Date: Fri, 8 Dec 1995 14:28:48 +0100 (MET)
From: Jacob Palme <jpalme@dsv.su.se>
To: Robert Elz <kre@munnari.oz.au>
cc: ietf-drums <drums@cs.utk.edu>
Subject: Re: Clarify amibuities 
In-Reply-To: <4643.818365358@munnari.OZ.AU>
Message-ID: <Pine.SUN.3.91.951208142216.15708P-100000@ester.dsv.su.se>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII

On Fri, 8 Dec 1995, Robert Elz wrote:

> That is exactly what I thought you were saying.   This is not
> an ambiguity.  It may be a deficiency, I won't argue about that,
> because the sender doesn't have all the facilities available
> that he might want, he cannot say "if you want to send a reply
> to just me, send here, but if you want to send to lots of people
> then send there instead".   I agree, that may need to be fixed.
> However if we define Reply-To: as "I want you to send replies (all
> replies) to these addresses, and only these addresses" there is
> absolutely no ambiguity left in the meaning of that statement.

You are right that with that definition, Reply-To does not become
ambiguous. Note however that that definition is not the way
Reply-To usually is used today, and that that definition will
cause major problems for many existing implementations.

To really clarify the meaning of the new definition, knowing
that it is radically new, I would like it to be formulated
in the following way (a sharper formulation of what you write):

  Reply-To is used by the sender to indicate where replies are to
  be sent. Reply-To should only be used in cases where the sender
  wants all kinds of replies sent to the same recipient(s). In particular,
  in cases where replies can occur intended for either only the
  author, or only the sender, or all recipients of the replied-to
  message, then "Reply-To" should *not* be used!

Is not that what you meant? But note that that is not the
way Reply-To is used today. Today, Reply-To is sometimes meant
for replies intended for the sender only, sometimes for replies
intended for all recipients, even in cases where both kinds of
replies are possible and can occur! So your new definition
means something radically new, it is not a clarification
of current practice. Current practive is still ambiuous!

------------------------------------------------------------------------
Jacob Palme <jpalme@dsv.su.se> (Stockholm University and KTH)
for more info see URL: http://www.dsv.su.se/~jpalme