- From: Larry Masinter <masinter@attlabs.att.com>
- Date: Thu, 22 Jun 2000 08:53:05 -0700
- To: "John Cowan" <jcowan@reutershealth.com>, "Rick JELLIFFE" <ricko@geotempo.com>, <xml-uri@w3.org>
> > The thing I find difficult is that "resource" seems to mean "anything
> > that has a URI"
>
> Just so.
There are resources that currently have no URI; the process of devising
a new scheme (RFC 2717 "Registration Procedures for URL Scheme Names",
RFC 2718 "Guidelines for new URL Schemes") is a process of creating new
ways of naming resources.
In natural language, it is common to use noun phrases to refer to abstract
concepts ("Honesty", "the color green") as well as concrete noun phrases
which have no physical referent ("the king of the USA"), and even
nonsensical
abstractions ("green ideas"). We're not very troubled with the distinction
between 'is there a name for it?' and 'does it exist?'.
I'm not sure why we're having trouble with similar distinctions between
a URI and the resource that might be identified by it. Words and phrases
are coined to name the previously unnamable, just as we invent URIs and
URI schemes.
The namespace specification chose to use resource identifiers as namespace
names; the philosophical debate we're having is about what inferences
you can draw about the relationship of the namespaces if you are given
some assertions about the relationship of the resources.
RDF cannot make assertions about resources except by using resource
identifiers. It is useful to be able to use RDF to make assertions about
namespaces. Can one use the namespace name as the resource identifier
used to make assertions about namespaces?
Larry
--
http://larry.masinter.net
Received on Thursday, 22 June 2000 11:54:15 UTC