Keith Moore wrote:
It's a perversion of RFC 822 which contradicts both the
intent and the way it has been used in practice.
It's probably clear that I hate PRA, because it's a kind of
FUSSP (worldwide "add Resent-Sender" at places where that
was never before required), because it claims erroneously
to be normally (default) "compatible" with Return-Paths
covered by a v=spf1 policy, and because it does not even
_reference_ 2476bis, let alone discusss 2476bis 8.1 details.
But I don't see where it's a "perversion of 822", IIRC the
author was present when it was evaluated in a MARID chat -
more than a year that I read the jabber log, I'm not sure.
If I get a mail Resent-From: C, Sender: B, From: A+B, then
822 tells me that A+B were the authors, B was the sender,
somehow C got it and resent it to me.
And if I then want to ask "why did you send this to me?",
I'd send that question to C. I don't need any "license" or
draft-lyon-senderid-pra for that task, but C is the "PRA",
the address claiming to be the sender.
So far no problem in _that_ draft. It could be a PS, if
they'd add some considerations for say mailing-lists. But
they say that mailing-lists and other redistributors MUST
also use that Resent-scheme in draft-lyon-senderid-core,
that's the point where senderid degenerates into a FUSSP.
But the "algorithm" itself to find an address of the last
entity claiming to be the sender of a mail (excluding all
"mediators" like mailing-lists on behalf of the receiver)
is sound. The recommendation that MUAs should display
that address is IMHO also okay (and maybe stuff for a PS).
Otherwise this PRA "algorithm" is only "informational",
it could as well be published as one of many answers in a
mail header FAQ. Nothing in draft-lyon-senderid-pra is
"experimental". It just reflects what (2)822 say about
Resent-From, Resent-Sender, Sender, and From.
So yes, the authors read and understood (2)822, and while
they were at it they published a summary of their findings
as draft-lyon-senderid-pra.
Bye, Frank
_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/ietf