On Tue, Jan 27, 2009 at 8:16 AM, Dave CROCKER <dhc(_at_)dcrocker(_dot_)net>
wrote:
note your reliance on "ISPs aware of this i= classification". It means that
the mechanism won't scale, since we cannot assume that an arbitrary receiver
will be aware of an arbitrary signer's scheme.
The nice thing about i= is that it doesnt have to mean a thing except
to the sender emitting it.
The sender should not count on recipients knowing about i= or caring
about it enough to matter.
So - if a sender wants to sign i= besides d=.. they can count on a
feedback loop based on d= coming back to them and then separating this
loop email easily into multiple client domains based on the i= they
have inserted. That's entirely internal to them and none of the ISP's
concern, if they choose to ignore that i=.
And if the ISP wants to sign i= to separate their mail streams up to
the receiver to decide whether they want to filter based on it or not.
i= can be freeform and any arbitrary value for all that I care (e&oe
some due diligence to stop mta exploit attempts being stuffed in i=)
--srs
_______________________________________________
NOTE WELL: This list operates according to
http://mipassoc.org/dkim/ietf-list-rules.html