On Monday, September 13, 2004, at 10:52 AM, Paul Iadonisi wrote:
-- Jim Lyon
mailto:JimLyon(_at_)Microsoft(_dot_)Com
tel:+1-425-706-0867
Internet commerce will never really take off until you can buy something
online without getting spammed by the vendor.
-----Original Message-----
From: owner-ietf-mxcomp(_at_)mail(_dot_)imc(_dot_)org
[mailto:owner-ietf-mxcomp(_at_)mail(_dot_)imc(_dot_)org] On Behalf Of Paul
Iadonisi
Sent: Monday, September 13, 2004 10:52 AM
To: ietf-mxcomp(_at_)imc(_dot_)org
Subject: RE: TECH-ERROR: SenderID sets recomendation for forwarders
thatare not compatible with RFC 2822
It [Resent-From] is not for 'manually or automatically'
re-inserted messages, but only manually, given that the
automatic re-insertion is excluded from using Resent-*
fields according to the above line I've quoted from
section 3.6.6 of RFC2822.
I think the bright-line test to decide whether Resent-From is
appropriate is:
Is the action occurring at the desire of the mailbox owner, or is
it occurring at the desire of the mail infrastructure owner?
Otherwise, using the reading that you propose yields an absurd and
unimplementable result:
If my MUA has a "Resend" command, and I manually execute it, there
should be a "Resent-From". But if an automated script executes my
MUA's "Resend" command, there shouldn't be a "Resent-From". This
appears nonsensical on its face, so an alternate interpretation is
required. (It also begs the question of whether my secretary
pressing
the button constitutes "manual" or "automatic", when he's doing so
under my standing orders with no discretion.)
You need to read specs like a lawyer, not like a programmer.
(Programmer: if the spec says "jump", I'll jump and damn the
consequences. Lawyer: if the spec says "jump" and that's either
ambiguous or nonsensical, I'll look beyond the words to the intended
meaning.) Admittedly, the best specs are ones that yield the same
result to both programmers and lawyers.
-- Jim Lyon