Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!seismo!topaz!nike!ucbcad!ucbvax!ucbarpa.Berkeley.EDU!fair
From: fair@ucbarpa.Berkeley.EDU (Erik E. Fair)
Newsgroups: net.news.adm
Subject: Re: garbage articles in net.micro.atari16
Message-ID: <15065@ucbvax.BERKELEY.EDU>
Date: Thu, 31-Jul-86 06:03:49 EDT
Article-I.D.: ucbvax.15065
Posted: Thu Jul 31 06:03:49 1986
Date-Received: Thu, 31-Jul-86 23:53:35 EDT
References: <627@polaris.UUCP>
Sender: usenet@ucbvax.BERKELEY.EDU
Organization: University of California at Berkeley
Lines: 39

To Herbie Chong and the other readers of net.micro.atari16, a belated
word of explanation: if the ARPANET burps sends the USENET gateway
duplicates, there are two scenarios:

	1. they assigned a message-id and continually retransmit
		the same message.

	2. they send out the same message over and over without a
		message-id.

In the former case (which all such burps have been to date), USENET
sees one copy of the offending message, and the gateway rejects the
rest, because it has the message-id in the history file and recognizes
the duplicates as duplicates.

In the latter case (as with the problem cited), ucbvax assignes a new,
unique message-id to each incoming copy of the message (feature of
sendmail). Unfortunately, there is no reasonable way for the gateway
to detect the duplicates in this case, so USENET loses big. Ideally,
all mailers in the ARPA Internet should assign network-wide unique
message-ids to all messages originating at the respective sites, but
since that is not as yet the case, we grin and bear the currently
imperfect world...

As for me being easily reachable... Herbie tried to get me during the
one week that I had a modem on the phone on my desk because we're
moving our VAX (styx (nee lll-tis-b.arpa), *not* ucbvax!) and other
machines, and had not set up reasonable hardwire access. So he
unfortunately got modem screech in his ear. For this, I apologize; I
should have set the thing to ignore incoming phone calls. I'm somewhat
puzzled by not getting any notice of the call from the group secretary,
though. After all, if he found my desk number at work (a cute trick,
given that we had been in the new building for only two weeks at the
time) he should also have had the general office receptionist's
number...

The quickest, most reliable way to reach me is what I sign off with:

	Erik E. Fair	ucbvax!fair	fair@ucbarpa.berkeley.edu