Actually, I've taken just the opposite tack: I'm now relying on SA for most of my spam filtering, and using TMDA as the "backup" to pick up any false positives that might slip through SA...I use the whitelisting/blacklisting fetures of TMDA heavily (much more user-friendly than the SA implementation), but no longer rely on the C/R features of TMDA for filtering mail. Too many valid e-mails were sitting unconfirmed in the queue, which leads me to believe the concept of C/R is still very foreign to (or maybe very resented by) many.
I've found this approach negates the need for generating keyword addresses for every conceivable sender. Anecdotal? Sure...but it's no less valid than saying that C/R is acceptable by the majority of e-mail senders. That's anecdotal as well, AFAICS. --Brian On Thu, Sep 16, 2004 at 02:45:15PM -0400, Godmar Back wrote: > I was contemplating a system that uses only a small subset of TMDA > features; and I'm wondering if it wouldn't be as efficient as TMDA in > preventing spam and helping people prioritize their email. In > particular, suppose you decided *not* to rely on sender verification > (which isn't available anyway) in any way whatsoever. Rather, you > require that all mail you read be sent to a keyword address. > Conversely, all email sent to your public address would be challenged > - unlike the way TMDA is currently used, however, a successful > response would also result in a keyword address being issued to the > responder that they must use for future communication if they want to > avoid repeated challenges. _____________________________________________ tmda-users mailing list ([EMAIL PROTECTED]) http://tmda.net/lists/listinfo/tmda-users
