Craig Hughes wrote:
> I think the probably most effective blacklist-type use of the AWL
> would be in
> calculating zero-frequency a-priori probabilities for new
> recipients who were
> not previously in the AWL. As I mentioned before, if after say 1 month of
> populating the AWL through normal
Matt Sergeant wrote:
MS> Personally I think the implementation of whitelisting is broken - if
MS> it's whitelisted or blacklisted we should be scanning period. But our
MS> white/blacklisting is implemented separately here, so you're unlikely to
MS> see a fix coming direct from me, I'm afraid (unl
Nathan Neulinger wrote:
NN> Was this changed recently? Cause it most definately did not work for me
I definitely think there's something weird going on in the short-circuit code.
I'll take a look at it and it'll probably be pretty clear what's up.
C
___
Well, AWL can't really run first. It more or less *has* to run last. But
there's no reason it can't run last, after the early-terminate has terminated:
while(early-terminate condition not met)
{
step through some rules
}
check awl here
as opposed to treating AWL as just another rule.
C
Mat
Nathan Neulinger wrote:
> Matt Sergeant wrote:
>
>>On Thu, 2002-05-02 at 19:31, Neulinger, Nathan wrote:
>>
>>>The biggest problem with -S is due to the ordering of the rule checks.
>>>If all of the negative rules (or at least the _large_ negative rules)
>>>were processed first, it would probably
Matt Sergeant wrote:
>
> On Thu, 2002-05-02 at 19:31, Neulinger, Nathan wrote:
> > The biggest problem with -S is due to the ordering of the rule checks.
> > If all of the negative rules (or at least the _large_ negative rules)
> > were processed first, it would probably be ok
>
> All the large
On Thu, 2002-05-02 at 19:31, Neulinger, Nathan wrote:
> The biggest problem with -S is due to the ordering of the rule checks.
> If all of the negative rules (or at least the _large_ negative rules)
> were processed first, it would probably be ok
All the large negative rules *are* processed first
Charlie Watts wrote:
CW> And, hey, if you've got ESP math working even to the point of a test
CW> release, you can quit your day job.
CW>
CW> I haven't actually noticed it to be a useful blacklisting tool, anyway.
CW> I've had it in my head that it could be useful as both, but haven't seen
CW> it
And, hey, if you've got ESP math working even to the point of a test
release, you can quit your day job.
I haven't actually noticed it to be a useful blacklisting tool, anyway.
I've had it in my head that it could be useful as both, but haven't seen
it dragging otherwise-uncaught spam across the
Neulinger, Nathan wrote:
NN> The biggest problem with -S is due to the ordering of the rule checks.
NN> If all of the negative rules (or at least the _large_ negative rules)
NN> were processed first, it would probably be ok, but right now (or at
NN> least with 2.20) - if you enabled it, the white
Sidney Markowitz wrote:
SM> The fact that the -S option is reasonable points out that the scoring is
SM> not a linear measure of spamminess. The function P(s) of the probability
SM> that a message with score s is spam stays near 0 until some small
SM> positive s, then asymptotically approaches 1
I think that AWL will just be whitelist-only if you use -S; you're not going to
be able to get around that I think. If you use -S, you're not going to be able
to guess what the score would have been if you'd let thing keep running and not
short-circuited. I can't think of any posisble adjustment
souri - Rolla Phone: (573) 341-4841
Computing Services Fax: (573) 341-4216
> -Original Message-
> From: Sidney Markowitz [mailto:[EMAIL PROTECTED]]
> Sent: Thursday, May 02, 2002 1:11 PM
> To: [EMAIL PROTECTED]
> Subject: Re: [SAtalk] AWL ve
On Thu, 2002-05-02 at 09:16, Charlie Watts wrote:
> It has just occured to me that this will adjust the AWL math because
> I won't be getting "big" positive numbers into the AWL any more.
The fact that the -S option is reasonable points out that the scoring is
not a linear measure of spamminess.
I've been running with the site-wide AWL and the spamd -S early-terminate
option.
It has just occured to me that this will adjust the AWL math because I
won't be getting "big" positive numbers into the AWL any more.
And I suppose this makes it more of an Auto-WHITE-list than an Auto-WHITE
& BLAC
15 matches
Mail list logo
