From: "Erik E. Fair"  (Time Keeper) <fair@clock.org>
Date: Thu, 5 Oct 95 09:26:42 PDT
To: cypherpunks@toad.com
Subject: Rethinking the utility of netnews "cancel" control messages
Message-ID: <v02110103ac99b201cba4@[204.179.132.4]>
MIME-Version: 1.0
Content-Type: text/plain


The day of unauthenticated netnews control messages of any kind is
basically over. We gotta:

1. turn off all automated system-wide control of netnews, or

2. properly authenticate all such messages (newgroup, rmgroup, cancel, etc.).

I think we also ought to think carefully about continuing to have a
"cancel" control message (and the Supercedes: header) any more,
authenticated or not - as useful as this mechanism occasionally is to
remove unsightly spams (and other Officially Troublesome Material), isn't
this kind of casual revisionism something that is, historically, to be
avoided?

If you had a netnews system which simply marked a message as cancelled in
some way, would you set your netnews reader to seek out cancelled messages?
Or ignore them?

If you were a librarian or historian operating The Official USENET Archive
of Everything, would you accept and process cancel control messages?

        "Backbone cabal? What's that?"

        Erik Fair <fair@noc.use.net>