On Wed, 11 Feb 2009, Jeff Macdonald wrote:
It would be equally valid for a signer to apply a different
pseudo-subdomain on each message, perhaps for tracking purposes.
I think that is actually a mis-use of DKIM. The message-id field covers
that nicely.
But Message-ID:'s semantics are defined in a different layer. As things
are currently defined, the DKIM module at either end can only make use of
what's in the signature itself.
My understanding of opaque allows identical opaque values to identify
the same "something".
Then you're arguing for something stronger than what the draft proposes.
The draft uses SHOULD, where to match your understanding, it would need a
MUST.
_______________________________________________
NOTE WELL: This list operates according to
http://mipassoc.org/dkim/ietf-list-rules.html