Re: are many URIs ultimately relative? was RE: are 'cid' URLs relative?

At 11:11 AM 2000-05-26 -0400, Jonathan Borden wrote:
>The point being that many commonly used URI schemes in fact employ context
>information for resource resolution and are in this sense relative.

Yes.  Not cid: or urn:, but plenty of others.  Is the following expansion
on this point consistent with why you are noting that?

The fact that a string conforms to the IETF requirements for a
URI-reference, or even an absoluteURI does not confer on the resource
identified nearly as much definition or stability as is desirable for a

If this is the point, I would agree, and hope others would agree.

While I am dubious that a suitably weak definition of "definition and/or
stability" can be crafted to properly serve as a global precondition for
namespaces [i.e. redefinable namespaces are possible, caveat_emptor], I
also agree that for well defined and stable namespace, a well defined and
[finitely] stable binding between "name collection identifier" and
"name-associated definitions as appropriate" is in order.  And this should
not be just any URI.

The Namespaces Rec is half a loaf in this regard because it only discusses
the consumption of namespace-marked XML instances, it does not discuss in
any way the creation or publication of namespace-usable markup
vocabularies.  Had the 'create namespace' and 'create conforming instance'
transactions been investigated more thoroughly, we might well have a
requirement for an 'fpi' URN in this spot, or something that works a lot
like that would.

Is the ISO FPI class of names registered yet, either as a recognized naming
authority under URNs or as a scheme under URIs?  

It would seem from this discussion that URI usage a) includes stuff you
wouldn't want to use and b) is arcane enough, so that some sort of BCP
document on how to name and document your namespace would be in order.


Received on Friday, 26 May 2000 12:03:08 UTC