On 06/Oct/11 03:37, Frank Ellermann wrote: > The concept of a FAIL report confuses me, because it's at odds with > the SPF FAIL semantics. As you said this would anyway only affect > receivers accepting FAIL.
There seems to be a semantic crossover here. On the one hand, some operators understand SPF as a means to ensure proper deliverability of any bounce that might occur, avoiding backscatter. On the other hand, ARF can be understood as an extension of DSN designed to overcome its restrictions, both target address and content. If an MTA rejects a message because of SPF FAIL, this may result in a bounce issued by the relay. OTOH, the same MTA may notice that domain owners are providing a fixed bounce address for SPF FAIL, and thus may safely issue the bounce itself. That is to say, we may alter the semantics of FAIL so as to allow both choices on equal standing. For example, some domain owners might be interested in tracking abuse of their domain name for forensic purposes, which may result in a stronger anti-spam effect than just avoiding the delivery. _______________________________________________ marf mailing list [email protected] https://www.ietf.org/mailman/listinfo/marf
