Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!seismo!munnari!kre
From: kre@munnari.oz (Robert Elz)
Newsgroups: comp.mail.headers
Subject: Re: reply-to
Message-ID: <1321@munnari.oz>
Date: Wed, 19-Nov-86 14:59:02 EST
Article-I.D.: munnari.1321
Posted: Wed Nov 19 14:59:02 1986
Date-Received: Wed, 19-Nov-86 22:37:52 EST
References: <904@brl-adm.ARPA>
Organization: Comp Sci, Melbourne Uni, Australia
Lines: 30

Jacob is clearly right, 2 reply-to headers are needed.  This has
no implication for control of to whom users may send replies,
obviously, someone who wants to reply to mail can send the reply
to anyone he feels like (including the return-path if he wants).

What's at issue here, is that its currently undefined what the
sender means when he inserts a reply-to header, that is, what
information he's trying to pass to the recipient of the mail.

There are two reasonable meanings.  One is that the sender is
saying "Don't reply to the From: address, its not where I
want replies to go, use this address instead".  The other is
"I would prefer if you would send replies only to this address,
and not to others on the From, To and Cc lines".

RFC822 doesn't make it clear which of these meanings a Reply-To
header is intended to have.  One solution would simply be to
make that clear.  But then we'd have a useful possible meaning
with no way to generate it.  The easy way to fix that is to allow
a new header for the meaning that's left out.

The only affect any of this will have on the recipient of a message
would be the default addresses that the "reply" command in the
mailer sets up in the outgoing message.  Nothing is going to prevent
users replying to messages from modifying those addresses in any
way they see fit.  (I sometimes "reply" to a message, and end up sending
the reply to no-one who saw the original message, or whose address
was mentioned in it anywhere...)

Robert Elz			kre%munnari.oz@seismo.css.gov