Re: [ietf-smtp] SMTP, DSNs, and enhanced replies

[Ned Freed <[e-mail]>]

> On 4/9/2020 4:00 PM, John Levine wrote:
> > My impression is that these days if you get a bounce it's usually a DSN.


> Right.  I thinking of the SMTP mechanism for a client's /requesting/ its
> use, which as far as I has not gained widespread use.

Widely supported, not widely enabled. THe main issue IME is success DSNs. The
way the standard reads you have to support them or be noncompliant. i tell
people security needs override that and to simply reject messages that request
them for security reasons.

Including the SUCCESS machenism was a mistake, but we were quite concerned
about feature parity with X,400 at the time - yes, this is that old - and
blowback wasn't on the radar.

But perhaps the real reason is that none of the mechanism the extension
provides have proven to be essential. Most applications for envelope ids and
original recipient addresses have been implement with more reliable VERP schems
instead, the ability to turn off DSNs completely has always been present, and
control over return of content is (a) Less useful because message sizes aren't
as relevant, (b) Something you don't want th sender to have control over
because of phishing/viral content, and (c) Something you don't want people to
be able to do for really large messages.

				Ned