Reasoner error? about owlsharp

[mdesalvo]

Hi constnick,
thanks as usual for the attention to the library and for your detailed reports!

This is an error of the reasoner, which should not consider blank properties during the execution of this specific rule: I'll patch this rule, after the patch you should not see any inferences coming from this rule. I think you should try instead the "InverseOfEntailment" rule, which reverses object assertions by exploiting owl:inverseOf relations between object properties. I'll also check this one in order to see if it is smart enough to detect the expected inference. I'll also patch this, in case it doesn't work.

Regards,
Marco

[mdesalvo]

Checked-in fixes in bd56416:

• Discard blank properties during "PropertyEntailment" reasoner rule, so that it will not crash anymore when encountering blank candidate properties coming from OWL2-full anonymous inline property expressions
• Recognize OWL2-full anonymous inline property expressions during owl:inverseOf computation, so that "InverseOfEntailment" reasoner rule will be able to infer "ex:iJunior ex:DescendantOf ex:iMajor"

[mdesalvo]

Hi @constnick,
as you are an expert ontologist, can you kindly explain me the purpose of the syntax:

ex:ParentOf rdfs:subPropertyOf [ owl:inverseOf ex:DescendantOf ] .

What is the advantage against an explicit statement:

ex:ParentOf owl:inverseOf ex:DescendantOf

Can this also exist in equivalency way:

ex:ParentOf owl:equivalentProperty [ owl:inverseOf ex:DescendantOf ] .

Thanks,
Marco

[constnick]

ex:ParentOf owl:inverseOf ex:DescendantOf — it is a wrong assertion within the model.
The inverse of DescendantOf is something like AncestorOf, and ParentOf is the subproperty of it.
Of course we can create the inverse property explicitly, but it is usually inconvenient to create inverse of every property for possible rare usage.

[mdesalvo]

Thanks for the clarification.

I checked on Protege with Pellet and HermiT reasoners emitting "ex:iJunior ex:DescendantOf ex:iMajor" inference. I confirm that now our rule named "InverseOfEntailment" is compliant. Expect the updated version soon.

Regards,
Marco