[ietf-dkim] user level ssp

[Hallam-Baker, Phillip]

> [mailto:ietf-dkim-bounces at mipassoc.org] On Behalf Of Michael Thomas

> The utility of "I sign some" is not in the policy itself, but in the
> *discovery*
> part of the protocol: when you find _any_ valid record, you 
> know that you can stop looking for one.  Depending on the 
> tree walking aspects of the discovery mechanism, this could 
> be a useful thing. Maybe it would be better to do this by not 
> expressing any  policy/practice in the otherwise valid to get 
> this functionality so as not to confuse the issue with the 
> semantics  of  "I sign some" which doesn't seem to mean much.
> 
> I have no idea what use "I sign no mail" has.

I suggest that we replace 'I sign no mail' and 'I sign some mail' with 'Undefined'.

A policy mechanism with two values is going to be much easier to administer than one with fve degrees of freedom.

One reason I want to insist on the binary choice here is that I want to encourage publication of the only policy that is useful to a receiver. If you allow for a weasel route you are going to have a hard time getting anyone to go all the way. Its like the problem we have with Draft standard and Standard here in the IETF.