Thanks for the explanation.  Working with ontologies is certainly a
different way of thinking, but I understand what you are saying.


On Wed, Oct 2, 2013 at 1:10 PM, Chris_Dollin <[email protected]> wrote:

> On Wednesday, October 02, 2013 10:47:45 AM Ralph Perniciaro wrote:
> > I believe that I have the same issue.  If I create an ontology in protege
> > and define an entity to have a property called title and specify that it
> > should have exactly 1 title.  If I load the model into Jena and then
> create
> > an individual of my entity type, I can add more than 1 title to the
> > individual.  I was going to add my own checks to prevent this, unless
> > someone can explain how Jena can enforce cardinality rules.
>
> Jena doesn't enforce cardinality rules, except in the sense that you
> can run validation checks to make sure the model is consistent.
> But note that if C is some class with a restriction that P has exactly
> one value, and you assert
>
>     c rdf:type C
>     c P a
>     c P b
>
> then you haven't violated a cardinality rule; you've asserted that
>
>     a owl:sameAs b
>
> which may or may not be generated by the inference you're using.
>
> (Since `title` is probably string-valued, it's a little trickier, "of
> course".)
>
> If you want on-the-fly cardinality checks in Jena you can always
> write them youself. That way you can be as pragmatic as you please.
>
> Chris
>
>
>
>
>

Reply via email to