problem with Inverse Property with "some" restriction

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

problem with Inverse Property with "some" restriction

Tran Binh
Hi all,

Maybe some ones already ask this question but I can not found.

I have a simple ontology file:

Class A
subclass A1, A2

Class B
subclass B1, B2

object property: O1
inverse object property O2

B1 has O1 some A1
B1 has O1 some A2

B2 has O1 some A2

when I search for: B1 and O1 some A. protege return a perfect result but with
A1 and O2 some B, it returns nothing. I suggest it should return B1

Is it something wrong with my ideas, or that is the rule of Open World?

Any ideas?

Thank in advance
Binh

Reply | Threaded
Open this post in threaded view
|

Re: problem with Inverse Property with "some" restriction

Thomas Russ

On Mar 23, 2011, at 3:08 AM, Tran Binh wrote:

> Hi all,
>
> Maybe some ones already ask this question but I can not found.
>
> I have a simple ontology file:
>
> Class A
> subclass A1, A2
>
> Class B
> subclass B1, B2
>
> object property: O1
> inverse object property O2
>
> B1 has O1 some A1
> B1 has O1 some A2
>
> B2 has O1 some A2
>
> when I search for: B1 and O1 some A. protege return a perfect result  
> but
> with
> A1 and O2 some B, it returns nothing. I suggest it should return B1
>
> Is it something wrong with my ideas, or that is the rule of Open  
> World?

You are getting correct results.  And it doesn't involve open world  
semantics.

   B1 has O1 some A1

means that every individual of type B1 must have at least one  
individual of type A1 connected by the O1 property.

Now, any individual of type A1 that happens to be the value of  
property O1 will have an individual that is the value of the O2  
property.  But nothing says that EVERY individual of type A1 has that  
sort of connection.

It would be perfectly consistent for your ontology to have the  
following individuals and properties:

   b11 type B1
   a11 type A1
   a12 type A1
   a21 type A2

   b11 O1 a11
   b21 O1 a21

Notice that there is an individual of type A "a12" that is not  
connected to any B1 individuals.  So it would be incorrect to say that  
every A1 must have a B for O2.









_______________________________________________
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