Dave Crocker wrote: > Tony, > > TH> a legitimate message from someone I have corresponded with in the past. > The > TH> only way to detect a fraud at the MUA would be to have a verifiable > TH> signature from Alain (this was trapped at my MTA due to the exe file). > > yes, but no. > > first, there is an increasingly heated debate between folks who want to > sign the message (TEOS, DomainKeys), versus others who want to secure the > channel between > sender and receiver (RMX, LMAP, SPF, etc.). Is there an obvious reason not to do both? Neither is a total solution, but either ties a message to an identity object. It is time to stop fighting over which is better and put both approaches out there. The market will decide if one is easier for Joe-sixpack to use. > > Once that debate is resolved, there is still the matter of compromised > system. The message might actually come from the purported author's > system, but still not be from the author because it has been taken over > by evil forces. So, even with perfect automated validation, the content > still might not be valid. Compromised systems are a problem, but the scope of the bogus mail originators is limited to the users of the compromised system. Without traceability it comes from everywhere, but with traceability at least one knows where to go to correct the problem. If we are stuck fighting over the 'perfect' solution, we will never get anywhere. The engineering community is particularly bad at figuring out what will catch on for the lowest denominator consumer, and a committee of engineers takes a bad situation and makes it hopeless. Tony > > > d/ > -- > Dave Crocker <dcrocker-at-brandenburg-dot-com> > Brandenburg InternetWorking <www.brandenburg.com> > Sunnyvale, CA USA <tel:+1.408.246.8253>