On Sat, 2005-11-19 at 11:30 -0500, Scott Kitterman wrote: > What you are saying is that just because a message meets an SSP requirement > is not a safe basis for an MUA marking them somehow good. I agree with > that, but I think it's outside the scope of what this almost working group > is supposed to do.
Of course design of an MUA would be beyond the scope of the DKIM WG. Facilitating security in the face of newer characters-sets and look- alike domains remains a desirable feature made possible by a binding approach. > IIRC, the farthest in that direction we go is an optional task for a header > to communicate DKIM results. The binding assertion would simply be an option within the base DKIM draft. This would not require an additional draft, however there could be some informational drafts to describe how to use this feature. > My view of restrictive SSPs is that messages that fail the restrictive > test should be rejected during the SMTP session. This will reliably > get the rejection notification back to a legitimate user and keep it > out of any bad message folder I have to periodically review. Automatic bindings could offer the same level of protection at the MTA without risking the side-effects produced by authorization records. > I think you miss the point about the potential value of restrictive SSPs to > the receiver. A binding recognition strategy does not forgo this style of protection. After many years, caching "broad" binding at the MTA/MDA could be depreciated. > I don't need better methods to sort messages into folders. > I will need better methods in the future to avoid having to deliver bad > messages at all. No mechanism, no matter how complex, will prevent Bad Actors from sending their messages. Don't even suggest SSP will reduce the number of bad messages! The DKIM signature will be useful at locating the source of abuse. The DKIM signature, in combination with "binding recommendations" can reduce a much wider range of spoofing without the recipient needing to pass an eye-test. > OK. So, bottom line is that you aren't wrong, but I think your 'threat' is > based on a false premise and out of scope. You agree that SSP does not provide a mechanism to prevent spoofing without reliance upon visual presentations, but that a scheme which avoids this reliance as an option within the DKIM signature is out of scope? The "broad" binding mode would offer the same ability to reject messages at the SMTP session as would the SSP 'o=!' policy, but in microseconds rather than seconds. -Doug _______________________________________________ ietf-dkim mailing list http://dkim.org
