Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Posting-Version: version B 2.10.3 4.3bsd-beta 6/6/85; site ucbvax.BERKELEY.EDU Path: utzoo!watmath!clyde!burl!ulysses!ucbvax!athena.mit.edu!jh From: jh@ATHENA.MIT.EDU Newsgroups: net.micro.atari Subject: Re: news being bounced off arpanet! (sort of long) Message-ID: <8510261755.AA04337@MIT-ARTEMIS> Date: Sat, 26-Oct-85 13:55:07 EST Article-I.D.: MIT-ARTE.8510261755.AA04337 Posted: Sat Oct 26 13:55:07 1985 Date-Received: Mon, 28-Oct-85 03:15:37 EST Sender: daemon@ucbvax.BERKELEY.EDU Organization: The ARPA Internet Lines: 81 The reason we can't follow your suggestion is simple. A mailing list like ours is a simple reassignment of names; my name and the names of the other arpa readers get put in a recipient field, and you are still the sender (Which is why I can reply to you now.). This is fine for small groups, as no one has to go through the hassle of moderating the list. Any mail you send is exactly like sending n messages, one to each person on the list. You would get a message from the appropriate mail demon if one of those messages were unsuccessful in getting to its recipient, and judging from your complaints, this is indeed what has been happening (Believe me, from a list this size, be glad there are only eight irresponsible individuals who didn't tell anyone when their mail addresses changed!). This format was really designed only to work for a small group. I am on and run several such groups, one for my living group and its alumni, one for an experimental program I'm in at school, etc. each with less than 100 people on it. It would be a real pain to readdress each message 80 times by hand in order to get to those same people, though. Using this format for a large list, especially one spread out over the country, is going to run into some trouble of the sort you just described. That is why there are digests. Digests are simply all the messages sent to the list put in a single (long) mail message, so that the recipient can "undigestify" them and read them. A typical digest will go out two to ten times a week. The benefit is that the person who "digestifies" (digests?) the messages is the one who resends them to the list, and is thus the one who receives all the mailbounce. The problem is that someone has to do a LOT of work (this list would require a new digest every 2-4 days). This arpa list is of the first kind, which gets us messages faster, but has obvious drawbacks. For one, we now have NO moderator, as Mark Crispin stopped some while ago, and no one has stepped forward to take his place. This means that there are no changes being made to the file containing our names. I have been trying to desubscribe for a month now, without success. It would be quite possible (and quite nice) if someone wrote yet a third kind of mailing list program which would do as you suggested -- automatically send out a message FROM info-atari-request TO (everyone on the list) each time it received a message TO info-atari; not unlike an automatic single-message digest. That way, mailbounc would go to info-atari-request, and the moderator could handle it there. This is indeed a part of a program called pmd (personal mail demon) which is written in C and runs on any system which looks for a .forward file before putting mail in someone's mailbox (e.g. any system running sendmail). A friend of mine wrote pmd, and I will send copies to anyone who wants one. Unfortunately, since we don't have a moderator, this won't happen for a while, so all of us (including usenet recipients) will have to suffer a while longer. This brings up your last suggestion. DON'T START A MAILER WAR! What will happen is this: you send a message into the net. 3 machines work for a few days trying to send your message, and send back one refusal message per day (all the others copies of your message get to us). That's nine messages you get back. You resend those messages to the "offending machines" (which are only doing their jobs). Each machine works for three more days trying to send the messages...you get 27 messages back. Your phone bill will blow up before the machines will. If you send to postmaster at those machines and complain, they will tell you not to expect mail to be delivered to nonexistent addresses, and they HAVE to return it (arpanet requires it). The source of the problem is not at what you call the "offending machines", but on our list and the way it is (isn't) being run. The only solution is to FIND A MODERATOR! I can't do it, because I don't have the kind of storage capacity necessary to hold the archives, and frankly, I want to get off this list. An arpa-arpa mailer war can bring (and has brought) down the Arpanet in a disastrous way. An arpa-usenet mailer war will kill the usenet recipient with phone bills long before the arpa machine dies from overload. If you bounce your bad mail to the machine the list is on, most likely it will slow things down there (as well as load up the list, which you receive) and the people using that machine will complain and the list will go away. There are two immediate solutions: 1. suffer (i.e. accept the mailbounce as the cost of posting.) or 2. don't post. I don't like it (think: I get 8-15 messages a day that I don't want, whether or not I post to the list!), but aside from writing to postmaster@score and telling him to completely kill the list (BAD IDEA!), we're stuck. --jh-- e-mail: post: jh@mit-athena.MIT.EDU (preferred) Joe Harrington jh%oz@mit-mc.MIT.EDU 69 Chestnut Street Cambridge, Massachusetts 02139