From what I've read on this forum, I get the impression that RBLs are pretty effective, if used judiciously. They stop a lot of spam (I'm not sure what percentage, but apparently not close to 100%) and generate few false positives. So I think I might try them out.
I've read elsewhere, though, that RBLs are "an outmoded tool, a very blunt instrument, that have outlived their usefulness" and "do vastly more harm than good". (Quoting Fred Langa of Information Week -- see
http://www.informationweek.com/story/IWK20021115S0018.) The article cites research (which I couldn't follow up on) to support its negative view of RBLs. Bottom line: Spammers are getting really good at evading RBLs, so they're gradually becoming less effective. Gosh, maybe I shouldn't try out RBLs after all. Hmm. What to do?
The same article mentions Bayesian email filters, a new tool with great promise. Bayesian filters use statistical analysis to identify spam and they seem to be able to stop nearly 100% of spam with a false positive rate of 0% or so. Those are my kind of numbers!
There are apparently more and more Bayesian email filters popping up on the Internet, mostly experimental, none yet for Notes mail. (For a good place to start reading about this, see
http://www.paulgraham.com/spam.html and other articles and links on the same website.)
Bayesian filters sound pretty good to me. I'd like to try one out on my Notes mail. I searched for "Bayesian" on this forum and in the sandbox. No hits. Darn! Of course I don't have the time (or, okay, the expertise) to write one myself.
Anyone interested in writing a Bayesian spam filter for Notes? If so, I'll be happy, no, thrilled to help you test it. Lotus? Anyone?
Return to top
. . RBL usage: No experience yet, but w... (~Rebecca Cisfre... 18.Jan.03) 
Print this page
