On   Fri,   20    Aug   1993   21:18:04   GMT    "Richard   W.   Wiggins"
<[log in to unmask]> said:
 
>I disagree. If I set up a robot reply tool I would never want this tool,
>as my agent, posting to mailing lists
 
I have  no problem with  that -  who the robot  writes to and  under what
conditions is clearly up  to you. However, if you DO  write such a robot,
it should identify its messages as being from a robot.
 
>I belong to  a lot of mailing  lists, some Listserv, some  others, and I
>cannot think of a  single mailing list that would want  me to inform all
>the members of my absence for every message someone posts to the list.
 
I can  think of  a lot of  mailing lists  where I'd want  a copy  of your
vacation/sickness notice as list owner, and  a handful of lists where I'd
want the  whole list informed  (once of  course - most  vacation programs
already know not to send multiple  notices to any given address). Part of
my job is to run a network along  with a small group of other people, and
we report all  problems through internal lists, even when  we know who is
going to have to fix it, so that everyone knows what is going on. We also
inform each other of vacations. If one of my colleagues forgot to do that
before leaving, I  do want the vacation message  distributed to everyone,
because it's important. We actually do that manually when it happens.
 
>Forgetting the auto-reply  tool question, then it's the  case that there
>is no suggested header that a user  (or a robot) can key on to recognize
>that a piece of mail came from a mailing list manager?
 
/usr/ucb/vacation was designed with unix and sendmail in mind. It keys on
the obsolete  '-request' string  and a number  of sendmailisms  to detect
list mail.  Adding a check  for 'owner-' in  both RFC822 header  and "the
RFC822 From_" field as they call it  (SMTP MAIL FROM:) would be a step in
the right direction, since this is what most lists are using nowadays. An
even  better step  would be  to  add 'X-Report-Type:  auto-reply' to  the
generated message. No, it's not standard,  there is no standard for that,
but it  works -  it will  prevent the message  from being  distributed by
LISTSERV  and any  other list  manager that  supports the  LMail delivery
errors format.
 
  Eric