On Fri, 17 Aug 2007 20:31:51 +0200 Iljitsch van Beijnum <iljitsch@xxxxxxxxx> wrote: > On 17-aug-2007, at 17:54, Steven M. Bellovin wrote: > > S/MIME would be a fine start. It also won't solve the problem until > > someone develops a user interface that DTRT for naive users who > > don't understand trust anchors, > Big yellow warning when S/MIME authentication fails in Apple's Mail > is hard to miss even if you don't understand exactly what it is... You'd be surprised what people will miss... You also have to account for people missing the presence of S/MIME, i.e., the bad guy just sends the email without any protection and hopes folks don't notice. > > > or how to distinguish myfinancialcompany.com from > > myfinancia1company.com when both have valid certificates. > > So I can register paypa1.com and then go to Verisign to get a > certificate for that domain? If that's true, then I think the law > makers in various jurisdictions have work to do. Given that paypa1.com was the very first phishing attack I saw, and that there was a cert... More recently, see http://blog.washingtonpost.com/securityfix/2006/02/the_new_face_of_phishing_1.html > > The very simple idea of having a .bank TLD for financial institutions > could also help a lot here. > Same failure modes. --Steve Bellovin, http://www.cs.columbia.edu/~smb _______________________________________________ Ietf@xxxxxxxx https://www1.ietf.org/mailman/listinfo/ietf