On 2002-05-30 23:46, "ext Wagner,Harry" <[log in to unmask]> wrote:
>> At the moment, I think (but I'm not completely sure!) we're
>> saying there
>> are circumstances in which we _would_ like an RDF application
>> to be able
>> to distinguish terms which in XML are associated with the namespace
>> named http://purl.org/dc/terms/ from terms which in XML are associated
>> with the namespace named http://purl.org/dc/elements/1.1/ ?
Well, of course, RDF would differentiate between the two, only because
the two namespaces will result in different URIs if concatenated with
the same local name. I.e.
"http://purl.org/dc/terms/creator"
"http://purl.org/dc/elements/1.1/creator"
but the namespace used in the XML is completely discarded by
the RDF parser. You could achieve the same results with
<ator xmlns="http://purl.org/dc/terms/cre">...</ator>
<or xmlns="http://purl.org/dc/elements/1.1/creat">...</or>
Insofar as RDF is concerned, they are semantically identical
as they result in precisely the same URIs in the RDF graph.
The fact that "http://purl.org/dc/terms/" is also being used
to denote a specific ontology seems to be the point of
confusion here -- that somehow, its use as a namespace in
the XML serialization is expected to result in some automatic
relationship being drawn between the term and the ontology.
But the reality is that XML and RDF have different naming
models, and RDF does not subscribe to the full XML naming
model. It discards all partitioning between namespace and
local name as well as the context of the local name (element,
element-specific attribute, or global attribute). RDF simply
uses namespaces as a means to an end, a hack really, to get
URIs into its graphs.
XML applications operate on qnames. RDF operates on URIs.
Thus folks should not look to RDF to maintain any significance
attributed to any namespace URIs occurring in the XML serialization.
Namespaces simply are not part of the RDF graph model and have
no meaning or realization there.
If there is a relation between a term and some resource, such
as a schema, ontology, whatever, then you have to define it
explicitly with rdfs:isDefinedBy. RDF will not grok it from
the namespaces occurring in the RDF/XML.
> It would be a lot simpler if RDF provided a 'schema' object... harry
Well, folks are free to define an ontology that provides such
a class. RDF itself is intended to be application neutral. But
you could certainly define or adapt a taxonomy of RDF classes
for differentiating the values of the rdfs:isDefinedBy property,
or alternately, define subproperties of it which carry the
additional semantics, e.g. foo:isDefinedByRDFSchema, etc.
Cheers,
Patrick
--
Patrick Stickler Phone: +358 50 483 9453
Senior Research Scientist Fax: +358 7180 35409
Nokia Research Center Email: [log in to unmask]
|