-----Original Message----- From: Jose Marcio Martins da Cruz [mailto:Jose-Marcio.Martins@ensmp.fr] Sent: Monday, November 15, 2004 3:27 PM To: SM Cc: Chris Santerre; Jeff Chan Subject: Re: [SURBL-Discuss] Re: [SURBL-Announce] SURBL stats: list hit ra tes measured in DNS
SM wrote:
Wrong link...
http://j-chkmail.ensmp.fr/jeff/urlbl/urlbl.txt
Jose-Marcio
Thanks, but.....
These stats are pretty useless: STATS BY URL BL
=================================================
* ab.surbl 139803 38226 * j-chkmail 207945 54310 * multi.surbl 175657 54486 * ob.surbl 990 260 * sc.surbl 220 180 * ws.surbl 354 246
Because you said it goes by alphabet and stops at first hit. Mutli.surbl contains everything in SC and WS. So everything after mutli is of no use.
Your other stats are cool though ;)
--Chris
Chris Santerre wrote:
Thanks, but.....
These stats are pretty useless: STATS BY URL BL
=================================================
- ab.surbl 139803 38226
- j-chkmail 207945 54310
- multi.surbl 175657 54486
- ob.surbl 990 260
- sc.surbl 220 180
- ws.surbl 354 246
Because you said it goes by alphabet and stops at first hit. Mutli.surbl contains everything in SC and WS. So everything after mutli is of no use.
You're wrong.
If some entries were found at WS and SC, this means that they aren't at MULTI.
Why ? There may be a race condition : as multi is created from other bls, it may be possible that when I rsync, SC and WS where already updated but not MULTI.
Quite interesting indeed...
I rsync data every two hours. It could be interesting to change periodicity and see what happens.
Joe
Your other stats are cool though ;)
--Chris
On Monday, November 15, 2004, 12:50:09 PM, Jose Cruz wrote:
If some entries were found at WS and SC, this means that they aren't at MULTI.
Why ? There may be a race condition : as multi is created from other bls, it may be possible that when I rsync, SC and WS where already updated but not MULTI.
WS updates approximately the same time as multi. SC updates more often than multi. In future versions, these will all be internally synchronized.
I rsync data every two hours. It could be interesting to change periodicity and see what happens.
You should rsync a lot more often, at least once an hour.
Jeff C. -- "If it appears in hams, then don't list it."
Hello,
Jeff Chan wrote:
On Monday, November 15, 2004, 12:50:09 PM, Jose Cruz wrote:
If some entries were found at WS and SC, this means that they aren't at MULTI.
Why ? There may be a race condition : as multi is created from other bls, it may be possible that when I rsync, SC and WS where already updated but not MULTI.
WS updates approximately the same time as multi. SC updates more often than multi. In future versions, these will all be internally synchronized.
I rsync data every two hours. It could be interesting to change periodicity and see what happens.
You should rsync a lot more often, at least once an hour.
I'll do it ! And I'll remove other bls already included in multi.
So far, here is a comment I've just received about integrating surbl into j-chkmail.
But Dennis shall note that I did the effort of integrating surbl into j-chkmail, but the great work of surbl comes from Jeff, Chris, Raymond and others.
I agree with Dennis. From all content checking spam detection methods, that one which seems to give less false positive rate is urlbl, no doubt about it.
*************************
Since I use only the left hand side (domains) having that as a separate file would reduce the load I creat even further. The Urlbl list is the most effective method of content filtering I've used so far and I like it a lot. Thanks for building this into the product.
Dennis
**************************
Best,
Jose-Marcio
Jeff C.
"If it appears in hams, then don't list it."
Discuss mailing list Discuss@lists.surbl.org http://lists.surbl.org/mailman/listinfo/discuss
-
Chris Santerre -
Jeff Chan -
Jose Marcio Martins da Cruz