URNs (Was: New Anchor attributes)

Carl Johan Berglund (f92-cbe@nada.kth.se)
Tue, 28 May 1996 09:58:26 +0200


Message-Id: <31AAB222.5586@nada.kth.se>
Date: Tue, 28 May 1996 09:58:26 +0200
From: Carl Johan Berglund <f92-cbe@nada.kth.se>
To: Abigail <abigail@tungsten.gn.iaf.nl>
Cc: www-html@w3.org
Subject: URNs (Was: New Anchor attributes)

Abigail wrote:
> I think a more general object naming scheme could solve that.
> Currently, one has to give the location of the server, and the
> address of the object on the server. A more general naming scheme
> could introduce an extra mapping layer. For instance:
> HREF = "xxx:CPAN" would point to the nearest mirror of the CPAN
> archive. It might even take unreachable servers into account.
> Could URNs deal with this?

As far as I have heard, this is exactly what URNs are for.
According to what Keith Moore from the University of Tennessee
said in Stockholm, 9 May, URNs (Uniform Resou are made to be stable 
names for a specific resource. Given a URN, you (or your browser) 
ask some cataloging service (Similar to DNS, I suppose) for a
URC (Uniform Resource Citation). The URC contain information 
about title, author, etc. and also a list of URLs, where the
resource can be found. If the list contains an URL in Sweden,
I wouldn't let my browser get it from New Zealand. If the 
Swedish server should be busy, my browser could ask a server
in France.

URNs look like "URN:name-space:stuff", for instance: 
"URN:isdn:1565921690" or "URN:xxx:CPAN". I don't think it is 
very wise to put this functionality into HTML itself, given 
this URN/URC sketch.

It would be nice to hear from those people who are working
with URNs and URCs, to hear how the work is progressing.

Maybe we shuld move this thead to www-talk, though.

-- 
Carl Johan Berglund <f92-cbe@nada.kth.se> 
http://www.student.nada.kth.se/~f92-cbe