On 9/29/2020 6:54 PM, Dave Crocker wrote:
On 9/29/2020 3:41 PM, Hector Santos wrote:
Do you have an algorithm that replaces the current one?
I've no idea what any of your note has to do with the DKIM protocol
specification.
wow.
By way of a small example, DKIM does not have o=.
Right, you were instrumental in attempting to "separate" policy from
DKIM to create DKIM-BASE, a success, it allowed progress to be made
with DKIM, but it never separated the signer::author identity
association primarily because, once again, DKIM-BASE is still
inherently bound to the 5322.From field. You never separated the DKIM
anchor identity and it was stated many times, until then, we will
always have the signer::author relationship and policy protocols based
on this relationship.
Until it is changed, DKIM will always have this self-signed
signer::author relationship. That goes back to DomainKeys with o=,
early DKIM with o=, removed in DKIM-BASE as you gracefully pointed out
but it moved to ADSP (now DMARC).
But really, nothing in your note concerns the published and approved
specification.
Published and approved, yet seeking further comments. From I had
already read and understood from the start, all in once sentence:
Extract 5322.Sender, if found, use this for DMARC lookup, if not
found, fall back to 5322.From
Correct? Anything else?
The only systems that this will work with is compliant downlink
receivers. Non-compliant receivers are still a problem. At the end
of the day, the Mailing List Server (MLS) still needs to support DMARC
on the inbound side.
--
Hector Santos,
https://secure.santronics.com
https://twitter.com/hectorsantos
_______________________________________________
dmarc mailing list
dmarc@ietf.org
https://www.ietf.org/mailman/listinfo/dmarc
