Errors-to: owner-drums@CS.UTK.EDU
Message-Id: <199505312010.QAA20373@CS.UTK.EDU>
Date:        Wed, 31 May 1995 22:00:06 +0200
From: Eric Thomas <ERIC@SEARN.SUNET.SE>
Subject:     Re: getting started
To: drums@CS.UTK.EDU, Dave Barr <barr@math.psu.edu>
In-Reply-To: Message of Wed, 31 May 1995 15:55:21 -0400 from Dave Barr
             <barr@math.psu.edu>

On Wed, 31 May 1995 15:55:21 -0400 Dave Barr <barr@math.psu.edu> said:

>What I meant to say is that if a gateway is set up such that it destroys
>the envelope  FROM information, that  it must not generate  any messages
>which  would be  sent the  address which  was destroyed  (e.g. bounces).
>Alternately the  system may direct such  messages to a human  for manual
>processing.

It's  still  not implementable  with  typical  non-Internet mail  systems
(other than  X.400). Simply imagine  a reduced version of  RFC822 without
"Reply-To:", "Sender:", "Resent-*" and without the MAIL FROM: field. Just
"From:" and this  is where everything goes - bounces,  replies, acks, you
name it.  This is what  the gateways  have to deal  with. You can  try to
require  them  to wave  a  magic  wand  and  cause bounces  to  magically
disappear into the  nature but that's just wishful  thinking, the reality
is that  if bounces go  to place X,  replies will go  to place X  and the
message will appear to come from place  X. You can choose to make place X
point to  a black hole  or to  a gatemaster, but  this in turn  means the
users won't  buy the  product, and  you'll be out  of business.  The only
viable products  are the ones  that do what  the users want,  which means
replies go  to the originator  and not to a  black hole. This  also means
bounces  go to  the originator  even when  there was  a MAIL  FROM: field
pointing elsewhere. Unfortunately, that's what the market demands.

  Eric