I-D ACTION:draft-ietf-nntpext-base-06.txt

[Bill Godfrey]

Hello Thomas

> And it raises other problems:

<snip>
Stop me if I've missed your point, but I think you missed Charles's point.

The whole point of Replaces: (as I read it) is to replace an article, but
where the update is not so important as to require people to re-read it.

For example, if I were to post a message...
------------
Newsgroups: example.example
From: bill at bacchae.demon.co.uk
Subject: NNTP information.
Message-ID: <bob-is-silly at bacchae.demon.co.uk>
References: <my-message at bobs-site.co.uk>

Bob wrote:
> Does anyone know which TCP port SMTP works on?

It works ocer port 25.
------------

A day later, I notice the spelling error. Many people would have already
read this message and I really wouldn't want to waste thier time and
network capacity going through a Supersedes: 

Instead, I would send out...
-------------
Newsgroups: example.example
From: bill at bacchae.demon.co.uk
Subject Re: NNTP information.
Message-ID: <replaces.bob-is-silly at bacchae.demon.co.uk>
References: <my-message at bobs-site.co.uk>
Replaces: <bob-is-silly at bacchae.demon.co.uk>

Bob wrote:
> Does anyone know which TCP port SMTP work on?

It works over port 25.
------------
(Notwithstanding that I'm not sure If I list the replaced message id
in the references list, but it's beside the point.)

News servers would probably copy this new message over the old one, and
give it the same internal message number. Anyone whose already read this
message wouldn't been served this new version. Anyone who hasn't yet 
would be given the new version.

If my error was serious, such as I had said port 26, a Supersede would
be in order, but it doesn't really matter that I typed ocer instead of
over. Anyone with half a brain and a good knowledge of English should be
able to make sense of the old version.

As for when to expire this new version, it's up to the admin. Personally,
I'd go for giving it the same *remaining* lifespan as the message it
replaced.

Bill, a veritable mine of information.