News: anywhere vs. NNTP: to-server, eh?

Al Gilman <asgilman@access.digex.net> Sat, 16 August 1997 21:44 UTC

Received: from cnri by ietf.org id aa11018; 16 Aug 97 17:44 EDT
Received: from services.bunyip.com (services.Bunyip.Com [192.77.55.2]) by cnri.reston.va.us (8.8.5/8.7.3) with ESMTPid RAA11083; Sat, 16 Aug 1997 17:47:27 -0400 (EDT)
Received: (from daemon@localhost) by services.bunyip.com (8.8.5/8.8.5) id RAA01723 for uri-out; Sat, 16 Aug 1997 17:16:40 -0400 (EDT)
Received: from mocha.bunyip.com (mocha.Bunyip.Com [192.197.208.1]) by services.bunyip.com (8.8.5/8.8.5) with ESMTP id RAA01718 for <uri@services.bunyip.com>; Sat, 16 Aug 1997 17:16:35 -0400 (EDT)
Received: (from daemon@localhost) by mocha.bunyip.com (8.8.5/8.8.5) id RAA09884 for uri@services; Sat, 16 Aug 1997 17:16:35 -0400 (EDT)
Received: from access2.digex.net (access2.digex.net [205.197.245.193]) by mocha.bunyip.com (8.8.5/8.8.5) with ESMTP id RAA09881 for <uri@bunyip.com>; Sat, 16 Aug 1997 17:16:32 -0400 (EDT)
Received: (from asgilman@localhost) by access2.digex.net (8.8.4/8.8.4) id RAA05889; Sat, 16 Aug 1997 17:16:27 -0400 (EDT)
From: Al Gilman <asgilman@access.digex.net>
Message-Id: <199708162116.RAA05889@access2.digex.net>
Subject: News: anywhere vs. NNTP: to-server, eh?
To: lynx-dev@sig.net
Date: Sat, 16 Aug 1997 17:16:27 -0400
Cc: uri@bunyip.com
In-Reply-To: <Pine.SUN.3.95.970816145553.16229A-100000@xochi.tezcat.com>
X-Mailer: ELM [version 2.4ME+ PL15 (25)]
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
Sender: owner-uri@bunyip.com
Precedence: bulk

> X-URL: http://www.flora.org/lynx-dev/html/month0897/msg00299.html
>
>      * Subject: Re: LYNX-DEV Re: news: vs. nntp: scheme
>      * From: Klaus Weide <[5]kweide@tezcat.com>

> 
> To me "news:" means "at any USENET site".  For practical use in accessing
> data, the "any" has to be translated into a specific machine and access
> protocol.  It's just a useful optimization to use a local server if you
> have one.
> 
> There are useful differences in meaning.  USENET vs. NNTP.  A community of
> people (which may not all use NNTP) vs. access to a specific machine
> (which may not exchange any messages with USENET).  Even though Lynx
> (following others) extends the syntax of RFC 1738 in various ways, so that
> it understands several strings to mean the same string, this doesn't mean
> for example that it's a good idea to give out to other people an URL for a
> USENET message in the "nntp:" form.
> 

I like it.  Let's see if I get it.  IF we were to put the
lynx-dev archive into a NewsGroup local to Flora.ORG, THEN that
would be a resource which should be referred to with an nntp: URL
(according to your semantics), right?  It would be useless to go
looking for the resource in any other location.

At least I like the fact that this makes the news: scheme a
de_facto example of an URI which is really a resource Name and
not a resource Location.  That's good.  The HTML 4.0 draft thinks
that URIs are all URLs and transport is all HTTP.  We need
somehow to broaden their vision...

On the other hand, I support the extension of "localhost"
defaulting across schemes other than http: so it would be easy
enough to fold these URNs and URLs under one syntax where if you
don't provide a path the URI is dereferenced by a
client-side-determined search strategy starting with [the usual
suspects].

This would give you permission to search for nntp:group.subgroup
anywhere you find an NNTP server and still know where to look
when provided with a resource identification with a path as in
nntp://news.flora.org/lynx.dev ?

The [news | nntp] converged scheme so defined would be always a
resource name with optional location cues embedded.

--
Al Gilman