To: djb@koobera.math.uic.edu (D. J. Bernstein)
Cc: drums@cs.utk.edu
Subject: Re: proposed agenda for 8 March WG meeting 
In-Reply-To: Your message of "05 Mar 1996 16:36:30 GMT."
             <19960305163630.4362.qmail@koobera.math.uic.edu> 
Date: Wed, 06 Mar 1996 12:18:56 +1100
Message-Id: <1134.826075136@munnari.OZ.AU>
From: Robert Elz <kre@munnari.oz.au>

    Date:        5 Mar 1996 16:36:30 GMT
    From:        djb@koobera.math.uic.edu (D. J. Bernstein)
    Message-ID:  <19960305163630.4362.qmail@koobera.math.uic.edu>

    Keep in mind that parallelism helps for _every_ recipient.

Yes, of course, as I said, it helps, I use it on my big lists, and
its a definite gain.   But simply doing parallel delivery is
different to failing to use multiple RCPT commands in one
delivery attempt, when  that does make sense.   Do both, do
parallel delivery where independant delivery is required anyway,
and use multiple RCPT commands where they make sense.

    Those few seconds pile up!

Only within the one message with the multiple RCPT commands.
Delivery to other hosts is proceeding in parallel.

    First, the bandwidth loss is insignificant, because most
    deliveries are to separate hosts.

Again,  this was not the original issue, which no-one would be
debating.   How you use your own system resources  to deliver
mailing list mail (whether you want 1, 20, or 200 connections
open in parallel) is up to the list manager (within the scope
of the tools available).   That has nothing to do with DRUMS.
Whether multiple  RCPT commands should be used or not, perhaps
whether they should be pielined or not, does have relevance.

kre