Re: INSEE releases OWL ontology and RDF data for geographical entities

Carl,

Le vendredi 04 août 2006 à 11:04 -0400, creed@opengeospatial.org a
écrit :
> I may be out in left field, but I was wondering (RDF aside), what is the
> relationship of this activity with other similar geo-tagging activities,
> such as http://geotags.com/geo/draft-daviel-html-geo-tag-00.html? 

I am probably not the best person to answer this question but if nobody
else does answer, I can give my very personal view on this project...

The INSEE is the French National Institute for Statistics and Economic
Studies. It is in charge of designing and maintaining a number of
statistical classifications in the economic, social and spatial area
(see
http://www.insee.fr/en/nom_def_met/nomenclatures/nomenclatures.htm).

One of these classifications is the COG (Code Officiel Géographique)
which is a repertory of all the French geographical administrative
entities. This repertory has been available on the Web for several years
in formats such as dbf and text files. More recently, it has also been
partially published as enumerations in a W3C XML Schema schema
(http://xml.insee.fr/schema/cog-enum.html#Code_R%E9gion_stype).

Bernard Vatant who needed this kind of information for a number of
Modeca projects came with the idea to translate these data into an OWL
ontology.

This ontology has been developed to be conform to the existing data
model of the COG with the assumption that it could be useful for the
community. It can be seen as a new expression of the COG but shouldn't
be seen as an attempt to propose a universal data model. 

Another facet of this project is that I have insisted to define a
recommended XML serialisation for the description of the entities that
keeps the "RDF tax" as low as possible so that these descriptions can
also be used by XML tools (that's my job since I am participating as a
consultant to a number of XML project for the INSEE).

And yes, this is the same old dream (or should I say error?) that we had
for RSS 1.0...

With the help from Bernard who has used all the possibilities of the
XML/RDF syntax, we have been able to produce very clean XML and, for
instance, the definition of the city "Anet" is expressed as:

<rdf:RDF xml:base="http://rdf.insee.fr/geo/"
         xmlns:geo="http://rdf.insee.fr/geo/"
         xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
         xmlns:dc="http://purl.org/dc/elements/1.1/"
         xmlns:rdfs="http://www.w3.org/2000/01/rdf-schema#">
<geo:Canton rdf:about="CAN_2801">
  <geo:code_canton>2801</geo:code_canton>
  <geo:nom xml:lang="fr">Anet</geo:nom>
  <geo:chef-lieu rdf:resource="COM_28007"/>
  <geo:subdivision>
    <geo:Commune rdf:about="COM_28001">
      <geo:code_commune>28001</geo:code_commune>
      <geo:nom xml:lang="fr">Abondant</geo:nom>
      <geo:voisin rdf:resource="COM_28064"/>
      <geo:voisin rdf:resource="COM_28098"/>
      <geo:voisin rdf:resource="COM_28267"/>
      <geo:voisin rdf:resource="COM_28321"/>
      <geo:voisin rdf:resource="COM_28375"/>
      <geo:voisin rdf:resource="COM_28377"/>
      <geo:voisin>
        <geo:Commune rdf:about="COM_27391">
          <geo:code_commune>27391</geo:code_commune>
          <geo:nom xml:lang="fr">Marcilly-sur-Eure</geo:nom>
        </geo:Commune>
      </geo:voisin>
      <geo:voisin>
        <geo:Commune rdf:about="COM_27543">
          <geo:code_commune>27543</geo:code_commune>
          <geo:nom xml:lang="fr">Saint-Georges-Motel</geo:nom>
        </geo:Commune>
      </geo:voisin>
    </geo:Commune>
  </geo:subdivision>
  <geo:subdivision>
    <geo:Commune rdf:about="COM_28007">
      <geo:code_commune>28007</geo:code_commune>
      <geo:nom xml:lang="fr">Anet</geo:nom>
.
.
.

    </geo:Commune>
  </geo:subdivision>
.
.
.
</geo:Canton>
.
.
.
</rdf:RDF>

One of the things I find very elegant is the use of xml:base which
allows to use rdf:about/rdf:resource almost like traditional id/idref
attribute without leading hashes in rdf:resource.

Hope this helps and gives more context of this project.

Eric

-- 
GPG-PGP: 2A528005
                         Generated by Signify v1.14.
                For this and more, visit http://www.debian.org/
------------------------------------------------------------------------
Eric van der Vlist       http://xmlfr.org            http://dyomedea.com
(ISO) RELAX NG   ISBN:0-596-00421-4 http://oreilly.com/catalog/relax
(W3C) XML Schema ISBN:0-596-00252-1 http://oreilly.com/catalog/xmlschema
------------------------------------------------------------------------