Fork me on GitHub
#rdf
<
2022-10-19
>
rickmoynihan08:10:51

@simongray yes that’s all true. We’ve often wanted for the same feature.

simongray14:10:03

Welp, at least I was correct in my assumptions. Thanks for letting me know.

quoll14:10:42

Does anyone else use reasoning, and if so, have they seen owl:irreflexiveProperty being treated correctly?

rickmoynihan15:10:35

@quoll: I’ve not come across that one before; is that just in OWL Full? Does irreflexiveProperty mean that :p can never be used to do :foo :p :foo ??

quoll15:10:51

That’s right

quoll15:10:02

But I’m finding that it’s being ignored by Pellet

rickmoynihan09:10:42

what OWL profile are you running? It looks like it should be included in EL QL and RL

quoll20:10:57

I tried in both SL and DL

quoll21:10:55

and Pellet is DL (i.e. when you select “DL” then you get Pellet)

rickmoynihan10:10:07

It’s always confused me a little how the OWL 2 profiles relate to OWL 2 DL… My understanding is that OWL 2 DL is essentially the set of computable OWL; drawn from OWL 2 Full with syntax restrictions. Then the profiles are all in OWL 2 DL; but are subsets with known computational guarantees. I don’t know though why OWL 2 DL doesn’t provide any (though OWL 1 DL does); is it because it’s unknown?? i.e. why isn’t OWL 2 DL listed in Table 10 here? https://www.w3.org/TR/2012/REC-owl2-profiles-20121211/#Computational_Properties

rickmoynihan10:10:47

Hmm I guess it’s because OWL 2 Direct Semantics is OWL 2 DL??

rickmoynihan10:10:18

OWL can be such a rats nest :rolling_on_the_floor_laughing:

quoll14:10:46

Sorry I wasn’t responding… Yes. When I first started working with OWL 2 I was trying to figure out “which level is DL?” And that’s what I came up with too. It makes sense, but not when you compare it to OWL 1

quoll15:10:18

and all the other reasoning levels in Stardog

quoll15:10:05

For instance:

:sibling a owl:IrreflexiveProperty, owl:SymmetricProperty;
         owl:propertyChainAxiom (:parent :child).
Given all the :parent and :child relationships, this should find siblings, but it always infers that people are their own sibling

rickmoynihan09:10:44

ah interesting; I was wondering how it was useful as it didn’t appear to make any entailments… so it’s main use is in essentially filtering out reflexive inferences you don’t want; makes sense.

quoll21:10:28

Well, I figured it would make sense… if it worked :rolling_on_the_floor_laughing:

quoll21:10:56

Protegé says you can’t have owl:IrreflexiveProperty on “complex” properties. Which seems to refer to any property that has any description on it