On Jul 24, 2008, at 4:16 AM, Robert A. Elliott, Sr. wrote:
> Thomas,
> All points are clear. I'm running Protege 3.4beta and I do have a
> direct connection to Pellet.
>
> 1. How can I address the Pellet connection issue in Protege 3.4?
This is selected from the "Reasoning" menu. Make sure the Pellet
1.5.2 reasoner is chosen and not DIG.
>
> 2. How can I restrict values on datatypes? i.e. If I want to make
> sure each individual had a unique value for a certain datatype, how
> can this be done in Protege 3.4?
There are some limitations to what you can restrict.
hasValue can be used to assign particular datatype values. In other
words, you can assign a specific filler value as part of a restriction.
Restricting the value to a particular range of integer or float can be
done, but it is more complicated. In OWL 1.0 you need to create a
custom datatype that is a subrange. But AFAIK that won't necessarily
be supported by the reasoners. Otherwise you need to use OWL 1.1/2.0,
which has syntax for such type restrictions.
As far as assuring a unique value for each individual's values, I
don't think you can do it using OWL. There isn't any construct for
expressing unique value, and no way to express cross-individual
constraints. And for datatype properties, you can't have inverses,
which is something that would work with object properties.
> 2a. If the answer lies within Protege 4.0 ....see item 3.
Possibly. I don't use Protege 4.
> 3. Will Protege 3.4 Owl ontologies "work"/import in Protege 4.0?
Yes, they should. They are all OWL files.
>
> Thanks.
>
> Robert.
>
>
>
> On Wed, Jul 23, 2008 at 6:52 PM, Thomas Russ <
tar@...> wrote:
> On Jul 23, 2008, at 9:31 AM, Robert A. Elliott, Sr. wrote:
>
> > All,
> > 1. I defined an integer functional datatype (RequirementNumber)
> > within a certain class with a min cardinality restriction of at
> > least 1.
>
> OK. That means there must be at least 1 filler of that datatype
> property. And since it is a functional property, that implies exactly
> one filler.
>
> > 2. I created two individuals in that class and assigned 0 and 5
> > (RequirementNumber) to each individual respectively.
>
> OK. That is fine. Each individual has one filler of the
> RequirementNumber property. That is fully consistent with the
> definition.
>
> Even without assigning a value, it would be consistent because of open
> world reasoning considerations.
>
> > 3. I run a consistency check and Pellet randomly passes and flags
> > the ontology. i.e. On some runs it flags and warnds and on other
> > occasions it does not warn...(0). (See attachments)
>
> It seems that sometimes you are using a direct connection to Pellet
> and sometimes going via DIG. You should make sure that the direct
> connection is always chosen in the Reasoning menu.
>
> > Shouldn't it NOT allow a value of 0?.....
>
> No. A value of zero is quite fine.
>
> Cardinality restrictions are concerned with the NUMBER of fillers, not
> with their VALUE. To restrict the value, you would need a different
> sort of restriction. You would need an appropriate type restriction
> to be something like positive integers. This is easier to do using
> OWL 1.1/2.0 instead.
>
> >
> > Am I correct in this assessment of the output from Pellet? Is this a
> > bug? Or am I missing something?
> >
> > Robert...
> >
> >
> >
> > Robert A. Elliott, Sr.
> > Ph.D. Student
> > Department of Computer Science & Engineering
> > Mississippi State University
> >
raesr20@...
> >
> <nopass.jpg><pass.jpg>_______________________________________________
> > protege-owl mailing list
> >
protege-owl@...
> >
https://mailman.stanford.edu/mailman/listinfo/protege-owl> >
> > Instructions for unsubscribing:
http://protege.stanford.edu/doc/faq.html#01a.03>
> _______________________________________________
> protege-owl mailing list
>
protege-owl@...
>
https://mailman.stanford.edu/mailman/listinfo/protege-owl>
> Instructions for unsubscribing:
http://protege.stanford.edu/doc/faq.html#01a.03>
>
>
> --
> Robert A. Elliott, Sr.
> Ph.D. Student
> Department of Computer Science & Engineering
> Mississippi State University
>
raesr20@...
> _______________________________________________
> protege-owl mailing list
>
protege-owl@...
>
https://mailman.stanford.edu/mailman/listinfo/protege-owl>
> Instructions for unsubscribing:
http://protege.stanford.edu/doc/faq.html#01a.03_______________________________________________
protege-owl mailing list
protege-owl@...
https://mailman.stanford.edu/mailman/listinfo/protege-owlInstructions for unsubscribing:
http://protege.stanford.edu/doc/faq.html#01a.03
nopass.jpg (98K)