Meng Weng Wong wrote:
the objection is detailed here:
http://spf.pobox.com/slides/molson/
What's CYA ? BTW, I've shown that PRA is a _real_ problem
with a legacy MUA behind a legacy MSA, and I've proposed a
simple way to fix it with the concept of a "default Sender:"
(if there is no Sender:) on the side of the recipient, more
than once.
some purists feel that a site's mail-from scenario may be
different from its PRA scenario
It's a fact, I use From: nobody(_at_)xyzzy with another MSA, and
of course this MSA insists on a MAIL FROM:<me(_at_)msa(_dot_)example>,
but it doesn't insert a Sender:, let alone patch the From:
it's important to let them make that distinction
Actually it would be good enough to treat the MAIL FROM as
default Sender: on the side of the recipient, because that's
the entity knowing all about PRA and Sender-Id.
The problem is RfC 2476:
| The MSA MAY add or replace the 'Sender' field, if the
| identity of the sender is known and this is not given
| in the 'From' field.
Before Sender-Id this MAY was good enough. With Sender-Id
it's a SHOULD on the border to a MUST. But if that's the
case and the problem, it can be also handled by the recipient:
"pretend that this Coke is Pepsi, if it's in a Pepsi glass".
(And maybe update RfC 2476 a.s.a.p.)
Bye, Frank (Coke classic fan, WTF is Pepsi)