> On Tue, 2003-06-03 at 16:44, Roland Schwaenzl wrote:
>
> > > "The class of LCSH headings were issued July 11, 2000",
> which is not
> > > true at all. Thus, Pete correctly notes that we have a case of
> > > conflicting semantics here!
> >
> > There could be many classes, whose class extensions
> contains things,
> > which have LCSH headings as literal values by means of the
> property,
> > RDF denotes with rdf:value.
> >
>
> So a possible interpretation would be "*this* class of LCSH
> headings was issued...". If this statement specifically does
> not extend to the instances, I guess I can buy it...
Crumbs.... If understand that correctly, I find it almost a subtlety too
far!
But I think I'm just about prepared to accept that in some way DCMI
issued that class of subjects.
> Still, in other cases this might not work. There is a
> difference between the term and what it denotes.
Yes, I agree there is a difference.
The URI http://purl.org/dc/terms/LCSH (which is what I think you are
calling the "term" - is that right please?) is not the same as the class
of subjects denoted by that URI.
However, I think maybe in Tom's earlier message he was using the term
"term" (!) slightly differently, i.e. Tom was using it in the sense DCMI
uses it here
http://dublincore.org/usage/documents/principles/
where a DCMI term is either
- a DCMI Element (a resource of type rdf:Property)
- a DCMI Element Refinement (a resource of type rdf:Property)
- a DCMI Encoding Scheme (a resource of type rdfs:Class whose members
provide values for an rdf:Property)
- a specific term/value of one of those vocabularies.
i.e. so a term in the sense used by Tom - a DCMI term - is _not_ a
URIref: it's a resource of type rdf:Property or rdfs:Class or of type X
(where X is a DCMI Encoding Scheme), that is identified by a URIref.
Maybe I'm hopelessly trying to square a circle (probably!)... but on
this basis I can continue to think of the resource
http://purl.org/dc/terms/LCSH as both
- an rdfs:Class (which denotes a class of subjects) _and_
- a type of "DCMI term" (specifically a "DCMI encoding scheme").
Pete
|