Message-Id: <199906081512.LAA07941@astro.cs.utk.edu>
From: Keith Moore <moore@cs.utk.edu>
To: "D. J. Bernstein" <djb@cr.yp.to>
cc: ietf@ietf.org, drums@cs.utk.edu
Subject: Re: Asinine anti-spam mechanisms 
In-reply-to: Your message of "08 Jun 1999 03:19:56 -0000."
             <19990608031956.2660.qmail@cr.yp.to> 
Date: Tue, 08 Jun 1999 11:12:37 -0400
Sender: moore@cs.utk.edu

   Sometimes the first attempt to deliver a message is stymied by a
   temporary problem such as a network outage or a full disk.
   High-quality mailers do not give up on delivery after a single
   temporary failure; they wait for a while and then try again. They
   give up only if the ``temporary'' problem persists for several days.


poor analogy.  when we suspect that a message may be spam we don't want 
to handle it in the same way as a network outage or full disk.  
in particular:

- we aren't willing to tolerate the same amount of max delay before
  deciding whether to send the message to the recipients.  we wait
  three days before giving up attempting to relay a message when the
  relay attempt fails for temporary reasons. suspected spam, on the 
  other hand, typically gets reviewed within a few hours.

- if we do decide that the message is spam, we want to discard it
  silently.  we don't want to return the message to the sender.

Keith