my point is that if the message is rejected a bounce will be
created, which is likely to be collateral spam.
Yes. And RfC 3834 replies, all "legacy" auto-responders, and
all broken-by-design C/R-systems cause more "collateral spam".
until someone figures out a reliable way to distinguish spam from
legitimate mail, any auto-responder is going to create some collateral
spam. there's no way around that. RFC 3834 does expect responders to
decide whether or not it's appropriate to respond to a message.
The essence of this problem is "bounces-to" versus "source" as
in <http://article.gmane.org/gmane.mail.spam.spf.discuss:14678>
no that's not the essence of the problem. and fwiw, you're wrong about
something you said in that article, in particular this:
It's extremely important to understand, that the MAIL FROM
address is not only some potentially unrelated Errors-To
address, it is THE source, THE sender, THE originator.
MAIL FROM is fundamentally the address where bounces go. It is NOT
necessarily any of the source, the sender, or the originator. There
are and always have been valid reasons for MAIL FROM to be set to some
other address than the source/sender/originator.
...and it's extremely important that you understand this.
Keith