[EMAIL PROTECTED] wrote:
On Wed, 24 Sep 2008, [EMAIL PROTECTED] wrote:
I was actually hoping to use it like I use zen.spamhaus.org and
dul.sorbs.net and just reject emails listed on those. It is very rare
that I get a false positive from either, but their efficacy isn't what
it used to be, either. So, I just configured my tcpserver to invoke
rblsmtpd using b.barracudacentral.org as well as the other two, and
after only a few seconds, the difference was astounding. Here is
perhaps 2 minutes worth of stats:
$ grep -c sorbs bl_stats
9
$ grep -c spamh bl_stats
228
$ grep -c barracud bl_stats
1321
Replying to myself, after I sent this, it occurred to me that the query
order is a huge factor...rblsmtpd stops scanning after the first hit.
Here is what I got when I put zen in front of barracuda and ran it for
maybe 30 seconds:
$ grep -c barracud bl_stats2
22
$ grep -c spamh bl_stats2
355
$ grep -c sorbs bl_stats2
3
In other words, zen is probably actually more effective by itself than
barracudacentral. Nonetheless, it helps a lot.
I see aproximately the same numbers, with a little more hits for zen (I
use a warn_if_reject for the BRBL). In percents (B & !Z)/(B+Z) ~= 10%,
and (Z & !B)/(B+Z) ~= 13%).
anyway,
- zen is widely used. so even if it has an FP, the originator will have
problems sending to a lot of places, and has enough incentives to get
delisted. In other words, the FPs caused by zen are "passed to the
originator" and are no more "our FPs"! (I hope you see what I mean).
- we don't have enough infos (yet?) about BRBL.
