Despite the discussion, the text still seems to say nothing about the case of a PSD=Y policy on the same-domain policy.
A guiding principle of the move to tree walk is to get accurate boundary data by giving domain owners the ability and responsibility to publish those boundaries. That means that every domain owner needs a way to explicitly state "this is my organizational domain". The suggestion to ignore an initial PSD=Y does not meet this requirement. For the case when PSD=Y and PSD=N, ignoring both tokens means that the walk proceeds out of the organization on expectation that the presence or absence of specific data, from outside of the organization, will steer the evaluator back into the organization. I had suggested that we could say that PSD=Y requires exact alignment. This should be no problem for PSDs, because the parent of a PSD is also a PSD, and therefore should also publish PSD=Y or publish nothing. This does inconvenience about 200 private registries. But if we do not want to inconvenience the private registries, then we are back to confronting the problem we created for ourselves by choosing a Y/N indicator to document roles that are not mutually exclusive. What is the plan for initial PSD=Y?
_______________________________________________ dmarc mailing list dmarc@ietf.org https://www.ietf.org/mailman/listinfo/dmarc