Suresh Ramasubramanian wrote:
Mike Adkins had an interesting suggestion where he'd use d=aol.com to
sign outbound messages, and i=good(_at_)aol(_dot_)com, bad(_at_)aol(_dot_)com
suspect(_at_)aol(_dot_)com to denote various reputation scores that aol's
assigned an outbound message.
This would allow aol to separate mail into different streams at their
end, and it'd allow receiving ISPs aware of this i= classification to
specialcase their filters to watch for i=bad or i=suspect email from
and why is that better, or worse, than d=good.aol.com, d=bad.aol.com,
note your reliance on "ISPs aware of this i= classification". It means that
mechanism won't scale, since we cannot assume that an arbitrary receiver will
aware of an arbitrary signer's scheme.
NOTE WELL: This list operates according to