Re: URN single or multiple variants (was: four-part harmony?)
Simon E Spero <ses@tipper.oit.unc.edu>
Message-id: <9310022206.AA29200@tipper.oit.unc.edu>
To: winograd@interval.com (Terry Winograd)
Cc: jak@violet.berkeley.edu (John A. Kunze), uri@bunyip.com,
www-talk@nxoc01.cern.ch
Subject: Re: URN single or multiple variants (was: four-part harmony?)
In-reply-to: Your message of "Sat, 02 Oct 93 10:35:45 -0800."
<9310021735.AA29337@interval.interval.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Date: Sat, 02 Oct 93 18:06:36 -0400
From: Simon E Spero <ses@tipper.oit.unc.edu>
[The Terry Winograd?]
The thing that has become apparent during the course of the entire URI
discussion is that there are certain parts of the problem for which there
is an obvious solution, and others which seem to be "open research problems"
- i.e. flame-bait.
URLs were pretty straightforward, since there was already a large body of
code in-place proving their viability.
URNs were slightly more troublesome, since people tended to overload different
meanings that were completely different to everyone elses. The current
definition is extremely minimalist; adding anything will break the rough
(way rough) consensus.
The definitions of sameness from allocation domain to allocation domain;
each allocation domain is only allowed one viewpoint. If an allocating
body wishes to use a different comparator, then a different allocating
domain should be used.
One idea might be to define an IAFA-style template to describe allocating
domains and bodies; we could have an equality-definition field, and
some standard definitions for use with that field. We could also have templates
for defining the definitions (this is getting recursive!)
Simon