Re: Enough DMARC whinging

[Dave Crocker <[e-mail]>]

On 5/1/2014 8:22 AM, Phillip Hallam-Baker wrote:
> Spam filters should know about things as important as mailing list
> subscriptions.

We have about 20 years of spam filtering experience in the industry.
The quality of modern filters is astonishingly good; it's the only
reason email remains viable for users, in spite of open-Internet spam
traffic being far above 90%.

I believe that little or none of that filtering includes awareness of
mailing list subscriptions.  So while the above is an interesting and
possibly useful line of pursuit, it's not clear how it can be elevated
to the status of 'should'.

Note that historically, mailing list operators have been resistant to
the imposition of technical or operational changes.


> It the mailing list has appropriate spam ingress controls, is
> authenticated using DKIM and there is evidence that the user has
> subscribed then the spam filter can whitelist all the messages from
> that list.

Nope.  Mailing lists can be sources of spam, too.  Even good lists.  So
blanket whitelisting would not be advisable.

However modern filtering engines are complex enough for nuance.  So they
can ratchet up or down just how stringent the criteria are. The above
scenario well might warrant less stringent criteria.


> And to the other conversations, we are talking about draft- here. And
> that isn't the same as standard. In fact one of the requirements for
> being granted standard would be to come up with answers to these
> issues.

Standards status matters, but not as an absolute.  The fact that the
text exists as an I-D also does not automatically impart importance or
utility, in terms of industry adopters.

Often it does, of course, but this doesn't seem to be one of those
cases, does it?


Sometimes we forget that the IETF is in the service of adopters, not
vice versa.

d/

-- 
Dave Crocker
Brandenburg InternetWorking
bbiw.net