Message-Id: <199712191835.NAA23237@spot.cs.utk.edu>
From: Keith Moore <moore@cs.utk.edu>
To: "D. J. Bernstein" <djb@cr.yp.to>
cc: drums@cs.utk.edu, moore@cs.utk.edu
Subject: Re: How to find a solution 
In-reply-to: Your message of "19 Dec 1997 03:25:40 GMT."
             <19971219032540.6866.qmail@cr.yp.to> 
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Date: Fri, 19 Dec 1997 13:34:59 -0500
Sender: moore@cs.utk.edu

[if Dan had specified reply-to, he wouldn't get a duplicate copy of
this message.]

> Keith Moore writes:
> > These mailers are already
> > broken, otherwise we wouldn't be having this discussion.
> 
> The problem is a lack of information. 

There's more than one problem at the root of the reply mess.  

Lack of information is indeed a problem.  Poor user interfaces are
also a problem.  No amount of information in the message header is
going to give the responder's user agent enough information to narrow
the responder's decisions into one or two constant-labelled buttons of
the responder's UA.

> When I'm following up to a
> mailing-list message, I don't know whether to send the author a copy.
> Erring on the safe side means that the author receives two copies.

You keep citing that particular problem.  Others cite other problems.
Part of the problem with Reply-To is that it's used for so many
different and conflicting purposes.  If we solve one problem without
trying to solve the others we'll end up in a worse mess than we have
now.

That's why I made a list of problems - so we can look at things
more-or-less comprehensively.

> Mail-Followup-To fixes this.

Mail-Followup-To only "fixes" this problem if user agents get a lot
more flexible than they are now, and only if people don't overload
Mail-Followup-To in the way that Reply-To has been overloaded.

But I see no reason to believe that Mail-Followup-To will not be used
by all of:

authors (as default behavior, to suppress duplicates)
authors (as exceptional behavior, e.g. to say 
         "please send replies to this message to my secretary" or
         "please send replies to these lists only" or
         "please send replies to me only" or
         "please send replies to president@whitehouse.gov")
lists (to set replies to go to the list, overriding author's preference)
lists (to set replies to go to the author, overriding author's preference)
lists (to set replies to go to the list, but only if author didn't
       state a preference)
lists (to set replies to go to the author, but only if author didn't
       state a preference)
etc.

In other words, Mail-Followup-To will be subject to exactly the same
set of multiple uses/abuses as Reply-To, and any attempt by the
responder's UA to automatically generate a recipient list for
``follow-up'' in the presence of Mail-Followup-To will be at least as
broken as the current follow-up behavior in the presence of Reply-To.

This is why I don't see a significant gain from simply adding
Mail-Followup-To.

Keith