Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!linus!decvax!ittvax!swatt
From: swatt@ittvax.UUCP (Alan S. Watt)
Newsgroups: net.news.adm
Subject: Re: This bothers me a bit, so...
Message-ID: <1191@ittvax.UUCP>
Date: Tue, 27-Dec-83 15:05:54 EST
Article-I.D.: ittvax.1191
Posted: Tue Dec 27 15:05:54 1983
Date-Received: Wed, 28-Dec-83 06:07:36 EST
References: cbosgd.759
Lines: 39

Regarding purported abuse of "net.test" newsgroup:

Actually, one can argue that the very nature of 'net.test' requires
automatic answer daemons -- How can you find out where the message is
getting if all human readers have un-subscribed to net.test.

We already have the various control messages to send data back to UUCP
and USENET map makers, why not an 'echo' control message whose only
function is to acknowledge receipt of the test message, perhaps including
the body of the original message to track down line-eating gremlins?

One could make this fairly cheap (in terms of total propagation) by
having a maximum path count (default of 1) which would only acknowledge
the message if the "From" path length were less or equal to the
specified maximum.  Thus if you thought there were a problem getting
news out past your most immediate neighbor, you could send an 'echo'
control message out with a path maximum of 2.  Systems at path length
== 1 would acknowledge and forward; systems at path length == 2 would
acknowledge and drop the message.
	
	(One could also argue that this kind of attenuation effect
	should apply to all USENET articles, but that's a topic for
	another article).

People would be encouraged not to specify a path-maximum of 99, or
something absurdly large, simply because more precise techniques for
locating faults would be available; if news admin on site A finds
that everything gets out to length 3 correctly except for stuff going
via site B, he can simply request news admin at site B to run tests
centered there instead of increasing the path-maximum.

People with a legitimate need to test article transmission would be
assisted by automated support; everyone could un-subscribe to 'net.test'
without hindering people tracking down problems (in fact, there would
be no need to actually insert the article in the local news storage);
and no one would need to send letters, polite or otherwise, to people
who might or might not be abusing this particular newsgroup.

    - Alan S. ("Every problem contains the seeds of its own solution") Watt