W3C home > Mailing lists > Public > public-owl-dev@w3.org > July to September 2007

Re: Some advice on inferring negated properties

From: Bijan Parsia <bparsia@cs.man.ac.uk>
Date: Fri, 17 Aug 2007 08:50:53 +0100
Message-Id: <4A2AB559-B635-4AD2-AA7B-070521FE3A90@cs.man.ac.uk>
Cc: "Swanson, Tim" <tim.swanson@semanticarts.com>, "Owl Dev" <public-owl-dev@w3.org>
To: Alan Ruttenberg <alanruttenberg@gmail.com>

On Aug 17, 2007, at 6:18 AM, Alan Ruttenberg wrote:

> On Aug 16, 2007, at 7:26 PM, Bijan Parsia wrote:
>
>>
>> 	<http://www.webont.org/owl/1.1/owl_specification.html#6.2>
>> 	DisjointObjectProperties
>>
>> Which, well, at this time at night at least, I don't see can be  
>> paraphrased into oldOWL. Hmm. Ok if you did some disjoint ranging  
>> and domaining, that would entail disjoint properties.
>
> True, though DisjointObjectProperties is stronger, yes, since it  
> doesn't require disjoint domain/ranges,

Hmm.

> just that for any instance, the sets of values of the disjoint  
> properties are disjoint. Do I have this right?

P domain C.
P range C.

Q domain D.
Q range D.

C disjointFrom D.

So, this rules out <x, y>:P and <x, y>:Q, but also <x, y>:P and <x,  
z>:Q. (Whether this is "stronger" because it rules out more things or  
"weaker" because it doesn't capture this state depends on your  
taste :)).

P domain C.
P range D.

Q domain D.
Q range D.
C disjointFrom D.

Gets the above case, but not the symmetric one.

So, yeah, looks like you can entail that DisjointObjectProperties(P,  
C) from domain hackery, but not the reverse.

Cheers,
Bijan.
Received on Friday, 17 August 2007 07:51:10 GMT

This archive was generated by hypermail 2.3.1 : Wednesday, 27 March 2013 09:32:55 GMT