Date: 5 Mar 1996 18:45:50 GMT
Message-ID: <19960305184550.5428.qmail@koobera.math.uic.edu>
From: djb@koobera.math.uic.edu (D. J. Bernstein)
To: drums@cs.utk.edu
Subject: Re: proposed agenda for 8 March WG meeting

> > Oh, really? What percentage of SMTP servers support it?
> All significant high-performance systems

That wouldn't answer the question even if it were true.

>		PIPE:	4-20    seconds for 100 recipients, 5 seconds for body
>		old:	200-500 seconds for 100 recipients, 5 seconds for body

Straw man. You're comparing zmailer+pipe to zmailer-pipe. Try qmail
(with sendmail or whatever on the other end) and you'll see that your
``essential'' gain is actually quite small.

> I know, "100 RCPTs max", however there was off-band agreenment.

How interesting. Do you back off if the other side keeps running out
of memory?

> I don't believe in splitting the message into parallel streams just
> to alleviate  latencies  while  sending large lists.

Whether or not you believe in it, it works, and people like it.

> From my experience I can say that minimally implementing PIPELINING
> at the SMTP-server is easy,

Not for sendmail, and not for some PC networking software, so it'll be a
long time before support is widespread.

For qmail the only hurdle is ESMTP parsing. Someone sent me code to try
to prove that this was easy; it took me about two minutes to find a case
where his code failed. I'm not interested in breaking my mailer, thanks.

> 	From looking at Dan's mailer (yep, I browsed the source), and his
> 	proposal documents, I get a first impression of "All the world is
> 	8-bit UNIX",

Please don't misrepresent my work. Try ``All the _Internet_ is 8-bit
DOS or Windows or UNIX or Macs or ...''

---Dan