To: M.T.Hamilton@lut.ac.uk, hoymand@joe.uwex.edu, J.P.Knight@lut.ac.uk,
In-Reply-To: M.T.Hamilton@lut.ac.uk's message of Thu, 17 Mar 1994 14:24:33 -0800 <94Mar17.142434pst.2732@golden.parc.xerox.com>
Subject: Re: Stab in the dark
From: Larry Masinter <masinter@parc.xerox.com>
Message-Id: <94Mar17.150931pst.2732@golden.parc.xerox.com>
Date: Thu, 17 Mar 1994 15:09:29 PST
Any scheme which embeds a 'location' in a URN is really begging the
question; URNs need to be location independent, and in general, the
lifetime of a URN needs to be longer than the lifetime of the DNS
entry of the original author.
It is possible that using DNS to implement some part of URN -> URL
resolution can be made to work, but not if you tie the ownership of
URN name entry to the ownership of the locations that go with it.
I'm worried a bit about authority in these URN resolution schemes;
you'd like the URN to be secure (in that someone couldn't spoof
urn:/us/government/whitehouse/speeches/clinton/9-10-93/a and supplant
it with some other document).
I don't think we've yet addressed the security aspects of URNs,
although I believe there ARE security requirements.