[redland-dev] RDFS/OWL and reasoning in Redland

Richard Newman r.newman at reading.ac.uk
Wed Sep 27 18:15:17 UTC 2006


On  27 Sep 2006, at 7:03 PM, Andre Meyer wrote:

> Well, I agree with the reasoner part. But these are two separate  
> things: an ontology enforcing the validity of triples in the store  
> on the one hand and a reasoner that adds triples based on rules and  
> ontological conclusions on the other hand. Does this make sense?

No, it doesn't.

In the case of the semantic web, ontologies and rules are used by  
reasoners and rule engines respectively to infer conclusions. These  
conclusions might be "Richard is a Person", or they might be "your  
store is inconsistent for X, Y, Z reasons". They will not say "you  
can't add this triple, it's invalid; People don't have wheels".  
Instead, your person is inferred to also be a car. It's up to you  
what you do with this information.

You are thinking in terms of schemas in the XML sense. Disabuse  
yourself of this notion.

Ontologies do not enforce the validity of anything (at least, not in  
the sense in which you are thinking of 'enforcing').

You can try to do validation by making your store logically  
inconsistent:

Richard is a Person
Richard has a Wheel
Anything with Wheels is a Car
 >>> People are disjoint from Cars  <<< this is your attempt at  
validation.

but it's hard to make this work.

-R
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.usefulinc.com/pipermail/redland-dev/attachments/20060927/7b85deb3/attachment.htm


More information about the redland-dev mailing list