At 11.53 -0500 04-12-09, Bruce Lilly wrote:
O can indicate a suggestion for where to send responses
via the Reply-To message header field (as described in
RFCs 2822, 822, 733; it is particularly applicable to
messages sent to mailing lists ("text message
teleconference") as described in those RFCs).
My belief is that we will never get a consensus with the
main developers on how to use "Reply-To". Because of that,
I feel that this is a header which you should not use, and
if you get it in incoming mail, should ignore, because your
mailer cannot be sure what its intention is. I know that
people in this mailing list mostly agree that only the
sender should specify "Reply-To" and that it should not be
set by mailing list expanders to refer to the list. But I
do not think you will ever get all mailing list expander
developers to agree with that view.
This is similar to the "Newsgroups" header in e-mail,
which also is so controversial with several very
incompatible meanings in widespread usage that the only
reasonable choice is to ignore it.
My original message only discussed how a receiving mailer
should handle incoming duplicate messages. Certainly, if
you can get sending and receiving mailer to co-operate in a
useful way, you could produce something even better, but
that is very difficult to achieve with the large number
of different mail clients used.
At 11.01 -0800 04-12-09, Russ Allbery wrote:
A related issue is if a message arrives in two copies, one with the
recipient as personal recipient, the other via a mailing list. When this
happens, the cause is almost always faulty handling by the sender's MUA,
and the intention is that the message should only be sent to/via the
list. So in this case, the best receiving MUA behaviour is probably to
completely suppress the personal copy of the message.
This is the way that I feel about it too, but as other people have
commented in this discussion, not everyone feels that way and some people
really want to get those personal copies.
Yes, some advanced users may want special handling of this,
I was mainly discussing the best default behaviour.
Keeping a data base of Message-Ids to recognize duplicates
is rather easy, if you do it the way I once did it. I kept
the data base to a fixed rather low limit. This means that
my software would not recognize duplicates if they arrived
at widely different times, but this does not happen in
reality. The fact that all duplicates typically arrive the
same day or at least the same week makes the implementation
of such a data base much easier.
--
Jacob Palme <jpalme(_at_)dsv(_dot_)su(_dot_)se> (Stockholm University and KTH)
for more info see URL: http://www.dsv.su.se/jpalme/