Re: geo ontology

From: David Huynh <dfhuynh_at_csail.mit.edu>
Date: Wed, 03 Aug 2005 21:07:38 -0400

Bruce D'Arcus wrote:

> Been playing with the new frbr ontology and representing
> subjects/topics. To wit, I went looking for an ontology I could use
> to represent place (I'm a geographer, so no surprise!), and found one
> from SIMILE!
>
> I have a few minor comments/questions:
>
>> geo:province
>> a rdf:Property ;
>> rdfs:label "province/state"_at_en ;
>> rdfs:comment "the state/province that includes the place"_at_en ;
>
>
> I know "state" is fairly U.S. centric, but isn't province also rather
> narrow?
>
> I don't have a better suggestion though.
>
>> geo:site
>> a rdf:Property ;
>> rdfs:label "site/city"_at_en ;
>> rdfs:comment "the site/city that includes the place"_at_en ;
>
>
> A site can be a point on the ground. Why not "city" (also?)?
>
> A broader question that may reflect my ignorance of RDF: why are these
> properties and not classes (e.g. subclasses of Place)?
>
> Finally, any examples of using this with SKOS?

It was a moment of "how can I most quickly show off Piggy Bank + Google
Maps?"... and I couldn't find any well-documented generic ontology for
geographic locations. And I'm not a geographer :-) So, I'm sure the
resulting ontology isn't that great.

I'd love to have a standard ontology for that to tell scraper writers to
use for recording addresses. In the same line of thoughts, perhaps this
(through PB scrapers) is how we get most people to end up with data in
the same set of ontologies for common things such as addresses, dates,
events, names, jobs,...

David
Received on Thu Aug 04 2005 - 01:02:09 EDT