On Tue, Jan 19, 2021 at 2:11 PM Douglas Foster <
[email protected]> wrote:

> No Murray, I was speaking to the PSD document.
>
> PSD's entire purpose is to detect abuse of non-existent organizational
> domains, so the definition of non-existent is crucial to its success.    I
> believe the current language will produce false positives, albeit probably
> a small number.    The current language is also more resource-intensive
> than mine, although that is not my concern.
>

What I mean is: If we say PSD experiment participants evaluate the notion
of "non-existent" differently than vanilla DMARC implementations, we have
to account for that when interpreting the results of the experiment.  But
the experiment as crafted is just to determine if the PSD algorithm as
proposed is a useful improvement.  It seems to me that changing the nature
of that test at the same time is scope creep that muddies the waters.

-MSK
_______________________________________________
dmarc mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dmarc

Reply via email to