No subject


Sun Dec 20 23:06:54 CET 2009


allocated in both RDF/XML and RDFa, and is used to hold a direct URI
identifier for the precise thing being described at that point in the
markup.

<p xmlns:foaf="http://xmlns.com/foaf/0.1/" xmlns:dc="http://purl.org/dc/terms/"
   typeof="foaf:Person" about="http://danbri.org/foaf.rdf#danbri">
  <span property="foaf:name">Dan Brickley</span>
  <span rel="foaf:made">
    <span typeof="foaf:Document" resource="http://danbri.org/">
      <span rel="dc:subject"
resource="http://example.org/geek-thesaurus#danbri-concept">
        <span rel="foaf:isAbout" resource="http://danbri.org/foaf.rdf#danbri"/>
      </span>
    </span>
  </span>
</p>

or in RDF/XML,

<rdf:RDF  xmlns:dc="http://purl.org/dc/terms/"
  xmlns:foaf="http://xmlns.com/foaf/0.1/">
  <foaf:Person rdf:about="http://danbri.org/foaf.rdf#danbri">
    <foaf:made>
      <foaf:Document rdf:about="http://danbri.org/">
        <dc:subject>
          <rdf:Description
rdf:about="http://example.org/geek-thesaurus#danbri-concept">
            <foaf:isAbout rdf:resource="http://danbri.org/foaf.rdf#danbri"/>
          </rdf:Description>
        </dc:subject>
      </foaf:Document>
    </foaf:made>
    <foaf:name>Dan Brickley</foaf:name>
  </foaf:Person>
</rdf:RDF>

I think naming a class or property so closely to one of the core
syntactic constructs in RDF/XML and RDFa is a recipe for confusion
(especially as this area is self-referential enough and confusing
already).

I don't think foaf:it is particularly transparent, <span rel="foaf:it"
resource="http://danbri.org/foaf.rdf#danbri"/> ... and in fact it
might be downright cryptic, but the conceptual machinery needed for
any of this is pretty heavy; not just understanding RDF classes and
properties; and then SKOS's semi-parallel model of concept
hierarchies; but how they relate to each other.

<skos:Concept>
  <skos:prefLabel>Bill Gates</skos:prefLabel>
  <foaf:it>
    <foaf:Person>
     <foaf:name>Bill Gates</foaf:name>
    </foaf:Person>
  </foaf:it>
</skos:Concept>

or

<skos:Concept>
  <skos:prefLabel>Bill Gates</skos:prefLabel>
  <foaf:isAbout>
    <foaf:Person>
     <foaf:name>Bill Gates</foaf:name>
    </foaf:Person>
  </foaf:isAbout>
</skos:Concept>

or even ...

<skos:Concept>
  <skos:prefLabel>Bill Gates</skos:prefLabel>
  <foaf:skos_concept_for>
    <foaf:Person>
     <foaf:name>Bill Gates</foaf:name>
    </foaf:Person>
  </foaf:skos_concept_for>
</skos:Concept>


It's odd but I think often in RDF vocab design, the biggest stumbling
point is finding a good name. If there was a natural good name (in
English) the property would exist already!

Any other name suggestions?

Re definitions: should it be owl:FunctionalProperty? I'd say not
explicitly, although typically a single value makes sense, these
mappings are naturally messy and so I'd stay quiet on this for now
while we get more deployment experience.

Should it be usable with classes as values? I think this might be
useful. Lots of concepts in thesaurus systems are class-like.

so: range: Thing, domain, skos:Concept. IFP: no, FP: no

Leigh do you have some examples we could put in the wiki as
documentation, once a name is fixed?

cheers,

Dan


More information about the foaf-dev mailing list