Errors-to: owner-drums@CS.UTK.EDU
Message-Id: <199509251702.NAA08722@augusta.math.psu.edu>
To: drums@CS.UTK.EDU
Subject: Re: note on Resent-* discussions 
In-reply-to: Your message of "Mon, 18 Sep 1995 20:02:28 EDT."
             <199509190002.UAA06070@wilma.cs.utk.edu> 
References: <199509190002.UAA06070@wilma.cs.utk.edu> 
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Date: Mon, 25 Sep 1995 13:02:09 -0400
From: Dave Barr <barr@math.psu.edu>

In message <199509190002.UAA06070@wilma.cs.utk.edu>, Keith Moore writes:
>When talking about things like Resent-*, it might be helpful 
>if we distinguish "automatic forwarding" (where a recipient
>says "forward all of my mail to address FOO"), and "manual
>forwarding" (where a recipient says "I'm going to forward
>this message to Bob so that he can act on it and reply to it.")

Whenever discussing email I often try to liken things into
the "real world".  This makes things easier to remember, and
I think results in a system which is easier for the naive and
new to understand.

I liken Resent-* to the scratching out of an "old" address and writing
in the new address on the envelope of a message.  The message isn't
opened up and the address inside scribbled out, rather notice is
given on the outside that so-and-so has a new address and that this
message is to be redelivered.  Nor is the whole message put into a new
envelope with a new address and postage.

No matter if the forwarding is automatic or manual, I think there is
a strong case that Resent-* is applicable.  Sometimes I wish
sendmail did this for .forward's, so the message can be better tracked.
(though nowadays Received: often now has enough information encoded
in, thanks to sendmail v.8.)

And so, I disagree with those who think better encoding of forwarded
messages (via MIME message/rfc822 or whatever) is an acceptable
replacement for Resent-*.  All the person wants to do is redeliver
a message -- simply scratching out the wrong address and putting
in a new one should be sufficient.  If the person wants to add
some explanatory text, fine -- then use MIME or whatever.

I think the DRUMS text should include a less ambiguous description
of Resent-*, including examples of acceptable current practices.

--Dave