RE: What to do about namespace derived URI refs... (long)

> > The key point here is that (a) we use URNs rather than
> > URLs to identify namespaces and abstract resources, [...]
> 
> Maybe, but note that this is only a "problem" where there is an
> overlap in the very concept of what a namespace "is". RDF treats it
> differently because it concatenates names onto the end to from a
> URI/URI-Reference... there's no problem at all to other RDF
> applications where a URI for a term is specified somwhere, even if
> this URI is also a URL. For example, the XSD specification tell you
> what URI to use to identify the terms within its namespace; no
> problem.

But if two different schema encodings are used to reify the same
concepts, belonging to the same conceptual namespace, but which have 
different URI fragment syntaxes, then the same RDF statements which
are valid according to one schema encoding are not valid according
to the other.

The point is that namespace URIs do not reliably define a single
MIME content type, and if the formula {namespace URI} + {fragment ref}
is supposed to provide a universal reference for the concept
reified by some schema, it will not work, because the {fragment ref}
portion may vary according to the encoding used by the schema in
a particular environment. 

I might be using XML Schema in one environment, a DTD in another,
a custom schema in a third, etc. and all might have different fragment
syntax (or none at all), so just how do I give a single unified
identity (URI) to the same concept in all three environments which
will work for all. I can't. Because RDF is dependent on the MIME
content type to create the URI ref.

And then, even if I use a URN, or tag URI, or HRN, or some other 
abstract non-URL URI for the concept, I still have to map from the 
schema defined identity to that generalized identity.

Patrick

Received on Wednesday, 6 June 2001 10:49:24 UTC