From: Simon Waters (no email)
Date: Mon May 16 2005 - 05:27:57 EDT
On Monday 16 May 2005 8:14 am, Hement Gopal wrote:
>
> I'd like to expand the list of rbls that I can check against.
Why?
There are a few (too few?) very good pages on comparitive performance of RBLs,
measuring false positives, and other types of reports.
After reading a selection of these, and pondering the goal of getting good
email in, without undue resources being allocated or undue false positives I
plumped for using just "sbl-xbl.spamhaus.org".
Every additional list brings its own problems, but will if it is any good
generally overlap extensively with the other lists where they are accurate.
So RBLs should quickly create a situation of diminishing returns.
If you measure performance by the proportion of sites sending email that are
blocked (rather than the percentage of spam blocked), many RBLs run over a 1%
false positive rate. This might be acceptable for personal use, but for most
business users this is probably too high.
If you want to use multiple RBLs effectively, use them in with a tool like
spam-assassin that can properly weight the significance of appearing on a
RBL, rather than just supply a "yes/no" decision.
|
|
|