From: dgd@cs.bu.edu (David Durand)
Date: Tue, 14 Sep 93 14:24:40 -0400
Message-Id: <9309141824.AA03609@csd.bu.edu>
To: uri@bunyip.com
Subject: Re: URL/URN Implementation Status (query)
The last time URNs came up I thought that there was pretty good consensus
on a number of those "fundamental issues"
> I don't think we've solved the fundamental problems of who and how
> will URN -> URL translation service be offered, or the fundamental
> question of whether URNs can point to mutable objects, immutable ones,
> or whether there are two kinds.
> (I.e., can the referent of a URN change depending on who you are,
> where you access it from, or when you access.)
> (IMHO) because there isn't an obvious 'right' answer, and every design
> choice has some advantages and disadvantages.
I thought that the point of URNs as opposed to URL was that they
don't change depending on where you are, when you access, or where you
come from. Allow any significant change from this as you've just got
URLs again: non-universal, short-lifetime IDs.
The interesting and unresolved issues are whether some kind of
version access is OK, like a URN to the current draft of X. This would
have to be supplemented with a way to get a particular version of
X. There's also the question of data format: What sorts of translation
give you the "same" document?
Am I way off base? I'm behind in some of the back list traffic.
Also is there a draft URN spec. floating about?
-- David Durand
BU Computer Science