On Mon 01/Aug/2022 21:15:36 +0200 John Levine wrote:
It appears that Alessandro Vesely <[email protected]> said:
You could make the argument that you "usually" find DMARC records at zone
cuts, but that's relying on convention or happenstance rather than a
standards-quality framework.
I'd agree, except that our definition of Organizational Domain as one
label beyond a PSD (PSD+1) poses a semantic limit on that.
Sorry, but this is beyond wrong.
A) Org domains are defined by what is in the draft, and that is not
what the draft says. Some org domains will found below a PSD, most
won't since most PSDs do not and will not publish DMARC records.
Oops, you're right. The PSD+1 sentence was removed in version -13.
B) PSDs have nothing to do with zone cuts. Some are at zone cuts,
some are not. For example, .uk .co.uk and .org.uk are all in the PSL,
and they are all in the same zone.
But those are PSDs, not org domains...
While we concentrate on existing scenarios, it may well happen that,
if DMARC algorithm sees widespread adoption, people will define
organizational boundaries that way, irrespective of actual DNS
control, as in your example.
You appear to be saying that people will ignore the spec and do
something else because they somehow imagine that we didn't mean what
we said and really meant something else. Aw, come on.
Eh? I don't know how you could extrapolate that meaning from what I said.
I didn't say people will ignore the spec. I said "if DMARC algorithm
sees widespread adoption" because that is what can make the spec come
true.
In that case, I said, "define organizational boundaries that way" can
become a practice. "That way" refers to the way Murray exemplified,
what else?. That is, following the spec.
The "while we concentrate on existing scenarios" was meant to
criticize our dismissing arguments because no case currently exist
that matches. For example, we fix the max depth at 5.
Best
Ale
--
_______________________________________________
dmarc mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dmarc
