Re: Persistent Documents and Locations

Brian Behlendorf (brian@organic.com)
Mon, 21 Aug 1995 22:38:57 -0700 (PDT)


Date: Mon, 21 Aug 1995 22:38:57 -0700 (PDT)
From: Brian Behlendorf <brian@organic.com>
To: "Karen R. Sollins" <sollins@lcs.mit.edu>
Cc: lazear@dockside.mitre.org, uri@bunyip.com
Subject: Re: Persistent Documents and Locations
In-Reply-To: <199508212321.TAA10110@lysithea.lcs.mit.edu>
Message-Id: <Pine.SGI.3.91.950821222205.600E-100000@fully.organic.com>

On Mon, 21 Aug 1995, Karen R. Sollins wrote:
> Well, no, I think you've misunderstood RFC 1737.  The intention for
> URNs is that they are valid in perpetuity.  If the object is still
> around, in 10 yrs, then the URN for it is still "valid".  I can't go
> into a long diatribe now, but there is a great deal behind the choice
> of making URNs global and long-lived.  But, suffice it to say that
> URNs don't expire or become invalid.   The resources they name may be
> deleted, but the URN should never be re-used or reassigned.  For each
> URN the assigment of it happens no more than once, ever.

And under normal circumstances, URL+Last-Modified time supply this, as Larry
pointed out.  Why not use this as a URN scheme?  It would seem to satisfy 
most of the URN requirements, and provide some interesting new 
functionality (hey, show me what this page looked like on this old 
date...), while having an incredibly easy transition.  Thoughts?

	Brian

p.s. - by "normal" I mean sites where Last-Modified isn't screwed up, as 
per recent discussions on http-wg.

--=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=--
brian@organic.com  brian@hyperreal.com  http://www.[hyperreal,organic].com/