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-discussion] Datatype Properties

Timothy Redmond tredmond at stanford.edu
Fri Sep 10 08:51:56 PDT 2010


On Sep 10, 2010, at 8:36 AM, Alex Shkotin wrote:

> Timothy: - Does any of this relate to what you are trying to do?
>
> Alex: - No. I hope we discuss situation when you need to assign data  
> type property value to class as object (something like individual).  
> Consider for ex- number_of_individuals, obsolete (true/false)...
> And in P4.1beta there is no this possibility.


There are two possibilities.  First of all, if you want OWL to  
associate some semantics with this, you can pun.  The OWL class can be  
declared as an individual and you can associate the data type property  
to that.  The downside of this is that, other than allowing it,  
Protege 4.1 does not provide very good support for this type of meta- 
modeling.  I believe that there is a gforge ticket out to fix this in  
Protege 4.1.

Second, if you do not want to assign a specific OWL semantics but  
either the human or a program reading the owl file has its own  
interpretation of your value, then you can use an annotation property  
value.  An example of this might be to assign an author to an OWL  
Class or a level of certainty to an assertion.  Neither of these will  
be used by a reasoner but they may be of use to people reading the  
file or programs using the file.

In your example, you did the pun - I didn't notice that right away.   
If you look in the individuals tab you will see the punned individual  
and you will see its data value.  The representation of this in  
Protege 4.1 closely corresponds to the owl meaning but it does not  
correspond as well to what  is intended by the human who is intent on  
meta-modeling.

My personal opinion is that you should think carefully before going  
down the meta-modeling path.  I think that meta-modeling is often over- 
rated and doesn't give the desired results.  But there are some very  
important ontologies that use meta-modeling heavily (e.g. fma).

-Timothy




>
> 2010/9/10 Timothy Redmond <tredmond at stanford.edu>
>
> I am not sure what this thread is about so I am attaching an  
> ontology.  I wrote it in Protege 4.1.  It has a datatype property,  
> p, and contains several assertions:
>
> 	• the class A consists of individuals whose p data value is greater  
> than or equal to 0f and less than or equal to 1f.
> 	• the class A has an individual, i, whose p data value is .18f.
> 	• the range of the class p is the set of data values greater than  
> or equal to 0.0f and less than or equal to 1.0f.
>
> Does any of this relate to what you are trying to do?
>
> -Timothy
>
>
>
>
>
>
> On Sep 9, 2010, at 9:56 AM, Julian Vincent wrote:
>
>> I'm having trouble with attaching numerical values to classes in  
>> P4.1.  I want to add an index to each class which I can then use to  
>> estimate how similar each class is to a set of criteria, which also  
>> define the class.  I have created a Datatype but can't work out how  
>> to attach a numerical value to it.  The form I have used is that in  
>> the OWL tutorial v. 1.2 (<DatatypeName> some decimal[=0.xxxx])  
>> where xxxx are digits.  This is pretty much what's written in the  
>> tutorial text, but doesn't work.  What am I doing wrong?
>>
>> Thanks
>> Julian Vincent
>> _______________________________________________
>> protege-discussion mailing list
>> protege-discussion at lists.stanford.edu
>> https://mailman.stanford.edu/mailman/listinfo/protege-discussion
>>
>> Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03
>
>
> _______________________________________________
> protege-discussion mailing list
> protege-discussion at lists.stanford.edu
> https://mailman.stanford.edu/mailman/listinfo/protege-discussion
>
> Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03
>
>
> _______________________________________________
> protege-discussion mailing list
> protege-discussion at lists.stanford.edu
> https://mailman.stanford.edu/mailman/listinfo/protege-discussion
>
> Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03




More information about the protege-discussion mailing list