Errors-to: owner-drums@CS.UTK.EDU
To: Eric Thomas <ERIC@SEARN.SUNET.SE>
Cc: drums@CS.UTK.EDU
Subject: Re: Comments and Resent-* header fields 
In-Reply-To: Your message of "Sat, 23 Sep 1995 01:01:06 +0200."
             <199509222327.TAA28462@CS.UTK.EDU> 
Date: Sun, 24 Sep 1995 05:02:54 +1000
Message-Id: <19963.811882974@munnari.OZ.AU>
From: Robert Elz <kre@munnari.oz.au>

    Date:        Sat, 23 Sep 1995 01:01:06 +0200
    From:        Eric Thomas <ERIC@SEARN.SUNET.SE>
    Message-ID:  <199509222327.TAA28462@CS.UTK.EDU>

    On Fri, 22 Sep 1995 03:02:29 -0400 Keith Moore <moore@CS.UTK.EDU> said:

    Case 1,  you are resending the message to a piece of software (a mail
    server) that will take action based on the contents.

This is actually a pretty boring, the automaton isn't going
to care, in your 1a case it simply makes no difference at all,
in the 1b case you probably shouldn't be forwarding a message,
but sending a new one, and the automaton should be just looking
at ordinary headers (that is, if you wanted the thing to be
looking at Resent headers, you probably shouldn't).   This is
not really an interesting case.

    Case 2,  you are resending to a human person.

This is the interesting one.

    While you do not want to annotate or edit the message in
    any way, you want the recipient to know that you resent it.

That's not quite how I would put it - if I wanted the recipient
to know I had resent it, I'd probably use an encapsulating
forward.   With Resent, its more like I want the recipient to
be able to discover I forwarded it, if they're interested enough
to look - that is, if they wonder why they received the message,
or something.   Otherwise it is best if the message just looks
as if it was sent to them.

It is important that the annotation appear for another reason,
that is, that I have actually looked at the message to decide
to forward it, the message should be marked in some way that
indicates I have seen it, it should probably be impossible for
a user to (easily) forward a message they have read (as opposed
to simply automatically redirected) without adding some kind of
marker to indicate that.   We should probably attempt to
deprecate UA's that allow simply wrapping an old message in a
new envelope and retransmitting without header alteration.

    Well, I contend that this is not an appropriate function.

I suspect almost everyone disagrees, it is very useful.

    If I forward a message to Joe such that it shows up as
    coming from me,

But it doesn't.   It shows up as coming from the original
sender, with a "redirected by" sticker on it, just as happens
when someone sends paper mail to your front office, and they
decide you're the right person to handle it, and pass it along.

    ... especially given that I can't annotate to bring
    this problem to Joe's attention.

The Resent-* headers are that annotation.

    All in all my impression is that the people who use Resent- fields for
    mail servers do so because there is no command in their mail program to
    forward a message "as is", without any header change.  I don't think the
    standard is the right item to change :-)

UA's should have no such function.   Ideally it should be
prohibited, unfortunately that would be meaningless, its not
possible to do.  We should certainly advise against making
anything like that easy - failing to add appropriate trace info
is a true evil.

    ... forwarding can be a complex issue.

Loke many things, this is because the is not one elementary
concept of "forwarding", it means different things to different
people.

    There's the well-known, silent forwarding where you just
    change the envelope address.

I don't think this is well known, it is certainly not something
I see frequently (if ever).

    I never use Resent-* for automatic forwarding.

Nor do I, I don't think we have any disagreement there.

    I'm just not sure we can make a generic statement about
    mail forwarding when there are so many specific cases.

Yes.  I think we could say that Resent-* headers should be
added when a human has viewed the mail, and wants to forward
it so the eventual recipient receives it looking basically as
if it had been sent directly to them.

I also believe we should say something about how to repeat
this - what should be done eith existing Resent headers when
a new set are to be added.

    >Alternatively, we could deprecate Resent-* altogether.
    In the long term, it may indeed become obsolete.

I actually doubt that.   It does have uses, even if being
applied for mych by default in a UA isn't one of them (the UA
could reasonably display a "This mail was relayed by xxx@yyy"
message when Resent-* headers are present.

kre