Re: [dmarc-ietf] 5.5.4. Publish a DMARC Policy for the Author Domain - dmarcbis-06

[Alessandro Vesely <[e-mail]>]

On Wed 06/Apr/2022 01:01:46 +0200 Scott Kitterman wrote:
> On Tuesday, April 5, 2022 12:22:57 PM EDT John R Levine wrote:
>>> Scott took the time to define PSDs and PSOs in RFC 9091, restated in
>>> Sections 3.2.8 and 3.2.9 of the current draft.  Since the definitions of
>>> Organizational Domain (both the current 3.2.7 an my proposed change)
>>> require PSD + 1, a PSD has to be a proper subdomain of another PSD in
>>> order to admin an Organizational Domain itself.
>> 
>> I don't agree, but whatever.


PSD + 1 != PSD, since 1 != 0.  No need to agree or disagree...


>>> _dmarc.ac.me TXT "v=DMARC1; p=quarantine; adkim=r; aspf=r; fo=0; pct=100; rua=mailto:[e-mail]"
>>> ac.me mail is handled by 10 mail.ac.me.
>>> ac.me TXT "v=spf1 mx ip4:89.188.43.10 ip6:2a02:4280:0:200:89:188:43:10 -all"
>>>
>>> All of the above admit no org domain.
>>
>> If I have a message from [e-mail] and it has valid SPF of mail.ac.me, is 
>> that aligned?  Why or why not?
>
> I think for a PSD that sends mail, the current definitions would require Mail 
> From for SPF or d= for DKIM to be the same because they wouldn't end up having 
> the same org domain.  It might be simple enough to say to ignore psd=y or a 
> lower level psd=n if the 5322.From has a domain with psd=y (they can always 
> set strict alignment if they don't want lower level stuff to match).


+1

OTOH, if the From: domain is without psd=y, then any identifier having psd=y 
cannot be aligned.


Best
Ale
--