Message-Id: <199402172148.QAA14065@wilma.cs.utk.edu>
From: Keith Moore <moore@cs.utk.edu>
To: Martin Hamilton <M.T.Hamilton@lut.ac.uk>
Subject: Re: Finding URN->URL servers
In-Reply-To: Your message of "Thu, 17 Feb 1994 15:44:22 GMT."
<199402171544.PAA10693@lust.mrrl.lut.ac.uk>
Date: Thu, 17 Feb 1994 16:48:31 -0500
>
> / Why don't we just use whois++ for this purpose? Then we can just registe
> r
> / a URN.ORG root and use ordinary A records:
>
> How about urn.int, or perhaps just urn? :-)
Fine with me, but we gotta get the powers that be to agree to it.
> Let's see... use of TXT (or MX, or UR-RR :-) records would give us a
> means of locating URN servers that's protocol independent (dare I say
> future proof?), with multiple independent and widely deployed
> implementations of URN->URL convertors (e.g. gopherd and httpd!).
>
> Sounds good to me!
I'm dead set against using MX records or overloading any existing type in a
way that's contrary to its original design. TXT is okay. I like A records
because they would save the extra step to look up the host name, and because
it's so easy to use gethostbyname() to do that part of the lookup.
It would take too long to define a new record type. If A records aren't
sufficient, we should use TXT and be done with it.
Keith