Date: Thu, 17 Jul 2008 15:11:18 -0300 (ADT)
From: "Marc G. Fournier" <scrappy@hub.org>
To: Alvaro Herrera <alvherre@commandprompt.com>
cc: Andrew Sullivan <ajs@commandprompt.com>, pgsql-www@postgresql.org
Subject: Re: Spam filtering on the mailing lists
In-Reply-To: <20080717155428.GE3934@alvh.no-ip.org>
Message-ID: <20080717150757.M69138@fserv.hub.org>
References: <200807161509.m6GF9RO14179@momjian.us>
	<1216227673.28961.139.camel@jd-laptop>
	<20080716180738.GF5758@commandprompt.com>
	<8C1E1C3655705FB9706E2E10@ganymede.hub.org>
	<20080717124653.GB6732@commandprompt.com>
	<20080717124111.P69138@fserv.hub.org>
	<20080717155428.GE3934@alvh.no-ip.org>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed


On Thu, 17 Jul 2008, Alvaro Herrera wrote:

> Marc G. Fournier wrote:
>>
>> On Thu, 17 Jul 2008, Andrew Sullivan wrote:
>>
>>>> I can change the quarantining to be >4 if ppl want, which should
>>>> greatly reduce the # of messages going through ...
>>>
>>> Sounds like a good idea to me.
>>
>> 'k, I'll put it down to 4 tonight, and increase the frequency I check for
>> false-positives accordingly, to make sure that I don't get a big spike as
>> a result ...
>
> Hmm, I think keeping it at 5 is a good idea, seeing how 2 of those were
> being substracted by the AWL and you've just cleared that.
>
> Maybe what we need is a way to remove a particular domain from the AWL
> instead?

Actually, I was thinking of disabling the AWL feature altogether ... it 
might make a few extra false-positives to need to release, but I doubt it 
will make a huge difference ...

Maybe disable AWL and leave score threshold at 5 ... ?  I can adjust 
after, based on # of false positives ... if false positives don't go up, 
then without AWL causes no adverse problems ... then I can look at 
adjusting score threshold down while watching for false-positives ...