From: Michael Thomas [mailto:mike(_at_)mtcc(_dot_)com]
Sent: Friday, April 29, 2011 4:37 PM
To: Rolf E. Sonneveld
Cc: Murray S. Kucherawy; ietf-dkim(_at_)mipassoc(_dot_)org
Subject: Re: [ietf-dkim] Output summary
Indeed, the chickens have come to roost. This was ill-conceived at the
time of the errata, and it is ill-conceived here. It is yet another reason
why I believe that the protocol described in 4871bis only bears passing
resemblance to 4871 and interoperation will be purely coincidental.
I don't agree. I don't know of any current implementations that would be
hampered by what's being done here.
An implementation, new or old, that provides both the SDID or AUID up the chain
is fully compliant with both versions. You would only have an interoperability
problem if you had written a DKIM verifier that exported only the AUID, or only
failed signatures, or something like that.
NOTE WELL: This list operates according to