[ietf-dkim] Responsibility vs. Validity

[John Levine]

>It has to be of the syntactic form of an <addr>. The usual implication of  
>such syntax appearing in a standard is that a thing that looks like an  
><addr> is expected to be an <addr>. If there is any other intent (e.g. it  
>was a Message-ID), then one would expect the standard to say so.

The local part is not a mailbox, and it's not a Message-ID.  If it
were a mailbox, it would say it was a mailbox, or if it were a
Message-ID, it would say it's a Message-ID.  We're not deconstructing
the Talmud here, the words mean what they say.

> The wording in 4871 strongly suggests, by implication, that the
> domain at least is supposed to be a domain.

Um, it's not an implication, its a MUST:

  The domain part of the address MUST be the same as or a subdomain of
  the value of the "d=" tag.

R's,
John