Message-Id: <199603270208.VAA21593@wilma.cs.utk.edu>
From: Keith Moore <moore@cs.utk.edu>
To: Dave Crocker <dcrocker@brandenburg.com>
cc: drums@cs.utk.edu, moore@cs.utk.edu
Subject: Re: Headers and agents 
In-reply-to: Your message of "Tue, 26 Mar 1996 14:14:58 PST."
             <v03005a01ad7e1b927061@[205.214.160.106]> 
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Date: Tue, 26 Mar 1996 21:08:28 -0500
Sender: moore@cs.utk.edu

> 	Here's a question:  Do we REALLY object to creation of non-existent
> Date or Message-ID headers?  If so, why?  Or is it that we never, ever want
> an MTA to REPLACE these fields when they already exist?

I have a moderate objection to MTAs adding a Date field, because it
misleads the recipient as to when the message was sent.

I have almost no problem with MTAs adding a missing Message-ID field.

However, I do think it's important to try to prevent further erosion of 
slippery slope between UAs and mail transport.  The particular case of MTAs 
adding Message-IDs doesn't cause nearly as much harm as the general problem of 
MTAs munging and/or interpreting headers.  My guess is that adding Message-IDs 
does more good than harm.

I'm not sure how best to specify this except to  (a) insist that UAs supply 
all of the required fields, and (b) document the reasons why MTAs sometimes 
add missing fields along with advice as to what is likely to do the least harm.

Keith