Tony Finch wrote:
On Wed, 13 Jul 2005, Jim Fenton wrote:
1. The suggestion that the DKIM-Signature be treated as a trace header field
needs to be a suggestion because it's hard to retroactively rule a
previously-unknown header field is a trace header field.
The way I view this kind of normative text is that it can only apply to
implementations that support the current specification, i.e. it implicitly
does not retroactively add new requirements that old implementations
cannot satisfy. Therefore in this case it is sensible to use "MUST".
The problem is that intermediate MTAs, which may not be DKIM-aware, may
be the ones that are not treating signatures as trace headers. I didn't
want to make it seem as though we were depending on the order of the
signatures. I agree that it is reasonable to require DKIM-compliant
MTAs to maintain signature order, though.
-Jim