[mail-vet-discuss] "none" as a possible result string?

[Michael Thomas]

J D Falk wrote:
> msk wrote:
> 
>> Someone has asked for "none" to be added as a possible result for an
>> authentication method.  An example use of this might be a message not
>> signed with DKIM, and a DKIM verification agent that wants to be very
>> explicit that it processed the message and that the message was
>> unsigned. One could argue though that this is redundant to "neutral",
>> which we already have, or that the implementor would simply omit to
> add
>> the header altogether in those cases.
>>
>> In any event, I thought I'd offer it up for discussion while there's
>> still a chance to make the change if someone is strongly in favour of
>> such.
> 
> I think it's a good idea.  This header is (as I see it) primarily
> intended for post-processing of upstream results, so there can be an
> important difference between verdicts of "processed, no result" and "not
> processed at all" (denoted by a complete lack of the header.)

I think I agree, but it should be made very clear that the difference
should be only of forensic value. Ie, don't build a kind-of locked
icon for a broken signature or some such other silliness that might
give users the wrong impression, yadda yadda.

		Mike