Errors-to: owner-drums@CS.UTK.EDU
To: Eric Thomas <ERIC@SEARN.SUNET.SE>
Cc: drums@CS.UTK.EDU
Reply-To: drums@CS.UTK.EDU
Subject: Re: From the Chair: the Reply-To issue 
In-Reply-To: Your message of "Sun, 27 Aug 1995 02:59:31 +0200."
             <199508270128.VAA04351@CS.UTK.EDU> 
Date: Sun, 27 Aug 1995 16:12:40 +1000
Message-Id: <13029.809503960@munnari.OZ.AU>
From: Robert Elz <kre@munnari.oz.au>

    Date:        Sun, 27 Aug 1995 02:59:31 +0200
    From:        Eric Thomas <ERIC@SEARN.SUNET.SE>
    Message-ID:  <199508270128.VAA04351@CS.UTK.EDU>

    I think we're missing an important point in this discussion

So do I, but in a different way.   That is, mailing lists have
basically nothing whatever to do with the discussion wrt Reply-To.
They are a peripheral distraction at best.

Even without mailing lists, I can send a message

	To: someone, another, and-more
	cc: one, tow, three, one-hundred, people

(all those are really fqdn's of course).

Whether one (or more) of those addresses happens to expand
to a few hundred, or thousand, other people really doesn't
matter in the slightest to the principles involved.

That is, when I send the message, I might decide that I prefer
that replies be sent just to me, without bothering everyone
else, or to all of the people on the To line, the cc's were
just to people I wanted to know that I had sent the message,
but who don't have much interest in the answers, or to all
the people I listed, or almost any combination.

Reply-To is the technique I can use to do that, I simply stick
in a Reply-To that nominates who it is I prefer should get
replies.   The recipient of my message can decide to acceed
to my preference, or not, as his mood takes him at the time.

If we start looking at this by our perceived view of how the
recipient's UA's behave, or should behave, or could behave, we're
doomed.   There are way too many of those things, and they're all
different.   The graphic UA I use has no "reply all" or "reply
personal" buttons, instead I have about 8 different reply functions
I can use (I can't enumerate them, I never remember them all, but
I do remember how to use each one when that one is called for ..
just as I can never remember my passwd, but never have any trouble
typing it).   Those 8 aren't much like most of you imagine exist.
Further, if I want more flexibility I have a command line
interface I can use where I have even more options, and that is
before I start editing the headers of the template message I
stick my reply in.

The headers are a protocol (languages are protocols, there is no
difference here, protocols don't have to be bi-directional things),
they are intended to convey specific information from the sender
to the recipient (and occasionally from other systems along the
way), the purpose of the spec (822) is to tell us (sender's UA
authors) how to encode the sender's wishes so they can be
understood and acted upon by us (recipient's US authors) so
the recipient understands the sender's desires, whether or not
they decide to take any notice.

For present purposes, we have to decide just what Reply-To:
contains, is it purely an "I don't like my From: line, use
this instead", or is it "Please, this is where I would most
replies directed".   Unfortunately, rfc822 doesn't say (its
examples can be interpreted either way) - that is why DRUMS
is here, to make that decision.

So, to go back to Keith's message ...

    So on this particular issue, I want people to very clearly state
    whether Reply-to expresses the author's preference about (pick one):

    [ ] where replies to the message should be sent

yes.

    [ ] where replies to the *author* should be sent

no (given that I interpret this question as meaning Reply-To
overrides From and has no other purpose).

kre