It appears that Richard Clayton <[email protected]> said: >We should also discuss removing "donotexplode" -- for which I think a >strong case has been made (basically, there's lots of sites which >operate role addresses and are going to explode and send offsite)
I think we could come up with a plausible distinction between forwarders that expand role accounts and mailing lists, but I also don't see how it would be of much use to verifiers so I agree it should go. As I understand it, our main defence against replay is that replayers can't sign as the address they're replaying from so the chain is broken. >The case for removing "donotmodify" is weaker ... I see the argument >that its just as aspiration by the sender but it is not the end of world >if mailing lists or alumni forwarders avoid modifying mail that should >not be passing through them anyway (such as account recovery info, >invoices and so on). Again, I could see a plausible definition as don't modify anything important, where important has the Potter Stewart definition.* At this point I'm not sure whether that would be useful. R's, John * - "I know it when I see it" _______________________________________________ Ietf-dkim mailing list -- [email protected] To unsubscribe send an email to [email protected]
