Search Mailing List Archives


Limit search to: Subject & Body Subject Author
Sort by: Reverse Sort
Limit to: All This Week Last Week This Month Last Month
Select Date Range     through    

[protege-owl] Domain/Range vs OWL Restrictions

Juan Sequeda juanfederico at gmail.com
Wed Aug 15 00:22:42 PDT 2007


On 8/14/07, Kaarel Kaljurand <kaljurand at gmail.com> wrote:
>
> Hi,
>
> On 8/14/07, Juan Sequeda <juanfederico at gmail.com> wrote:
> >
> > I have a modelling question.
> >
> > Lets say for example I have Class Student and Class Course and there is
> an
> > objectProperty isEnrolled whose domain is Student and range is Course.
> >
> > Now, if I had an OWL restriction on the Class Course saying that
> >
> > :Course
> >  a owl:Class ;
> >  rdfs:subClassOf
> >  [ a owl:Restriction ;
> >  owl:onProperty :isEnrolled;
> >  owl:allValuesFrom :Student
> >  ] .
>
> Domain and range in your example can be expressed by the following
> natural language sentences:
>
> Everybody that is-enrolled-in something in a student. (domain)
> Everything that somebody is-enrolled-in is a course.  (range)
>
> or in Manchester Syntax as general SubClassOf-axioms:
>
> (is-enrolled-in some Thing) subClassOf student
> (inv(is-enrolled-in) some Thing) subClassOf course


True.. that is exactly what domain and range means.

Now you additionally have:
>
> :Course
>   a owl:Class ;
>   rdfs:subClassOf
>     [ a owl:Restriction ;
>       owl:onProperty :isEnrolled;
>       owl:allValuesFrom :Student
>    ] .
>
> which I understand as (in Manchester Syntax):
>
> course subClassOf (is-enrolled-in only student)


But then how  does this OWL restriction differ from the restriction that the
domain/range creates. It seems that it is the same.

which can be rewritten as:
>
> (inv(is-enrolled-in) some course) subClassOf student
>
> which in natural language reads:
>
> Everything that a course is-enrolled-in is a student.
>
> That's weird! Did you intend to use the inverse of
> is-enrolled-in instead? I.e.:
>
> course subClassOf (inv(is-enrolled-in) only student)
>
> which can be rewritten as:
>
> (is-enrolled-in some course) subClassOf student
>
> which in natural language reads:
>
> Everybody that is-enrolled-in a course is a student.
>
>
> > Does that make my ontology "stronger"?
>
> What do you mean by "stronger"?


When I mean stronger... what are the implications of using only the
domain/range or only the owl:restriction, or having both. Is this more
expressive? What does this imply on the reasoner

> Does this have to do with the open world assumption, meaning that if I
> only
> > use the domain/range (which is RDFS) and I use a reasoner, then
> something
> > else could be classified in that group?
>
> What group?


If I don't state that the Course can ONLY have students... I am leaving the
world open (open world assumption) and with this restriction I can close the
world. Is this the correct way to do it?

> What if I only use this OWL restriction and don't use the domain/range.
> What
> > implications does that have.
>
> Things can be enrolled in other things than only courses, and the
> enrolled things are
> not necessarily students.



But if I only need  student?

Sincerely,
> kaarel
> _______________________________________________
> protege-owl mailing list
> protege-owl at lists.stanford.edu
> https://mailman.stanford.edu/mailman/listinfo/protege-owl
>
> Instructions for unsubscribing:
> http://protege.stanford.edu/doc/faq.html#01a.03
>



-- 
Juan Sequeda

Research Assistant                             Project Manager
Dept. of Computer Science                  Solutions Superbacano.com
University of Texas at Austin                Austin, Texas, USA
http://www.cs.utexas.edu/~jsequeda    http://solutions.superbacano.com
jsequeda at cs.utexas.edu                    juan at superbacano.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.stanford.edu/pipermail/protege-owl/attachments/20070815/42350a29/attachment.html>


More information about the protege-owl mailing list