Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!mailrus!wuarchive!texbell!uhnix1!splut!jay
From: jay@splut.conmicro.com (Jay Maynard)
Newsgroups: news.software.b
Subject: Re: cancel propagation (was Re: Supersedes problems with rapid-fire articles)
Message-ID: <2852@splut.conmicro.com>
Date: 1 Sep 89 08:55:11 GMT
References: <5200@looking.on.ca> <...> <6233@looking.on.ca> <1989Sep1.004629.17171@utstat.uucp>
Reply-To: jay@splut.conmicro.com (Jay "you ignorant splut!" Maynard)
Organization: Confederate Microsystems, League City, TX
Lines: 39

[Side note: It looks like I'm gonna have to rebuild rn to increase the
interp buffer size; these Cnews 47-foot-long article-IDs are expanding
the references: line faster than used to happen with Bnews. Sigh.]

In article <1989Sep1.004629.17171@utstat.uucp> geoff@utstat.uucp (Geoff Collyer) writes:
>I don't believe one can assume, in general, that a site will get an
>article and its cancel from the same neighbour, nor even assume that
>they will travel the same routes.  Remember that notions of "upstream"
>and "downstream" are fuzzy and don't even always apply to leaf sites.
>Thus it is possible for a cancel to arrive before its target.  Throwing
>away cancels (due to absence of its target) will tend to cause some
>sites to never see a cancel even though they will see the target.

How would this happen?

There are either two possibilities, with respect to any site, backbone
or leaf:
1) The cancel arrives before the target. In this case, the cancel, by
causing an entry to be created in the history file, will cause rejection
of the target when it arrives as a duplicate. The site will not feed the
target to any of its downstream sites, since it doesn't have the article
to feed.
2) The cancel arrives after the target. In this case, the cancel message
does its thing, and deletes the article. Propagation of the cancel is
needed in this case, since the article has already been fed to the
downstream sites.

In general, the cancel will travel the same paths as the target; this
would not be the case only if someone fed foo.bar.ctl to a site that
didn't get foo.bar, and the .ctl newsgroup was specified in the header.
(Does Cnews use this feature? I thought it was obsolete). It does not
help to remove an article from the net's machines faster to propagate
cancels where the article has not gone, and will never go.

-- 
Jay Maynard, EMT-P, K5ZC, PP-ASEL   | Never ascribe to malice that which can
jay@splut.conmicro.com       (eieio)| adequately be explained by stupidity.
{attctc,bellcore}!texbell!splut!jay +----------------------------------------
"Rabid rerouters *love* to route mail to devnull@hell.org" - Brandon Allbery