[
Lists Home |
Date Index |
Thread Index
]
Bullard, Claude L (Len) wrote:
>
> But that doesn't fix the problem. As I said before,
> the issue comes up in the context of RDF attempting
> to use URIs for one to one mappings. It would seem
> that if they are using a name which must map, they
> must be responsible for specifying the selector
> mechanism which as you point out on the Web, is
> the protocol technology (eg, HTTP). What does
> RDF do at that stage of identifying?
RDF -- the new version of 12 Nov 2002, not the original --
handles URIs fairly sensibly, I think.
In the new RDF model, URIRefs are atomic, logical constants;
no assumptions are made about the nature of resources [*].
An RDF graph is just a graph; it has no intrinsic meaning
beyond the graph structure. URIs label nodes and edges,
nothing more.
Any additional meaning ascribed to the graph comes from an
"X-interpretation," for various values of X, and, for the
most part, "X-interpretations" are Outside the Scope of
This Document.
[*] This is somewhat true of the original M&S Rec as well,
but reading the original I get this persistent, nagging
feeling that an RDF triple *must* be saying *something*
about *something*. I don't get that feeling reading
the new version. Less metaphysics, more math; a big
improvement.
--Joe English
jenglish@flightlab.com
|
