Message-Id: <199804102158.RAA19620@spot.cs.utk.edu>
From: Keith Moore <moore@cs.utk.edu>
To: "Gregory J. Woodhouse" <gjw@wnetc.com>
cc: Keith Moore <moore@cs.utk.edu>, Jacob Palme <jpalme@dsv.su.se>,
        IETF working group on revision of mail standards <drums@cs.utk.edu>
Subject: Re: A solution for group reply redirection which everyone might accept? 
In-reply-to: Your message of "Fri, 10 Apr 1998 09:13:32 PDT."
             <Pine.BSF.3.96.980410090707.13229D-100000@shell3.ba.best.com> 
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Date: Fri, 10 Apr 1998 17:58:51 -0400
Sender: moore@cs.utk.edu

> Regardless of whether or not this is a function of users becoming
> accustomed to one set of semantics, I've found that the Reply-To semantics
> in MS Exchange (where the equivalent of Reply-To, if set, overrides both
> types of replies) can be disconcerting to people accustomed to other UAs.

In general, I've found that:

+ significant changes to UA interfaces are disconcerting

  when changes are necessary, it's very important to do them in such
  a way that they don't confuse or frustrate experienced users.

+ when previously disjoint groups begin to communicate, if they have
  different notions of how things like "reply" work, both groups find 
  this disconcerting.  

  e.g. the BITNET community was accustomed to user agents whose 
  notion of reply was limited to "reply to From/Reply-To" only 
  (often constrained to be a single address). Hence it was common
  for BITNET-hosted lists to rewrite the From address or mung 
  Reply-To.  Whereas the Internet community, not being constrained
  to CMS NOTE, PROFS, or VMS MAIL, was much more accustomed
  to "reply to all" and their lists were less likely to mung Reply-To.
  We're still fighting some of the battles that resulted from the
  Great Migration from BITNET to Internet.


> On the other hand, it is useful in some situations. I would still prefer
> to have either two types of "Reply-To" fields, or an exclusive flag that
> I can use to change the semantics.

Which two types do you mean?  I'm assuming that Reply-To or something
like it will still exist to allow the author to redefine the
responder's notion of "reply to author", so there will be two "types"
in any event.

The question seems to be whether the two types are:

1) allow the author to redefine the responder's notion of "reply to author"
2) allow the author to redefine the responder's notion of "reply to group"

or

1) allow the author to redefine the responder's notion of "reply to author"
2) allow the author recommend a new discussion context for responses,
   which isn't typed as "list", "group", "all", "personal" or
   anything else, and doesn't redefine the responder's notion of 
   "reply to group" or "reply all" etc.

Keith