RE: Need compelling story on the value of ontologies in fusing location-based data

> The example that I would love to show is this: suppose that document 1
> has data for a particular region, expressed as a "bounding box":
>
>    <rdf:Description>
>        <region>
>           <top-left-coordinate>
>               <lat>73.000</lat>
>               <lon>123.000</lon>
>           </top-left-coordinate>
>           <bottom-right-coordinate>
>               <lat>75.000</lat>
>               <lon>125.000</lon>
>           </bottom-right-coordinate>
>        </region>
>        ... document 1 data ...
>    </rdf:Description>
>
> And document 2 contains data for an area that is within the bounding
> box:
>
>    <rdf:Description>
>        <latititude>74.000</latitude>
>        <longitude>124.000</longitude>
>        ... document 2 data ...
>    </rdf:Description>
>
> It would be *extremely* cool if, by consulting an OWL Location ontology,
> an application could recognize that the location specified in document 2
> is located *within* the bounding box specified in document 1.  Such an
> example would get a *lot* of people in my community very excited about
> OWL and RDFS.  Unfortunately, I can think of nothing in OWL or RDFS that
> would help with this (I am eager to be proven wrong).  /Roger

An application could easily do just that - it's only simple arithmetic after
all. Consider the application a service, orthogonal to the logic. The
difficult part is expressing the information in an unambiguous fashion that
could be shared amongst the applications. As well as the input data you have
here, presumably you would need a way of expressing the result returned from
a SWS, A isInside B or whatever - thanks RDF/XML. To reason with this you'd
presumably need to be able to state things like {X isInside A} disJointFrom
{Y isInside A} - thanks OWL. The good bit is that none of this system is
hard-wired (apart from the specific service), it's all entirely pluggable
using declarations.

I don't know whether this *looks* extremely cool ("I could do the same in
one line of Python"), but I reckon it is ;-)

Cheers,
Danny.

Received on Tuesday, 24 June 2003 09:47:23 UTC