vendors.owl#Microsoft>

 
 -Timothy
 
 On 05/15/2011 01:10 PM, Андрей Михайлюк wrote: > 
Hi all.
In Protege Frames I could have two semantic relations of the same kind between differend classes using slots:
OntologyOS:
Slot "Vendor" with domain "Windows","MacOS" and allowed supperclasses "Company"
Class "Company" with subclasses "Microsoft" and "Apple"
Class "Windows" with slot "Vendor" restricted to allowed supperclasses "Microsoft"
Class "MacOS" with slot "Vendor" restricted to allowed supperclasses "Apple"

I thought it is valid ontology. But OWL cannot contain two objectProperties with the same name "Vendor" and differend domains/ranges. I mean, I cannot define two triples "Windows"-"Vendor"-"Microsoft" and "MacOS"-"Vendor"-"Apple". Is it true? When I create OntologyOS in CLIPS format - everythig is ok, when I convert it to OWL - everything is bad. Is it the OWL shortcoming or there is a way to define such information in OWL?

Thanks
_______________________________________________
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


_______________________________________________
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: OWL vs Frames: Semantic relations of the samekind between differend classes

Андрей Михайлюк
Oh... Unfortunately, the task I'm working on was posted by people, who does understand the ontology only as classes connected by relations (something similar to John F Sowa's ontology model but much more simpler), so as I said, I tried to find a correspondence between OWL and <Class-Relation> models - in the most sensible way. As I understand, this is a variant of SubClassOf(:Windows ObjectSomeValuesFrom(:Vendor :Microsoft)) representation. I've implemented its translating to Class-Relation model and vice-versa just few minutes ago. Hope, that this is the most semanticaly correct way.

So, thanks for all your answers!

--- Исходное сообщение ---
 От кого: "Timothy Redmond" <[hidden email]>
 Кому: [hidden email]
 Дата: 17 мая 2011, 00:48:49
 Тема: Re: [protege-owl] OWL vs Frames: Semantic relations of the samekind between differend classes
 


> >
> > So I tried to find a correspondence between OWL and <Class-Relation> models.
>  
>  The first step in doing this is to figure out what the "<Class-Relation> models" really mean.    To make an analogy, if someone comes to me and asks
>  
>  >
> > I love my pet. Which is right - "I love my cat" or "I love my dog"?
>  
>  the response would be to ask "what kind of pet do you have?"   Similarly with the OWL language, once you know what you are trying to say - provided it can be expressed in OWL - the OWL will flow without problems.
>  
>  -Timothy
>  
>  
>  On 05/16/2011 10:33 AM, Андрей Михайлюк wrote: >
> > Thanks for the answer.
> >
> > I supposed something like that, but I believed there is a simpler way.
> > Actualy, I am developing some ontology tool using OWL API. The one should have OWL as the ontology format but algorithms should be expressed in <Class-Relation> model - not the OWL. So I tried to find a correspondence between OWL and <Class-Relation> models. Now I see that I have to parse SubClassOf expressions to find out is this an inheritance between two classes or an expression of class-to-class relation.
> >
> > Regards,
> > Andrii
> >
> > --- Исходное сообщение ---
> > От кого: "Timothy Redmond" <[hidden email]>
> > Кому: [hidden email]
> > Дата: 16 мая 2011, 08:26:46
> > Тема: Re: [protege-owl] OWL vs Frames: Semantic relations of the same kind between differend classes
> >
> >
> >
> >
> > >
> > > There are several issues mixed in this question. I suspect that the short answer is that the problem you are seeing is with the frames to OWL translation and that OWL is plenty expressive enough to say what you want.
> > >
> > > One of the issues in the question is the distinction between classes and individuals. When you suggest that the Vendor slot/property has domain Windows it sounds like Windows is going to be a class. But then you talk about the triple Windows-Vendor-Microsoft, it sounds like Windows and Microsoft have either become individuals or the Vendor property is now a map between classes of classes.
> > >
> > > So I will choose an interpretation and try to express what you are saying in OWL. It seems reasonable that Widows and MacOS are classes - we can very reasonably put several different operating systems in both categories. I would have been tempted to make Microsoft and Apple be individuals of type Company but I know this is a tricky area. To be consistent with your example we can make them classes with the idea that perhaps they each only contain one individual. And finally Vendor will be an object property that takes products and returns the vendor of that product. For simplicity I will assume that the Vendor property is functional.
> > >
> > > So one of the things that you might be wanting to say is
> > >
> > > >
> > >
> > > >
> > > > Any member of the Windows class is an operating system whose vendor is some company in the Microsoft class.
> > >
> > >
> > > This can easily be expressed in OWL as
> > >
> > > Class: <http://protege.org/ontologies/Vendors.owl#Windows>
> > > SubClassOf:
> > > <http://protege.org/ontologies/Vendors.owl#Vendor> some <http://protege.org/ontologies/Vendors.owl#Microsoft>
> > >
> > >
> > > -Timothy
> > >
> > > On 05/15/2011 01:10 PM, Андрей Михайлюк wrote: >
> > >
> > > >
> > > > Hi all.
> > > > In Protege Frames I could have two semantic relations of the same kind between differend classes using slots:
> > > > OntologyOS:
> > > > Slot "Vendor" with domain "Windows","MacOS" and allowed supperclasses "Company"
> > > > Class "Company" with subclasses "Microsoft" and "Apple"
> > > > Class "Windows" with slot "Vendor" restricted to allowed supperclasses "Microsoft"
> > > > Class "MacOS" with slot "Vendor" restricted to allowed supperclasses "Apple"
> > > >
> > > > I thought it is valid ontology. But OWL cannot contain two objectProperties with the same name "Vendor" and differend domains/ranges. I mean, I cannot define two triples "Windows"-"Vendor"-"Microsoft" and "MacOS"-"Vendor"-"Apple". Is it true? When I create OntologyOS in CLIPS format - everythig is ok, when I convert it to OWL - everything is bad. Is it the OWL shortcoming or there is a way to define such information in OWL?
> > > >
> > > > Thanks
> > > > _______________________________________________
> > > > 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
> > >
> > >
> >
> > _______________________________________________
> > 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
>
>
_______________________________________________
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