[R] R list problems ; "r-help-bounces" etc

[Martin Maechler]

First things first :

  Our mail server had experienced a bad fit for a bit more than 16 hours, 
  from about 2005-01-11 15:24 GMT aka UTC
	  to 2005-01-12 07:37 GMT
  Of course, a timeout of such an extended period of time should
  have never happened, and we are very sorry about it.


>>>>> "AndyJ" == Andy Jaworski <apjaworski at mmm.com>
>>>>>     on Tue, 11 Jan 2005 13:27:26 -0600 writes:

    AndyJ> Could somebody please tell me what does the
    AndyJ> r-help-bounce address do?

r-help-bounces (final "s") that is.

All our mailman mailing list e-mails are sent with a header
"Sender: <mailinglist>-bounces at ...."
[The following is not 100% precise, since e-mail protocols are
 even a bit more complicated]
This is the address that your mail server (not your "mailer" aka
your mail client!) should use when it finds that your
own address is not valid and it cannot deliver mail to you.
The mail server should return the mail to the
"Sender" and tell the reason why the mail couldn't be delivered.

[BTW, also ``decently configured'' vacation programs should reply to
 the "Sender:" address when that differs from the "From:" one ..]

On our side, the mailing list software "mailman" receives all
the mails to *-bounces at ... and tries to deal with it properly.
In most cases mailman, registers that an address has had
problems, and if an address produces too many such bounces, it
will be automatically be "disabled".
mailman will not trying to forward mailing list mails to that
address, but still send a daily reminder for a while
before its completely unsubscribed.

Here is a version of the "R-help bounce processing" page that
usually only I as list administrator see :

  http://stat.ethz.ch/~maechler/r-help-bounce-processing.html

You can see the current setting of R-help's bounce processing
parameters.


    AndyJ> When I try to respond to an r-help post, my mailer
    AndyJ> (Lotus Notes) generates the r-help bounce return
    AndyJ> address automatically in addition to the original
    AndyJ> sender address and the r-help address.  

this is a clear mis-feature (aka "bug") in your mailer.
Without knowning the exact protocols, I know that a user's reply
shouldn't go to "Sender:" if there was a "From:"

    AndyJ> I responded to an r-help message last night and I
    AndyJ> never saw my response posted.  Could this be due to
    AndyJ> the r-help-bounce being there?

no, see above.

    AndyJ> Thanks in advance,

you're welcome.

Martin Maechler, ETH Zurich