I have stumbled into the same position as Barry, or maybe a more extreme one. I send no bounce messages, and very few 5xx responses.
In my environment, one of the more common causes of non-delivery is a terminated employee. It has not seemed like automated messages sources pay much attention to NDR reports, so sending them seemed like a waste of time. I tried reject during SMTP based on sender verification, but it went badly because of a software bug, and a manager silently lost subscription to a critical service. This leaves me reluctant to try again. I also see clear evidence of spammers trying to do directory harvesting by guessing account names, and I don't want to help them. More generally, there is the need to perform both Sender Authentication and Sender Reputation checking when trying to decide if any particular sender is worthy of a notification. It just became easier to block silently. Doug On Mon, Jul 24, 2023 at 4:10 AM Alessandro Vesely <ves...@tana.it> wrote: > On Sun 23/Jul/2023 22:12:55 +0200 Barry Leiba wrote: > >> Without bounces the sender is in the dark. > > > > Yes, if the sender is a human. > > > > Not so, if the sender is a mailing list and that sender will then > > unsubscribe the intended recipient. > > Also not so, if the sender is a malfeasant who may use the bounce > > message for bad purposes. > > > > It's very clear to me that if I think a bounce message will be > > harmful, I will not send one. I will happily prefer silent discard > > over a bounce when I think that's a better approach for that > > situation. Bouncing a legitimate mailing list message is just bad. > > If you have reason to believe you're going to do that... don't. > > Either deliver the message or silently discard it. But don't bounce > > it. > > > Living aside the malfeasant case for a moment, do you think the worthiness > of > bounces can be stated depending on the type of sending? Always? > > The only meaningful signal a mailing list can get out of a 5xx response is > to > deduce that that mailbox doesn't exist any more, so it must be > unsubscribed. > > For alias expansions, there is the case that the author can appreciate to > learn > that her correspondent's mailbox is full, or that her writings was > considered > spam. But again, consider that friends most likely write directly to the > target mailbox, while newsletters deserve the same treatment as mailing > lists. > > Is List-Unsubscribe: an indicator of drop vs. reject? > > > Best > Ale > -- > > > > > > _______________________________________________ > dmarc mailing list > dmarc@ietf.org > https://www.ietf.org/mailman/listinfo/dmarc >
_______________________________________________ dmarc mailing list dmarc@ietf.org https://www.ietf.org/mailman/listinfo/dmarc
