Re: regarding illegally formed address and commands

John C Klensin <> Sat, 28 December 1996 00:56 UTC

Received: from cnri by id aa12595; 27 Dec 96 19:56 EST
Received: from by CNRI.Reston.VA.US id aa19473; 27 Dec 96 19:56 EST
Received: from localhost ( []) by (8.7.6/8.6.12) with SMTP id TAA21011; Fri, 27 Dec 1996 19:12:45 -0500 (EST)
Received: from ( []) by (8.7.6/8.6.12) with ESMTP id TAA20997 for <>; Fri, 27 Dec 1996 19:12:37 -0500 (EST)
Received: from by MAIL1.RESTON.MCI.NET (PMDF V5.1-5 #8388) with SMTP id <01IDIKKGQQEU000BKK@MAIL1.RESTON.MCI.NET> for; Fri, 27 Dec 1996 19:12:07 EST
Message-Id: <>
Date: Fri, 27 Dec 1996 19:10:16 -0500
Precedence: bulk
From: John C Klensin <>
To: Ned Freed <>
Cc: Jack De Winter <>,
Subject: Re: regarding illegally formed address and commands
MIME-version: 1.0
Content-type: text/plain; charset="us-ascii"
X-Mailer: Windows Eudora Pro Version 3.0 (16)
X-Listprocessor-Version: 8.1 beta -- ListProcessor(tm) by CREN

At 09:36 96.12.24 -0800, Ned Freed wrote:
>almost always unused is just plain silly. Try explaining it sometime to a
>senior technical person at a large company who is trying to get some
>mail delivered whose return address is both bogus and unfixable. 

Even better, try explaining it to her even more senior, and non-technical,
boss. The technical term for doing that is "death wish".

While I've been an advocate, for many years, of tough "enforcing the rules
is the only way to shake out the trash and make forward progress" policies,
the reality, as Ned points out, is that the vast majority of users feel a
lot more strongly about the mail being delivered if that is at all possible
than they do about any degree of rule-enforcing and moralizing.  Even if
they agree with rule-enforcing in principle, they are "willing" to make an
exception every single time there is a question about something of
importance to them getting through.

There are some important principles here which apply to SMTP extensions
work, to DRUMS, etc.

  *  Ultimately, the worst thing that can happen to a standard is for
people to ignore it as irrelevant.  The fastest way to become irrelevant is
to "require" people to do things that violate their good sense and/or
business interests.

  * Any rule that is made that _requires_ bouncing mail that could
otherwise be delivered will be ignored by almost every for-profit vendor
who intends to stay in business and by almost every not-for-profit producer
who wants to see his or her code in general use.  The exceptions arise with
vendors who have purists as customers (purists are few and usually
unprofitable) and with reference implementations that are written to
demonstrate The Right Way to do something, rather than how to get the mail

  * And any rule that _requires_ massive inefficiency in mail handling will
be ignored by nearly as many vendors/producers as a rule that requires
bouncing mail.   As with delivery, the users tend to prefer performance to
religious arguments.