[MlMt] Too many bounces?

[Bill Cole]

On 28 Jan 2019, at 3:35, Benny Kjær Nielsen wrote:

[...]
> I'm now at the point where I would automatically become suspicious if 
> anyone (any webpage) seems to claim to fully understand the 
> consequences of any kind of setup/setting related to this :-)

There are definitely some people (maybe dozens) with very deep 
understandings of DMARC and the consequences of all of its options, but 
they know better than to think there's a universal right way to deploy 
it trhat can be boiled down to a web page. I'm not one of those people 
but I know a few of them, and none are fond of general p=reject use, 
especially for domains used to provide mass-market retail/free 
mailboxes. Everyone who participated significantly in the DMARC 
definition process knew that all existing discussion group mailing lists 
operating across independent domains would be damaged by unwise use of 
p=reject. I'm pretty sure that some people see that as a feature, either 
because traditional mailing list behavior is inherently problematic and 
needs an incentive to change OR because they see an opportunity to 
advantage their own captive discussion groups. It's not accidental that 
the first significant mailbox provider to use p=reject was Yahoo.

The only real fix for mailing lists is to munge From headers, at least 
for list members who have p=reject domains. Disabling bounce processing 
is an unsustainable option and banning users in p=reject domains is 
impractical.

-- 
Bill Cole
bill at scconsult.com or billcole at apache.org
(AKA @grumpybozo and many *@billmail.scconsult.com addresses)
Available For Hire: https://linkedin.com/in/billcole