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] Comparing individuals

Ely Edison Matos ely.matos at
Tue Aug 7 12:19:00 PDT 2007

Hi Thomas and fellows,
Thanks for the comments and sugestions. Actually, i'm converting xml files
that already has a schema. There is a notion of "standard units" (base units
of International Systems SI, like meter, ampere, kilogram, etc.) and "user
defined units", defined as operations over the base units (e.g. centimeter
is a operation over meter). At the ontology I created a class "Unit" and two
subclasses "StandardUnit" and "UserDefinedUnit". Using your example,
"kilometer" will be a UserDefinedUnit with a property "isDerivedFrom"
The notion of "isSameAs" will be applied to variables. Variables are
associated to a Unit. I'll need to know if variables A and B are associated
with the same units (or variations of the same Unit), among others
similarities, they are "equals" or "the same" variables.
I'm accepting sugestions and I will be happy to discuss the "ontology
engeneering"...Only I dont't know if this is the adequate forum or there are
other lists to this matter.


----- Original Message ----- 
From: "Thomas Russ" <tar at ISI.EDU>
To: "User support for the Protege-OWL editor"
<protege-owl at>
Sent: Tuesday, August 07, 2007 3:26 PM
Subject: Re: [protege-owl] Comparing individuals

> On Aug 5, 2007, at 6:27 AM, Ely Edison Matos wrote:
> > Hello semantic fellows,
> > I have a begginer's doubt, i hope you can help. I need to know thar
> > 2 individuals are "equals" or "the same" or "equivalent" in a owl
> > ontology. Sample:
> >
> >   <owl:Ontology rdf:about=""/>
> >   <owl:Class rdf:ID="Measure"/>
> >   <owl:Class rdf:ID="Unit"/>
> >   <owl:ObjectProperty rdf:ID="measures">
> >     <rdfs:range rdf:resource="#Measure"/>
> >     <rdfs:domain rdf:resource="#Unit"/>
> >   </owl:ObjectProperty>
> >   <Unit rdf:ID="meter">
> >     <measures>
> >       <Measure rdf:ID="distance"/>
> >     </measures>
> >   </Unit>
> >   <Unit rdf:ID="metre">
> >     <measures rdf:resource="#distance"/>
> >   </Unit>
> >
> > How can I test (or "discover") that "meter" and "metre" are the
> > same (cause it shares the same value to ObjectProperty, in this
> > case). Do i need a some closure at class definition? Can I use a
> > reasoner or SWRL?
> You might want to reconsider the current structure of your ontology,
> since I don't think it really will scale well as you expand it.
> In particular, the inference you suggest above (and in subsequent
> messages) doesn't seem like one that you will really want to use.
> Consider if you decide to introduce a new unit, the kilometer.
> Presumably this would then be modeled using
>   <Unit rdf:ID="kilometer">
>      <measures>
>        <Measure rdf:ID="distance"/>
>      </measures>
>    </Unit>
> But if you do this, it will then also share the distance Measure
> instance as the value of its "measures" property.  And by the
> inference rule you describe, your system will conclude/discover that
> "meter" and "kilometer" are the same as well.  I'm pretty sure you
> don't want that conclusion.
> A more robust and IMHO better ontology design would have only a
> single instance for any particular unit and associate text strings
> with that unit, thus allowing alternate mappings from the same
> canonical instance:
> <Unit rdf:ID="meter">
>      <measures>
>        <Measure rdf:ID="distance"/>
>      </measures>
>      <textstring rdf:datatype="
> XMLSchema#string">meter</textstring>
>      <textstring rdf:datatype="
> XMLSchema#string">metre</textstring>
>    </Unit>
> <Unit rdf:ID="kilogram">
>      <measures>
>        <Measure rdf:ID="distance"/>
>      </measures>
>      <textstring rdf:datatype="
> XMLSchema#string">kilogram</textstring>
>      <textstring rdf:datatype="
> XMLSchema#string">kg</textstring>
>    </Unit>
> and so on.
> _______________________________________________
> protege-owl mailing list
> protege-owl at
> Instructions for unsubscribing:

More information about the protege-owl mailing list