OWL Protege modeling question

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

OWL Protege modeling question

Nancy Wiegand
We are prototyping an application in Protege OWL that I am more or less
modeling after the Wine Agent example from the end of the W3C OWL guide
(http://www.ksl.stanford.edu/projects/wine/explanation.html). My
question is why was it necessary to define (an extra) meal course class
rather than just modeling a "needs" relationship between the main food
and drink classes? That is, if the point is to associate a type of drink
with a type of food, why not just model an objectProperty (such as
"needs") that has the appropriate restrictions between the food class
and the drink class for each type of food? It seems that would
accomplish the same thing in a more straightforward manner, i.e., allow
inferencing to find the drink needed for each food type. Is there
something wrong or lacking in some way, such as maybe modeling the
restrictions on an object property versus a class?

Thank you,
   Nancy Wiegand
-------------------------------------------------------------------------
To unsubscribe go to http://protege.stanford.edu/community/subscribe.html

Reply | Threaded
Open this post in threaded view
|

[protege-owl] Re: OWL Protege modeling question

Nick Drummond
Nancy,

A simpler model could indeed implement what you have suggested.
I suppose one argument is that it is not a property of the food itself
that it goes with a particular style of wine.

It may be the other properties of the food which determine this. By
seperating this knowledge off into the notion of a course its possible
to define a general course type that puts together SpicyFood and
FragrantWine for example (and we may not want to have to force SpicyFood
and FragrantWine to be named classes).

This seperation also gives us an advantage that we have no dependencies
between our food and wine ontologies - this information is stored in a
third "mapping" ontology that described courses.

Nick


Nancy Wiegand wrote:

> We are prototyping an application in Protege OWL that I am more or
> less modeling after the Wine Agent example from the end of the W3C OWL
> guide (http://www.ksl.stanford.edu/projects/wine/explanation.html). My
> question is why was it necessary to define (an extra) meal course
> class rather than just modeling a "needs" relationship between the
> main food and drink classes? That is, if the point is to associate a
> type of drink with a type of food, why not just model an
> objectProperty (such as "needs") that has the appropriate restrictions
> between the food class and the drink class for each type of food? It
> seems that would accomplish the same thing in a more straightforward
> manner, i.e., allow inferencing to find the drink needed for each food
> type. Is there something wrong or lacking in some way, such as maybe
> modeling the restrictions on an object property versus a class?
>
> Thank you,
>   Nancy Wiegand
> -------------------------------------------------------------------------
> To unsubscribe go to http://protege.stanford.edu/community/subscribe.html
>
>

--

Nick Drummond

http://www.cs.man.ac.uk/~drummond/ <http://www.cs.man.ac.uk/%7Edrummond/>
-------------------------------------------------------------------------
To unsubscribe go to http://protege.stanford.edu/community/subscribe.html