[dmarc-discuss] policy overrides and 'what-if'

[Paul Midgen]

I would add the general guidance to Ben and others seeking to do the same thing that the policy override is the result of receiver-side local policy, subject to change without notice.

So, one could model the behavior of individual receiver systems (e.g. in case X Hotmail does Y and Gmail does Z) but I think it's probably a better idea to:

a) assume that the policy override decision is independent of p=, AND
b) provide some kind of "this is a best guess and YMMV based on receiver" advisory text.

-p

-----Original Message-----
From: Michael Adkins [mailto:madkins at fb.com] 
Sent: Thursday, April 05, 2012 10:33 AM
To: Ben Clifford; Paul Midgen
Cc: dmarc-discuss at dmarc.org
Subject: Re: [dmarc-discuss] policy overrides and 'what-if'

Ah, interesting point.  As Paul mentioned, there will be a list of all the expected override types.  It might take a little more operational experience before we can conclusively say we've covered them all, but I think we should make the effort to document an exhaustive list in the spec.

The answer to your question seems like the sort of thing we should include in an operational guidelines document for domain owners, but I don't think we've documented them yet.




On 4/5/12 5:30 AM, "Ben Clifford" <benc at cqx.ltd.uk> wrote:

>
>On Apr 5, 2012, at 8:03 PM, Paul Midgen wrote:
>
>> I may be missing your point about the information you'd rather have, 
>>but when a PolicyOverrideReason is present, the disposition associated 
>>with it is the result of the override. For example, if you have 
>>p=reject and mailing_list is present, you could expect to find 
>>disposition == none. Are you asking about something else?
>
>
>I'm asking about how I can infer what the disposition likely would be 
>in a what-if.
>
>Hypothetically:
>
>I have p=none, but I am doing a what-if for p=reject.
>
>I see records with disposition=none, and DKIM and SPF all failing, and 
>policy override reason = mailing_list.
>
>Intuitively I can say "ok, this is safe - it will still be a 
>disposition = none when you change the policy".
>
>I see records with disposition=none, and DKIM and SPF all failing, and 
>policy override reason = sampled_out.
>
>Intuitively I can say "oh dear, if that message wasn't sampled out, 
>this message would be a disposition = reject".
>
>The only difference in those two reports is the policy override reason 
>token.
>
>So where do I get the meanings for how I should do the above two 
>what-ifs differently?
>
>Should I be expecting to get that from the spec, or should I be trying 
>to maintain my own table based on my own observations?
>
>Ben
>
>--
>
>
>_______________________________________________
>dmarc-discuss mailing list
>dmarc-discuss at dmarc.org
>http://www.dmarc.org/mailman/listinfo/dmarc-discuss
>NOTE: Participating in this list means you agree to the DMARC Note Well 
>terms (http://www.dmarc.org/note_well.html)