Re: (Round 2) Proposed Extensions to OWL

[Roger L. Costello

> I will take a stab at it.  Let me take a complete example.  Suppose that
> I want to express this: The length of the Yangtze River is 6300
> kilometers.  Here's how I think it would be expressed using the above
> triple:
>
> <River rdf:ID="Yangtze">
>     <length>
>         <LengthMeasure>
>             <transform>
>                 <LengthInKilometers>
>                     <value>6300</value>
>                 </LengthInKilometers>
>             </transform>
>         </LengthMeasure>
>     </length>
> </River>
>
> Read this as: "The value of length is a resource of type LengthMeasure.
> LengthMeasure has a transform to LengthInKilometers.  The
> LengthInKilometers is 6300."
>
> Do you agree that this is a faithful XML expression of the above
> triple?  If not, how would you express it?  /Roger

Not quite.  My original version had two properties for the  transform - its
type and its value.  Your version has the actual numerical value being a
property of the transform instead of being a property of the actual physical
quantity.  The transform exists as a mapping independently from any
particular value, so the 6300 should not be one of its properties.

Here is what I think is a faithful reflection of the original intent (modulo
namespaces, of course) -

<River rdf:ID="Yangtze">
    <length-measure>
        <value>
            <transform rdfs:type='#LengthInKilometers'/>
            <number>6300</number>
        </value>
    </length-measure>
</River>

Notice how the length-measure bnode (which could be given an ID of its own
to remove its "b-ness") is a resource in its own right, so you could make
other statements about it.

As I think about it, I think the transform bit would better be written like
this instead -

<transform rdf:resource='#LengthInKilometers'/>

Notice in this version that the transform is referred to as actually being
another resource.  The idea is that each of these transforms would be an
object or concept in its own right - like a singleton - it gets defined once
and reused as often as necessary.  I think this approach would be
preferable, but either way would work OK.

Clear?  Comments?

Cheers,

Tom P

Received on Monday, 30 June 2003 10:45:58 UTC