Re: sameAs vs. differentFrom

On Thu, 18 Oct 2018 at 18:57, Bijan Parsia
<bijan.parsia@manchester.ac.uk> wrote:
>
> On Oct 18, 2018, at 18:46, Cristian Cocos <cristi@ieee.org> wrote:
>
> I was surprised to see that two individuals that were previously asserted as owl:differentFrom were inferred to be, at the same time, owl:sameAs (due to some unrelated axioms).
>
>
> That can happen. Consider the trivial case:
>
> a sameAs b.
> a differentFrom b.
>
> Or slightly more elaborately
>
> a sameAs b.
> a type C.
> b type not C.
>
> Are these two properties not supposed to be disjoint?
>
>
> Yes but that doesn’t prevent me from saying silly things.
>
> Do reasoners know they are disjoint?
>
>
> Yes.
>
> I certainly have not seen these two explicitly declared as disjoint,
>
>
> It’s in the semantics.
>
> but I thought that that was something that goes without saying.
>
>
> It never goes without saying but it has been said (not as such, but as a consequence of the semantics).
>
> (And if they have not been conceived as disjoint, why so?)
>
> Anyway, what struck me was that the reasoner did not deem the ontology inconsistent.
>
>
> Then the reasoner is incomplete.
>
> You should always say which reasoner. Hermit or pellet or FaCT++ will definitely find the inconsistency.


One question lingers.

Christian said 'the reasoner'.
Complete reasoners should not have behaved that way (modulo bugs of course).
Which reasoner misbehaved? Ideally, which version? So that a proper
bug report can be raised.

Cheers,
Ignazio

>
> Any W3C OWL insiders care to explain this?
>
> Many thanks,
>
> Christian
>
> --
> "People think that I must be a very strange person. This is not correct. I have the heart of a small boy. It is in a glass jar on my desk." -- Stephen King

Received on Thursday, 18 October 2018 20:47:43 UTC