Re: [ietf-smtp] DSNs

[Valdis Kl ē tnieks <[e-mail]>]

On Sun, 26 Apr 2020 17:59:18 -0400, Scott Kitterman said:

> > The other big problem is that even if we chunk out an RFC that clarifies
> > that they *can* pick-and-choose which DSN events to react to on a case by
> > case basis, it's still going to take a decade before people actually do
> > something about it.
>
> I think that's backwards.  People already have done something about it and do
> pick and choose.  The question is if and how we should align standards to that
> reality.

I was unclear.. I meant specifically the "We turned it off because we thought that's
what we had to do to comply" crowd.   Once it gets in config files, it's there *forever*.

(Not SMTP, but demonstrating the problem - I had an IP address for an AIX
workstation on my desk that was running a stratum-2 NTP server. That IP address
was removed from the clocks.txt file in late 1998, and went permanently offline
in June 1999 when we moved from "workstations on desks doing servers" to proper
rack-mounted boxes with all the bells and whistles.

In July 2018, a test over 6 hours found 501 machines that *still* thought that IP
address was a good place to send NTP requests...   Gaak.)