Message-Id: <9310270133.AA01487@expresso.bunyip.com>
From: Peter Deutsch <peterd@bunyip.com>
Date: Tue, 26 Oct 1993 21:33:17 -0400
In-Reply-To: Mitra's message as of Oct 23, 1:03
To: mitra@path.net (Mitra), uri@bunyip.com
Subject: Re: URN definition -- a way out?
[ You wrote: ]
>
> I agree with Dirk that it would be a big mistake either to abandon
> "sameness" as property of URN's or to define what "sameness" means.
>
> However Dirk, you are reopening a decision made a long time ago. URN's do
> not contain ANY meta information. When we had meta information in URN's
> and URLs we had endless arguments as to which meta information they should
> contain. This was solved (to everyone's satisfaction I think) by removing
> ALL meta information from URN's and putting it in the URM/URCs. With URNs
> coming closer to consensus we are only now starting to define what
> URMs/URCs look like.
Hear! Hear!
I think this divorcing of meta-information from the URN
was one of the cleverer things we agreed to do in the past
year or so, primarily because I think it illustrates the
value of the divide and conquer approach. Find the minimal
subset you agree on and get something out. Tempus Fugit
and all that.
There are people with real projects that will use URNs and
URLs and that are waiting for things to gel. We don't need
or want to wait for the debate on meta-info to congeal
before deploying something. We've tried to be good scouts
and not muddy the water with competing URL formats or
functionality that lacks a wider consitutency but we do
want to get on with things at this point.
My own personal decision to accept (despite all my
reservations) the current URL document was made for the
same reason. I'd rather see _something_ out there, however
flawed I might believe it to be, so that we can get
experience with it. The easiest way I see to do that is to
take a salami approach (ie. do it one slice at a time).
- peterd
--