Date: 06 Apr 1998 10:10:00 +0200
From: mail-std-list@faerber.muc.de (=?ISO-8859-1?Q?Claus_Andr=E9_F=E4rber?=)
To: drums@cs.utk.edu
Message-ID: <6rM9PZL3cDB@faerber.muc.de>
In-Reply-To: <4382.891240263@aussie.cs.mu.OZ.AU>
Subject: Re: Reply-To issue list
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
REFERENCES: <v04003a00b14c17fa2b7d@[198.94.11.101]>

Jacob Palme <jpalme@dsv.su.se> schrieb:
> At 07.44 +0100 98-03-30, Robert Elz wrote:
> > We should stop making arbitrary distinctions between two kinds out of
> > the considerably larger potential sets of types of replies.
>
> Why has it been so difficult to agree on a solution to the Reply-To
> issue? Is it because we have basic difference in underlying values.
>
> I believe that Robert Elz statement above (Keith Moore has written
> similar things) captures the essence of the basic disagreement. And
> that this is a disagreement on underlying values, and that is why it
> is so difficult to agree.
>
> I believe that (correct me if I am wrong) Robert and Keith feel that
> it is wrong for a sender to give direction on where replies are to be
> sent.

No, as I understand it, they feel that the author (not sender) may give  
advise (better use that word instead of direction) on where replies  
should be sent. However, they (and I) also believe that it is impossible  
to give advise for any case that can be thought up.

To be most useful, the reply to header should hold all recipients the  
author(s) suggest(s) as targets for replies. This allows him/her/them to  
exclude To and CC recipients from the reply address list.

Everything else should be left to those replying or their user agents,  
which is sending Replies to the author (From), sender (Sender ? Sender :  
From), all (From + Reply-To + To + Cc - duplicates - own).

What I'd like to add is that it IS possible to automatically detect the  
use of the Reply-To header as "Reply-To replaces From". Iff the From  
address holds a single address, there is no Sender header and the Reply- 
To header holds an address that is not given anywhere else, it is very  
likely that Reply-To was used this way.

Further, I believe that mailing lists are just like newsgroups  
transmitted via an alternate medium. So we should define similar headers  
as Newsgroups and Followup-To for them. Reply-To would handle personal  
replies, while List-Followup-To would handle public replies (followups).  
Please note the distinction between reply (private, directly to a known  
set of users) and followup (public, to an unknown set of users through a  
distribution system).

Due to the similar nature of the two systems, many other problems that  
arise in mailing lists also appear in newsgroups or are already solved  
there. We should use the same mechanisms (headers) for both:

  Supersedes            (proposed draft-ietf-mailext-new-fields-xx)
  Expires               (  -"-                 -"-                )
  Control: cancel       (proposed hereby for mail, no draft yet)
  named articles        (proposed in the USEFOR working group,
                         http://www.landfield.com/usefor/)
  Followup-CC-To        (proposed hereby for both, no draft yet)
  References            (semantics already changed to match[?])
  See-Also              (proposed hereby for mail, no draft yet)
  ...

> They feel that the writer of the reply should be forced to make
> a conscious and thought-through decision on where to send every
> reply. They feel that the two common commands in most mail clients,
> named something like "reply-to-personal" and "reply-to-all", should
> not be there, because these commands relieves writers of replies from
> carefully choosing where to send their replies.

The problem with these two commands is that they (a) often don't work  
correctly as Reply-To is used differently and (b) having these buttons  
often leads to an user agent design where it is hard to change the list  
automatically created. In my opinion, it is fine to have such reply-to  
buttons as long as it is easy to manually add and remove addresses from  
the list.

Finally, I believe that the two distinct buttons should be "personal  
reply" and "public followup", with "personal reply with copies to all"  
being special (i.e. additional functionality) just as "reply to author"  
or "sender". Of course followups could only be available in public  
media, i.e. mailing lists and newsgrous.

-- 
Claus "3247" Andre Faerber <http://www.muc.de/~cfaerber/> fax: +49-8061-3361
PGP: ID=1024/527CADCD FP=12 20 49 F3 E1 04 9E 9E  25 56 69 A5 C6 A0 C9 DC