On Feb 25, 2008, at 10:43 AM, Steve Atkins wrote: > > On Feb 25, 2008, at 10:17 AM, MH Michael Hammer (5304) wrote: >> >> While there might be some domains that take that stance I wouldn't >> expect "mainstream domains" to take that stance. I expect financial >> institutions to be somewhat early adopters and any ISP/mailbox >> provider that decides discardable means don't bother trying to >> deliver the mail should be prepared to beef up their Customer >> Service staff to handle complaints. >> >> There is a significant difference between sendedr saying "I >> recognize that a fraction of a percent of legitimate mail might not >> get delivered if my assertion is followed" and "sender doesn't care >> whether the mail gets delivered". > > They're the same statement. If they cared about mail they send being > delivered they wouldn't deploy something which will, by design, > cause some fraction of it to not be delivered (and provide no other > benefits).
"No not send DSNs upon signature failure" is not the same as "Expect this domain's messages to be "Signing Complete". A measure of delivery integrity includes receipt of Delivery Status Notifications upon failure to validate. Presumably, once DKIM becomes broadly implemented, back-scatter will be curbed by reduced success rates of spoofed messages. Rather than achieving consensus for the "discard" feature within the requirements phase, the required "Signing Complete" assertion has been replaced by a new and completely different assertion previously not discussed on this list. The usefulness of the "discard" assertion appears limited and perhaps unneeded once DKIM becomes more broadly adopted. These are not the same assertion. Make a case where both assertions are allowed, or return to the previously agreed "Signing Complete" that omits advice on verifier actions. -Doug _______________________________________________ NOTE WELL: This list operates according to http://mipassoc.org/dkim/ietf-list-rules.html