ISSUE-10 (defined roles): REPORTED: ability to have defined roles

ISSUE-10 (defined roles): REPORTED: ability to have defined roles

http://www.w3.org/2007/OWL/tracker/issues/

Raised by: Peter Patel-Schneider
On product: 

Reported by kashyap.vipul, May 23, 2007

Suppose we need to define a role called "fatherOf" which is a subrole 
of "parentOf". Would like the ability to define it as follows:

of course, subPropertyOf(father_of, parent_of)
But a more precise specification could be:

fatherOf = parentOf INTERSECTION crossProduct(Male, Person)

Another approach could be:

fatherOf = parentOf AND (ALL inverse(parentOf) Male)
However, I am not sure if the above is a well defined OWL 1.1 expression 
as it sort of mixes concept and role descriptions.

Look forward to suggestions on this issue.

Delete comment
Comment 1 by pfpschneider, Oct 11, 2007

You are asking for the ability to define roles.

A similar definition would be 
brother = sibling INTERSECTION range(Male)
using range(x) as a shorthand for crossProduct(owl:Thing,x)

This expands expressive power (I'm pretty sure).  I don't know whether there is a
reasoning algorithm for this construct.


Summary: Need a role descrption based on the "cross product" of two concepts
Comment 2 by kashyap.vipul, Oct 11, 2007

Peter,

Thanks for the response. 
Agree with you is that what I am looking for is role definitions.

So the other thing you bring up is that are we focused only on reasoning as a 
functionality?

Would we want to consider other functionalities such as "loss-less round trip and 
model interchange" or querying and data retrieval in addition to reasoning?



Delete comment
Comment 3 by pfpschneider, Oct 12, 2007

Clarification to comment 1:

The shorthand is confusing.  I meant to define brother as "those sibling
relationships whose object is Male".

Delete comment
Comment 4 by pfpschneider, Oct 12, 2007

Yes, there are other things that need to be considered besides reasoning.  However,
without the ability to do reasoning (reasonably) effectively, how are you going to do
the other things you want?  It seems to me that data retrieval is a kind of
reasoning, and lossless round-tripping depend on reasoning.


Comment 5 by kashyap.vipul, Oct 12, 2007

As far as querying and data retrieval is concerned, there have been effective and 
scalable ways of specifying algebras and coming up with query optimization 
techniques. Would view this as an alternative model of computation in contrast to 
reasoning.

I get the feeling that I am now going out of scope for the WG, so feel free to 
scope this discussion:

Would it make sense to consider say an algebra underlying OWL 1.1 with the goal of 
identifying optimizations for data and query retrieval? Of course the tableaux 
technique seeks to optimize an inference operation and there might be some common 
ground between the two approaches?

Comment 6 by kashyap.vipul, Oct 12, 2007

Response to comment 3:

Let's park the issue of role definitions for now.

Is it possible to specify "those sibling
relationships whose object is Male"

Will range(brother) = Male
and brother subPropertyOf sibling do the trick?

Received on Wednesday, 24 October 2007 21:01:51 UTC