Date: Fri, 16 Jul 1993 13:12:10 -0400 (EDT)
From: Rob Raisch <raisch@ora.com>
Subject: Re: URNs in the DNS
To: Michael Mealling <ccoprmm@oit.gatech.edu>
In-Reply-To: <199307161422.AA17937@oit.gatech.edu>
Message-Id: <Pine.3.03.9307161308.A1469-b100000@amber.ora.com>
On Fri, 16 Jul 1993, Michael Mealling wrote:
> > Perhaps we can get away with something as simple as this...
> >
> > URN-TO-URL-MAPPING:
> > URN: // as per current internet draft
> > Multiple occurrences of the following block of fields
> > URL: // physical address of object
> URM: // CACHED but not permanent metainformation
> > end of multiple occurrences
> > END URN-TO-URL-MAPPING
>
> Note the precedence that this structure generates. A URN 'owns' each URL
> below it and a URL 'owns' each URM below that. What you then can do is put
> a new URN at the end of that template with it's respective list of
> URLs/URMs. This would then give you some way to denote a relationship
> between URNs. You could have a template that was designated as all URNs for
> each version of a document, each with it's own URN. Then give that template
> it's own URN as a pointer to the whole set.
Ummmm.... Perhaps I should mention that the way Registrar is set up is almost
a one-for-one match with IAFA templates. What this means is that there is a
template for each URN, and each template contains a list of URL records, as
well as mucho other meta information. The only reason I wrote registrar to
begin with is that whois++ is, (as far as I am aware), currently vapourware,
and I felt it was really important for people to actually see something which
approached the problem.
</rr>