> >(4) The Microsoft SMTP gateway uses the address from the SMTP MAIL FROM
> >command (which may differ from the address in the From header) as the
> >address in the Microsolft Mail From field.
>
> OK, before Eric gets all fired up on this, let me try for first blood :-)
> As a long-time postmaster and listserv co-ordinator, I'm TOTALLY DISGUSTED
> with the behavior of these "cutesy PC" mailers that think that such mail
> behavior is OK.  It is OUTRIGHT LAZINESS on the vendor's part not to parse
> the "real" mail headers for details rather than taking the TOTALLY INVALID
> transport layer information as representative of the mail item.  Until
> these newfangled mailers accept the fact that there is more to mail than
> a simple "from" and a "recipient list" this nonsense will continue.
>
> Not picking on you Roger, just venting some pent-up bile on the subject :-)
>
> [\] Jeff
 
I won't argue most of your point with you.  Microsoft Mail and others
of its ilk were never designed to be on the Internet.  Thus they don't
have Internet Mail concepts such as an envelope from address, reply-to,
etc.  Gateways are trying to make a square peg in a round hole.  There
is no right way to do the job.  There is hope that Exchange will be
better, but that's a ways off.
 
I do have to disagree that the Microsoft Mail SMTP gateway is wrong in
using the envelope MAIL FROM address in the MS Mail From header.  This
is necessary to insure that bounce messages get sent to the place that
Internet Mail standards call for.  The only other way I've heard of to
deal with that is for the gateway to have a data base of all valid MS
Mail users, so it can issue the bounces.  But that doesn't work as well
as one might like either.  Besides the problems of maintaining such a
data base, there are other kinds of bounces that MS Mail can generate
besides an unknown user.  As I said, there is no right answer.  Gateway
developers have the choice of violating Internet Mail standards,
inconveniencing the user, or both in some cases.
 
However, as Eric is fond of pointing out on IETF email working group
mailing lists, products such as LISTSERV have to deal with the real
world.  Thus my request for a configuration option.  Eric and I have
already corresponded privately about this.  But I felt that others
should be made aware of the problem so they don't have to figure it out
themselves.