Date: Tue, 19 Oct 93 13:18:48 -0400
From: Leslie Daigle <leslie@beethoven.bunyip.com>
Message-Id: <9310191718.AA06898@beethoven.bunyip.com>
To: uri@bunyip.com
Subject: forwarded message
-----Forwarded Mail-------
> From dgd@cs.bu.edu Sat Oct 16 11:27:49 1993
> Return-Path: <dgd@cs.bu.edu>
> Received: from mocha-int.bunyip.com by beethoven.bunyip.com (NX5.67c/NX3.0S)
> id AA04370; Sat, 16 Oct 93 11:27:47 -0400
> Received: from CS-MAIL.BU.EDU by mocha.bunyip.com with SMTP (5.65a/IDA-1.4.2b/CC-Guru-2b)
> id AA03205 (mail destined for leslie@beethoven.bunyip.com) on Sat, 16 Oct 93 11:38:55 -0400
> Received: from CSD.BU.EDU by cs-mail.bu.edu (5.61+++/Spike-2.1)
> id AA09346; Sat, 16 Oct 93 11:38:46 -0400
> From: dgd@cs.bu.edu (David Durand)
> Received: by csd.bu.edu (5.61+++/Spike-2.1)
> id AA02506; Sat, 16 Oct 93 11:38:43 -0400
> Date: Sat, 16 Oct 93 11:38:43 -0400
> Message-Id: <9310161538.AA02506@csd.bu.edu>
> To: uri-request@mocha.bunyip.com
> Subject: URN/URL discussions
> Status: RO
>
>
> I am always hesistant about weighing in here as I'm from outside -- I have
> no support to be part of IETF meetings, and as a university type, I'm not
> trying to build systems in the same nitty-gritty way. But...
>
> The distinction between URL and URN is really pretty clear. A locator is a
> recipe for how to find something on the net. For instance it may have
> relatively ephemeral (though essential) information like internet host names
> in it.
> A name is a permanent identifier for a thing that will not change depending
> on the thing's location. It may or may not change based on data formats and the
> like, but that's a subsidiary issue. If I'm doing distributed documents I want
> to refer to Moby-Dick, or some particular edition of it. The issues about
> which host to go to are separate.
>
> URNs attempt to resolve a data issue, not a programming issue. You don't
> _have_ to change URNs when data moves -- you do have to change URLs. When I
> first saw WWW the lack of something like URNs is what disturbed me.
>
> Of course we want to make it relatively easy to find URLs given a URN, but
> the fact is that they might not exist: the URN might refer to a document that
> is not currently available on the net, or that is redundantly available on
> the net (ie. pick the closest FTP site).
>
> Is this wrong, or have I really misunderstood things here?
>
> -- David Durand
> BU Computer Science
>