Re: The UR* scheme registry, Citing URL/URI specs

David G. Durand (dgd@cs.bu.edu)
Sat, 25 Oct 1997 15:08:56 -0500


Message-Id: <v03007800b0780021d22a@[205.181.197.103]>
In-Reply-To: <Pine.WNT.3.95.971025143511.-3977301J-100000@hazel.hpl.hp.com>
Date: Sat, 25 Oct 1997 15:08:56 -0500
To: Dave Raggett <dsr@w3.org>
From: dgd@cs.bu.edu (David G. Durand)
Subject: Re: The UR* scheme registry, Citing URL/URI specs
Cc: Tim Bray <tbray@textuality.com>, Dan Connolly <connolly@w3.org>,
        Keith Moore <moore@cs.utk.edu>,

At 1:37 PM -0500 10/25/97, Dave Raggett wrote:
>On Fri, 24 Oct 1997, David G. Durand wrote:
>
>> Location independence is really useful (demonstrated fact). The
>> counter-argument that location independent names can break is very
>> weak given that location dependent identifiers also break
>> regularly.
>
>But there is nothing to stop regular URLs being used as location
>independent identifiers. This is just a matter of how you use
>the URLs, e.g. you could use a directory service to find a copy
>without caring as to where the copy is held.

Amaxing how every surfacing of this topic results in recapitualtion of the
same onld arguments. The difference is perhaps best given by an analogy:

   We often cite books by their "names" A (title, author) pair, but
sometimes something more complex. The location-based naming scheme of
books, by library, floor, shelf and position would also work. Further, my
local librarian might be smart enough to take a copy of his shelf rather
than sending to the British Museum. That doesn't change the fact that the
_normative_ interpretations of the names are very different -- one is
satisfied only by a particular document, the other by whatever object
resides at a region of space.

   My librarian might be making a mistake if the BL changes buildings (as
they are in the process of doing now). Once they've changed buildings the
meaning of floor 1, shelf 2555, book 23 is different. The goal of URN
schemes is to insure that a social (and technical) infrastructure are in
place to prevent such changes of meaning insofar as possible, and to enable
the construction of name resolution mechanisms (whether social or
technical).

This is rather irrelevant to the question of what to call a locator that is
intended to be either a URL or URN. Martin's point that applications that
"just fetch things" should not care seems very sensible to me.

Are we getting closer to finding out if there's an official definition of
the term URI?

   -- David

_________________________________________
David Durand              dgd@cs.bu.edu  \  david@dynamicDiagrams.com
Boston University Computer Science        \  Sr. Analyst
http://www.cs.bu.edu/students/grads/dgd/   \  Dynamic Diagrams
--------------------------------------------\  http://www.dynamicDiagrams.com/
MAPA: mapping for the WWW                    \__________________________