Is this really how bayes+autolearn works?

[Scott Silva]

Furnish, Trever G spake the following on 12/13/2006 7:14 AM:
>  
> 
>> -----Original Message-----
>> From: mailscanner-bounces at lists.mailscanner.info 
>> [mailto:mailscanner-bounces at lists.mailscanner.info] On Behalf 
>> Of Denis Beauchemin
>> Sent: Wednesday, December 13, 2006 9:59 AM
>> To: MailScanner discussion
>> Subject: Re: Is this really how bayes+autolearn works?
> 
>> They scored 10 here:
>>
>> 	SpamAssassin (not cached, score=10.448, requis 4.5, 
>> BAYES_50 0.00,
>> 	INFO_TLD 1.27, SARE_LWHUGE 1.00, SARE_LWSHORTT 0.79,
>> 	SARE_MLB_Stock1 1.66, SARE_MLB_Stock2 1.66, 
>> SARE_MLB_Stock6 1.66,
>> 	TVD_STOCK1 2.40)
>>
>> The first one comes from base SA, all but the last one come 
>> from 70_sare_stocks and the last one comes from 80_additional 
>> in /var/lib/spamassassin/3.001007/updates_spamassassin_org 
>> (auto updated by SA).
>>
>> Denis
> 
> Thanks very much.  I added 70_sare_stocks and got all the same hits,
> except of course the TVD_STOCK1 rule.
> 
> I'm not presently using rules_du_jour or sa-update.  Am I correct in
> guessing that your
> /var/lib/spamassassin/3.001007/updates_spamassassin_org directory is a
> sign that you're using sa-update, or is that generated from something
> else?
> 
> I've been avoiding SARE, rules_du_jour, and sa-update on the theory that
> I don't have time to catch misbehaving auto-updated rules before they
> wreak havoc, and that  there aren't as many eyes watching those rules to
> raise an alarm if a rule is bad.  I guess it's time to reconsider that
> theory...
> 
> - Trever
There is a very good setup of rulesdujour with the patches to
bogus_virus_warnings that help with MailScanner false alarms at Fortress
Systems.  http://www.fsl.com/support/Rules_Du_Jour.tar.gz
And sa-update is as reliable as installing a new version of spamassassin, but
the rules get fixed or adjusted faster.

-- 

MailScanner is like deodorant...
You hope everybody uses it, and
you notice quickly if they don't!!!!