Re: Is there an implementer's convention for indicating false?

Graham Klyne,

> I'm wondering if there's any common convention among 
> implementers of RDF inference systems for indicating 
> that some offered premis is unsatisfiable, in systems 
> where all results are presented as some form of RDF.

I don't work on inference systems, but here's an educated guess:

> e.g.
> 
>     _:a owl:sameAs "1" .
>     _:a owl:sameAs "2" .
> =>
>     ?
>
> (expressed as an RDF graph.)

This might work, but it seems too _verbose_ and uses the vaguely defined
rdf:value property.  Is there a simpler way with more defined semantics?

] _:o1 rdf:type owl:Ontology .
] _:o1 owl:incompatibleWith _:o2 .
] _:o1 rdf:value t1 .
]   _:t1 rdf:type rdf:Statement .
]     _:t1 rdf:subject _:a .
]     _:t1 rdf:predicate owl:sameAs .
]     _:t1 rdf:object "1" .

] _:o2 rdf:type owl:Ontology .
] _:o2 owl:incompatibleWith _:o1 .
] _:o2 rdf:value t2 .
]   _:t2 rdf:type rdf:Statement .
]     _:t2 rdf:subject _:a .
]     _:t2 rdf:predicate owl:sameAs .
]     _:t2 rdf:object "2" .

I wish we didn't need to define new ontologies to contain each triple.
Furthermore, the below feels good, but won't work because of the precise
semantics (not to mention domain and range) of owl:incompatibleWith.  I
still think the below _should be_ allowed:

] _:t1 rdf:type rdf:Statement .
]   _:t1 owl:incompatibleWith _:t2 .
]   _:t1 rdf:subject _:a .
]   _:t1 rdf:predicate owl:sameAs .
]   _:t1 rdf:object "1" .

] _:t2 rdf:type rdf:Statement .
]   _:t2 owl:incompatibleWith _:t1 .
]   _:t2 rdf:subject _:a .
]   _:t2 rdf:predicate owl:sameAs .
]   _:t2 rdf:object "2" .

Again, I never needed to answer such a problem, but that's what my initial
thinking would be.  Hope it helps!

--
Jimmy Cerra

] "A good decision is based on knowledge 
]  and not on numbers" - Plato

Received on Wednesday, 12 November 2003 22:54:05 UTC