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] does protege allow this kind of check consistency?

Thomas Russ tar at ISI.EDU
Thu Apr 29 09:49:39 PDT 2010


On Apr 29, 2010, at 9:25 AM, Sara Paiva wrote:

> Yes i can understand that.
>
> But like has been said in all this replies it would require (which  
> is logical too) that i have a domain knowledge base where the rules  
> of my domain were established.
>
> For example:
>
> i can´t have two restrictions over number of pages if i use  
> "and" (because it makes no sense)
>
> i can have two restrictions over number of pages if i use  
> "or"  (therefore the  number of pages property cannot be functional)

This isn't a problem, because the AND or OR happen outside the  
restriction.  So it can and should be functional, even in the or  
case.  What you end up with is the following:

   intersection([number-of-pages has 1], [number-of-pages has 2])
   union([number-of-pages has 1], [number-of-pages has 2])

That works just fine with functional properties.  I suggest you try  
some of these things out in Protege before deciding whether it will  
work or not.

> For what i can see i can perform this checks, but not in an  
> independent domain way, like I am performing.

The content of your ontology cannot be domain-independent and be  
actually useful.  You have to know the meaning of the terms you define  
for your domain.  That will be true no matter what technology you use  
to accomplish your task.  Even if you end up programming your own  
special-purpose reasoning system, you will still need to encode that  
knowledge.


> On Thu, Apr 29, 2010 at 5:18 PM, Thomas Russ <tar at isi.edu> wrote:
>
> On Apr 29, 2010, at 9:00 AM, Sara Paiva wrote:
>
> I am working on semantic validation of queries (imagine Google  
> searches: "books which title is Da Vinci Code and which title is  
> Digital Fortress). If they don´t make sense why perform them? They  
> will always return empty results.
>
> Right know i am checking if owl has a way to check this form me  
> through unstatisfiable classes.
>
> but from you help i am getting to the conclusion that is not possible.
>
> Exactly the opposite is true.
>
> If you read the replies again, carefully, you will see that OWL can  
> identify problems with classes definitions that require more than  
> one different filler for properties that ARE KNOWN TO TAKE ONLY ONE  
> FILLER.  But you have to tell the system which properties are like  
> that.
>
> Otherwise, how is the reasoner supposed to know that a book can only  
> have one title but it can have more than one author?  You have to  
> tell it that.
>
> Now, you might run into other places where OWL does not have the  
> expressive power for the constraints you want, but this particular  
> one is a place where you can, in fact, get the inconsistent result  
> you desire.
>
> [I do still disagree that an unsatisfiable class should lead to an  
> inconsistent ontology as long as it doesn't have any individuals.]
>
>
>
> On Thu, Apr 29, 2010 at 4:52 PM, Thomas Schneider <schneidt at cs.man.ac.uk 
> > wrote:
>
> On 29 Apr 2010, at 16:45, sop wrote:
>
>
> Oh i get you point.
>
> But i want to go a little further than that and that´s what i don´t  
> know the
> reasoner can do.
>
> Let me clarify for you.
>
> Suppose i want to transform a query into a owl class description  
> just to see
> if the query can be satisfied (not obtain results).
>
> now suppose restriction 1 is "number of pages is 1" and the other
> restriction is "number of pages is 2". this query makes no sense as  
> no book
> can have 1 page and, at the same time, 2 pages. Right?
>
> Well, for us as humans it doesn't make sense to claim that a book  
> has exactly one and exactly two pages at the same time -- because we  
> have the background knowledge that no book can have two different  
> numbers of pages. If you're writing an ontology, you need to assert  
> that information as Timothy or I have suggested. The reasoner cannot  
> make any general-world inferences from the names of your properties  
> because it focusses on the logical structure of the ontology.
>
> I still don't know which queries you mean and how you want to  
> translate them into class descriptions.
>
> Cheers
>
> Thomas
>
>
>
> are you understanding me? do you think this can be done?
>
> plus: is there the notion of "and" and "or" because the same two
> restrictions above with "or" already made sense.......
>
> hope I am not making this confusing.
>
>
>
>
>
>
> -- 
> View this message in context: http://protege-ontology-editor-knowledge-acquisition-system.136.n4.nabble.com/does-protege-allow-this-kind-of-check-consistency-tp2075649p2075726.html
> Sent from the Protege OWL mailing list archive at Nabble.com.
> _______________________________________________
> 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
>
> + 
> ----------------------------------------------------------------------+
> |  Dr Thomas Schneider                    schneider (at)  
> cs.man.ac.uk  |
> |  School of Computer Science       http://www.cs.man.ac.uk/ 
> ~schneidt  |
> |  Kilburn Building, Room 2.114                 phone +44 161  
> 2756136  |
> |  University of  
> Manchester                                            |
> |  Oxford Road                                             _/// 
> _       |
> |  Manchester M13 9PL                                       
> (o~o)       |
> +-----------------------------------------------------oOOO--(_)-- 
> OOOo--+
>
> Sconser (n.)
>  A person who looks around them when talking to you, to see if there's
>  anyone more interesting about.
>
>                 Douglas Adams, John Lloyd: The Deeper Meaning of Liff
>
>
>
>
>
>
>
>
> _______________________________________________
> 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
>
>
> _______________________________________________
> 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
>
> _______________________________________________
> 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
>
> _______________________________________________
> 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




More information about the protege-owl mailing list