First let me try to wind this up (for now, we all know it'll resurface
again) by pointing out that my original comment a week ago that
(re)started this discussion said "IF ... and IF ... THEN ...".  The
IFs *were* capitalized there.  One of those major IFs was "If all
Bitnet mail delivery was via (B)SMTP.  I don't expect LISTSERV to
become completely RFC822 compliant until the transport is reasonable,
which it currently isn't.
 
Second, I _did not_ propose keeping all the Received: lines in the
messages that LISTSERV receives and resends.  I have modified our copy
of the UA for UCLA/Mail to support suppressing user-specifiable
header fields.  I personally suppress Received: and a number of other
normally junk lines.  I think that LISTSERV should keep the complete
original message in the logs, in case there's a question of how it got
there, or why it was delayed.  (Ok, who's going to be the first to
complain about the extra disk space?)  For various reasons, I have at
least one of my subscriptions set to FULLHDR, so I get all the junk
anyway.
 
Third, I didn't have any fights with the "Internet freaks".  I just
described for them the function of LISTSERV, and it's use of Sender:,
and asked for their interpretation of RFC822.  RFC822 very explicitly
states that Sender: should "not simply include the name of a mailbox
from which the mail was sent."  I propose that LISTSERV set Sender: to
either "LISTSERV@wherever", (which might require adding some X-field
to identify the message origin?) or else "listname-ERRORS@wherever".
I like this last solution, but the VM-ers are going to complain about
it being too much trouble to create the fake mailing addresses...
BTW, the Bitnet redist of the Space Digest has Sender: pointing off
somewhere like "Space-errors" or something.
 
/Leonard