Message-Id: <199709241953.PAA15523@spot.cs.utk.edu>
From: Keith Moore <moore@cs.utk.edu>
To: Jamie Zawinski <jwz@netscape.com>
cc: drums@cs.utk.edu, moore@cs.utk.edu
Subject: Re: Reply-To (was Re: Issues from the past few weeks) 
In-reply-to: Your message of "Mon, 22 Sep 1997 17:15:07 PDT."
             <34270A0B.BA327658@netscape.com> 
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Date: Wed, 24 Sep 1997 15:53:18 -0400
Sender: moore@cs.utk.edu

> Let's suppose I'm writing an MUA.  Let's suppose I have decided that 
> in my program, I would like to have two commands, "Reply to Author" 
> and "Reply to All".

Here's one way to address that situation:

If there's no reply-to header, you use From for "Reply to Author"
and From+To+CC for "Reply to All".

If there is a reply-to header, you pop a box that says something like:

  +---------------------------------------------------------------------+
  | The sender has requested (using Reply-To) that replies to this      |
  | message be sent to: <foo@bar.com>                                   | 
  | Do you want to send this reply to that address?  [yes] [no]         |
  |                                                                     |
  |[ ] <- In the future, honor this reply-to request without asking me. |
  |[ ] <- In the future, ignore this reply-to request without asking me.|
  +---------------------------------------------------------------------+

The check boxes at the bottom is a sop to those individuals or mailing
lists that always supply a reply-to field....for those lists or
individuals, you can tell your UA to Just Do It  (or Just Ignore It).

People will be surprised or even annoyed that their UA is asking them
such things, simply because it's a change in behavior that "used to
work".  But they won't be surprised by where the message gets sent!
(of course, most people won't notice the improvement...)

> It's not the place of DRUMS to tell one what a user interface should
> look like; but it *is* the place of DRUMS to make it clear to an
> implementor how to map abstract concepts like the above commands onto
> the specifics of what is in the message headers.  That is, DRUMS should
> make it clear how I should interpret the Reply-To header.  

This isn't a problem.

The problem is that current practices, and user expectations, are
fundamentally at odds with having "Reply" work in a sane fashion.  
Any reasonable solution to the problems with Reply and Reply-To 
will necessarily violate user expectations to some degree.

Keith