Lists Home |
Date Index |
Didier PH Martin wrote:
> ... If this is the
> case, I am having tremendous difficulties to see where the linkage between
> the resource and its ontology is located?
You are not alone in this. There has been discussion of this just this
past week on the RDF lists. One camp wants all RDF resources to be
indicated with a dereferenceable URL - the URL would provide information
about the resource, somewhat like RDDL does. Another camp is worried
about confusing the abstract meaning of a URL with the dereferenced
resource, and so thinks that no such URI should actually point to anything.
> At the end of its namespace URI?
> If yes, the semantic web starts to be operational. A resource would then
> a) A classification of its information units
> b) A reference to the model behind this classification.
> Is there any talks in that direction?
One thing that comes to mind is Patrick Stickler's URIQA approach
(unfortunately the Nokia server has been or is being moved and doe snot
seem to work right now).
URIQA has two parts -
1) a new header for an HTTP request, to signal that special information
is wanted about the resource
2) A subset of an RDF database - namely, all information known about the
resource (where the resource is the subject of an RDF statement) - but
only to a very limited depth. Of course, this information is still in
the form of RDF triples, but you could discover a resource's
classification this way.
Topic Maps have the so-called Public Subject Indicator. The author of a
PSI description could discuss the classification scheme and the model in
the referenced uri, but there is no standard way to make such
descriptions machine-understandable at present.