On Fri, 8 Jul 2005, Terry Fielder wrote:
For communications/messaging systems forwarder is a facilitator that
helps in establishing communication path from data origin to data
destintation, it does not replace who the data is from, merely helps
it on the way.
Agreed, 100%. But along the way it *must* identify itself.
I agree. That is why I supported SUBMITTER proposal, but not in the way
its being pushed by MS (i.e. being tied to incorrectly used Resent fields).
I think its useful concept on its own for such identification.
Hence why the RFC2822 From should be left as the original sender, and the
RFC2821 should be the forwarder identifying itself (unless of course its
known and trusted by the final recipient, then it doesn't need to identify
itself).
In the way RFC2821 FROM is used, it is more like bounce address, it may
have been meant to be different but that's what happened. In either case
even if we change it to be address of previous hop (forwarder) we would
still need "address of where to send bounces to" and I think address of
who the message is from originally is also useful at the SMTP session
stage without seeing any of the data.
Frankly I'm actually rather tied of this RFC2821 & RFC2822 identities
separation. A lot cleaner would have been to consider part of the header
data to be mail session and transmit it separately, then rest of the
header being trace info and last content body being actual data.
--
William Leibzon
Elan Networks
william(_at_)elan(_dot_)net