[ietf-dkim] Proposal for specifying syntax and semantics for multiple signatures

[Eric Rescorla]

Dave Crocker <dhc at dcrocker.net> writes:

> Stephen,
>
>
>> So, in an attempt to move towards that,  let me
>> try to ask for opinions on this discrete part of
>> the issue: When an n-th signature(*) is added by
>> some signer, does that mean:
>> a) I take independent responsibility for having
>>    sent/fowarded (the bits of) this email (that I've
>>    signed), or,
>> b) Me too, whatever that previous good signer
>>    meant - it's not gotten worse.
>
> The meaning of a DKIM signature has been kept intentionally minimal
> and vague. Something like alternative (b) moves towards complex
> relationships among signers, whereas (a) retains basic simplicity.
>
> So, I hope (a) is the choice.
>
> To the extent that a signer includes the DKIM-signature header of a
> previous signature, that ought to mean nothing more than ordering, not
> semantics.

I have to agree with Dave here. (b) sounds way too confusing.

-Ekr