W3C home > Mailing lists > Public > public-owl-dev@w3.org > October to December 2010

Re: class and inviduals

From: Bijan Parsia <bparsia@cs.man.ac.uk>
Date: Tue, 23 Nov 2010 12:50:32 +0000
Message-Id: <F9CC70D1-E9D2-45B8-8C07-629F77BDB35E@cs.man.ac.uk>
To: Owl Dev <public-owl-dev@w3.org>
On 21 Nov 2010, at 19:00, Bijan Parsia wrote:

> On 21 Nov 2010, at 09:49, Birte Glimm wrote:
> 
>> Hi all,
>> I am a bit lost in this longish discussion  for which I didn't yet
>> check the whole history and I am on holiday at the moment with very
>> limited internet access. It would be great if there was a more concise
>> problem description, then I'll get back to it at the end of next week.
> 
> The first point is to confirm that your metamodelling approach *doesn't* support the canonical Hilog entailments.
> 
> That is, from:
> 	C sameAs D.
> 
> We do not get:
> 	C equvialent D.
> 
> My first guess is that we *do* get this at the metalevel (due to the MatSubClass axioms) assuming we can state that o_someclass sameAs o_someotherclass, but I don't see how we'd propagate that back down to the object level. Indeed, it seems to conflict with Theorem 1 of your paper.


Sebastian Rudolph wrote me (he couldn't send to the list:
[snip]
> Let me step in for Birte here (as I co-authored the paper). In fact, Theorem 1 just states that if you just rewrite the KB in the presented way, the consequences on the original vocabulary do not change. However, if you additionally postulate axioms on the extended (reified) vocabulary this will change things. In particular if you assert o_houseMouse same as o_fieldMouse, this will have exactly the effect that Bijan assumed, namely extensional equivalence of the two classes. Feel free to check with any OWL-compliant reasoner.
> 
> Cheers,
>   Sebastian

I stand corrected. Hilog semantics (of some form, i.e., straitified; for classes) is "as easy" to implement as punning (i.e., via a preprocessing step). This is definitely new information.

I still question whether it is *desirable* (and, of course, you still need to be careful with roles). But, we could at least try out real KBs!

So, Hilog semantics fans. Please send your favorite example ontology. If this does, in fact, work out, it might be possible to add it to the standard or at least to the de facto standard.

Cheers,
Bijan.
Received on Tuesday, 23 November 2010 12:50:52 GMT

This archive was generated by hypermail 2.3.1 : Wednesday, 27 March 2013 09:32:58 GMT