Re: Certified Mail Delivery agent

At 05:34 2001-12-12 +0000, John Conover wrote:

> > Replies according to some special text, or just replies?  What if the 
> reply
> > comes in the form of a bounce (and we know how many variations there 
> are on
> > those), complete with a copy of the sent message (and therefore, 
> containing
> > the phrase that pays)?
> >
>
> How about "To certify that you sent this message, write 375 in the subject
> of the reply", 375 being randomly generated.
At least in the subject, you don't have to contend too much with MIME 
encodings...  Unfortunatley, the message becomes less of a "just reply" to 
"do this if you want us to ever read your email."
> > >Smartlist script forwards the original message on to the recipient,
> > >and enters the e-mail address in the "whitelist" database.
> >
> > Q: how well does this work for _mailing lists_ -- does it send the request
> > to the _sender_ (the mailing list, oh, thrill me), or to the _author_ (who
>
> If mailing lists are in the "whitelist", any message from a mailing
> list will be passed.
Uhm, follow me here:  someone installs the scripts.  They're happily 
motoring along, then use some webform to subscribe to a mailing list.  The 
first message arrives from the listbot asking you to confirm your 
subscription.  As I read it, your agent would reply to that message telling 
the sender (a bot, not a human) to reply in some certain way.  Of course, 
that never happens.  So, a while later, the user is wondering why they're 
not seeing anything, and they go rummaging through their mail archive 
stuff, and discover that the agent got this message but didn't receive a 
meaningful reply.  User retrieves the message, replies to the listbot, and 
is added to the list.  Okay, we're subscribed.  Another message or two 
arrives from the listbot welcoming you to the list and providing posting 
guidelines, etc. -- in all probability, this message comes from yet another 
address, which is subjected to the agent once again, and probably annoys 
the heck out of the listowner if they receive the auth 
requests.  Eventually (perhaps very soon), a message arrives from JoeUser 
on the list, and then another from BettyUser, etc. -- neither the From: or 
the From_ (sender) appear in the acceptlist, so the agent once again sends 
out an auth request -- this time to whom?
UNTIL that list is in the acceptlist, it will presumably be subject to the 
actions of your agent (unless you have other logic in place).  An autoreply 
to the list, or to the list user is going to be a poor way to introduce 
yourself on a new list.  In many cases, when subscribing to a new list, you 
don't know the address the list is mailing out on until you receive that 
first series of messages - by which time, you've "introduced" yourself to 
everyone...
Thus, mailing lists need to NOT be subject to the agent.  If you rely on 
the _sender_, lists will pose special problems for you: if you mail the 
auth request to the _sender_ you may send it to the list owner, who might 
just as well remove you from their list, or you may send it to the list - 
in which case, the listadmin might just as well remove you from the list - 
if you send to the From:, then you're not certifying the sender, and EACH 
user on a mailing list will be subject to harrassment by the agent.
Of course, plenty of spew comes _through_ mailing lists, which is why a 
good ruleset for catching spew is going to be needed anyway, unless you 
plan to discard most list mail.

---
 Sean B. Straw / Professional Software Engineering

 Procmail disclaimer: <http://www.professional.org/procmail/disclaimer.html>
 Please DO NOT carbon me on list replies.  I'll get my copy from the list.

_______________________________________________
procmail mailing list
procmail(_at_)lists(_dot_)RWTH-Aachen(_dot_)DE
http://MailMan.RWTH-Aachen.DE/mailman/listinfo/procmail