Message-Id: <26010121484612@vms3.macc.wisc.edu>
Date: Mon, 01 Jan 96 21:48 CDT
From: "Eric Norman (MACC)" <EJNORMAN@macc.wisc.edu>
Subject: Re: Header to label lists (was untitled)
To: jgm+@CMU.EDU
CC: drums@cs.utk.edu

> I have heard numerous people claim a need for a header to identify
> that a message has undergone list expansion, but I have yet to hear a
> clear, rational statement of what problem such a header is supposed to
> solve.

If you have a header that clearly establishes that this message was
a result of expansion of a mailing list, you can:

(1)  Safely implement a "reply to list" function in UAs.  There are
     arguments that Reply-To: should not be used for this purpose; not
     that I agree with them.  In any case, Reply-To: may have been usurped
     before the message got the the list expander.

(2)  Have a rule that any program that automatically responds to
     mail messages without human intervention (e.g. vacation programs)
     MUST NEVER respond if such a header is present.

(3)  Detect loops in mailing lists that end up pointing back at
     themselves somewhere along the line.

(4)  Give agents a method of not bothering list owners with "been
     trying 1 day; will try 7 more" messages.  I agree that NOTARY
     if probably a better way to handle this one when and if it's
     successfully deployed.

In general, since mailing lists are so common and seem to cause
so amny problems, I consider a header that clearly indicates that
a message passed through a mailing list to convey very useful
information; I personally consider it more useful that other extant
headers (e.g. In-Reply-To).

For what it's worth, I'll voto to call such a header "Expansion-Of".
I could argue that I think it makes more sense, but the real (silly)
reason is that I wouldn't have to change a constant in my code and
recompile.

Eric Norman