Re: URN Usage

Martin Hamilton <M.T.Hamilton@lut.ac.uk> Fri, 17 September 1993 12:31 UTC

Received: from ietf.nri.reston.va.us by IETF.CNRI.Reston.VA.US id aa02682; 17 Sep 93 8:31 EDT
Received: from CNRI.RESTON.VA.US by IETF.CNRI.Reston.VA.US id aa02676; 17 Sep 93 8:31 EDT
Received: from mocha.bunyip.com by CNRI.Reston.VA.US id aa12323; 17 Sep 93 8:31 EDT
Received: by mocha.bunyip.com (5.65a/IDA-1.4.2b/CC-Guru-2b) id AA05150 on Fri, 17 Sep 93 05:31:30 -0400
Received: from avarice.lut.ac.uk by mocha.bunyip.com with SMTP (5.65a/IDA-1.4.2b/CC-Guru-2b) id AA05146 (mail destined for /usr/lib/sendmail -odq -oi -furi-request uri-out) on Fri, 17 Sep 93 05:31:23 -0400
Received: from genie.lut.ac.uk by avarice.lut.ac.uk id <14130-0@avarice.lut.ac.uk>; Fri, 17 Sep 1993 10:36:53 +0100
Date: Fri, 17 Sep 1993 10:17:10 +0100
Sender: ietf-archive-request@IETF.CNRI.Reston.VA.US
From: Martin Hamilton <M.T.Hamilton@lut.ac.uk>
X-Orig-Sender: Martin Hamilton <M.T.Hamilton@lut.ac.uk>
Reply-To: Martin Hamilton <M.T.Hamilton@lut.ac.uk>
Subject: Re: URN Usage
To: Rob Raisch <raisch@internet.com>
Cc: Uniform Resource Identifier discussion <uri@bunyip.com>
In-Reply-To: <Pine.3.03.9309162316.B5114-c100000@hmmm.internet.com>
Message-Id: <Pine.3.07.9309170939.C13654-b100000@avarice>
Mime-Version: 1.0
Content-Type: TEXT/PLAIN; CHARSET="US-ASCII"

Rob Raisch said:

> We really need to stop running around wrangling with abstract concepts and
> get something done.  I for one would like to see a URN scheme which had a
> number of usable features that could be used simply to differentiate one
> *product* from another. (In fact, I suggested such a scheme quite some time
> ago which generated almost no comment whatsoever.  Not sour grapes, mind
> you.  Just a very tired mind, up late at night, watching people argue over
> details which will never be used.) 

Ok, here goes...  I'm putting together a whois++ server just now.
Here is a whois++ template for converting URNs to URLs - question
is whether this information is enough (Admin- stuff too maybe?).
I think it would be best to isolate the URN->URL mapping from
everything else, particularly since whois++ might not be the
only way of doing this lookup!

  Template-Type: URN2URL
  Handle: Unique ID
  URN: Uniform Resource Name (one)
  URL*: Uniform Resource Locator (many)

(I'm not sure if URL* is the right syntax, tho :-)

If we say the idea is that some of the entries in whois++ reference
URIs, which are (still?!?) URN+sequence number, then the whois++
side of things looks to be sorted out, but there are still at least
a couple of issues in the UR* world:

  o given a URN (from wherever), how to find a server for it 

      URNs are going to be unwieldy if they have to incorporate
      access protocol, host name/ip address, port number et al -
      but without these there can only be one way of resolving URNs?

      Perhaps a workable solution would be to use one of the DNS
      related scenarios, e.g. some component of the URN translates
      to a domain name, under which is a URL (or URLs) for the
      URN registry?  How will that scale?? :-}

  o what the syntax of a URI is - any objections to this...

      URI: <URN:...whatever>+number

Cheers,

Martin