From: brennan@hal.com (Dave Brennan)
Message-Id: <9402102123.AA12857@hysteria.hal.com>
Subject: Re: URN functional spec
To: uri@bunyip.com
Date: Thu, 10 Feb 94 15:23:57 CST
In-Reply-To: <CKz0qq.8HA@pandora.sf.ca.us>; from "Mitra" at Feb 9, 94 6:59 pm
> I believe those naming authorities which use accessible names will see
> a much faster uptake of those URN's while those using things like MD5
> checksums will see them ignored in favour of URLs.
Probably true, but what makes one URN more "accessible" than another if
they both, strictly speaking, need to be resolved before they can be used?
Here's an example of a URN-like value I'm using today: UQ3zf5B-BM9X3cS.
The value itself tells me nothing about the document to which it refers,
but all I have to do is feed it to the document browser software to view
the thing to which it refers. IMHO, this is much better than something
like "/usr/local/lib/doc/foo.txt," which although has more meaning to me
as a human, is wide open to abuse. Even though the latter value needs
resolving if we call it a URN, it is far too tempting for a user to go
looking for the document as though it were in what looks like a path even
though it may not be (or have never been) there.
So I'm all for discouraging URN values with "explicit semantics."
While they might be good in the short-term, there's just too much
potential for trouble.
-- Dave Brennan HaL Computer Systems brennan@hal.com Austin, TX, USA