[exim] Spam control via ratelimiting.

We have been toying with the concept of ratelimiting for some time. We
use it create log entries and in some cases we use the control=freeze
option to stop outgoing spam. I would use the defer option if email
clients reacted sanely to the event, but since they don't we are really
hesitant to use it. Apple Mail is the worst - it opens up a dialogue
window to the configuration pane where the user can re enter their
password (which of course many of them no longer know) or pick different
port numbers. It's all a recipe for a run on our friendly Help Desk people.

Enough complaining, here is what I and my Director would like to be able
to do.

-- Identify likely spammers via ratelimiting
-- Freeze these messages using control=freeze
-- Allow them to be delivered very slowly. And this is where it gets
ugly - I could just let these messages sit in the queue ( I don't have
autothaw enabled so now thats what they do.) I could collect message ids
and deliver them via a script one every 90 seconds or so, but that
doesn't solve the problem of one messages to 100 recipients. As soon as
I release one message hotmail or yahoo gets hit with a hundred incoming
chunks of spam and then we are on the blocklist. Is there anyway
currently to control the release of a message (addressed to multiple
recipients ) one recipient at a time?
I expect the short answer is no - probably violates a couple of RFC's,
but I wanted to ask before I moved on to other ideas. I've been
trolling the docs and haven't seen anything yet.
Thanks
/sd

--
Steve Devine
Systems& Infrastructure
Academic Technology Services
Michigan State University