on people.
How strong are the safeguards?
Andrew
On Mon, 26 Jul 2004 [EMAIL PROTECTED] wrote:
Date: Mon, 26 Jul 2004 11:52:53 -0700
From: [EMAIL PROTECTED]
To: [EMAIL PROTECTED], [EMAIL PROTECTED]
Subject: RE: [Razor-users] Too many false positives
Andrew,
Can you give me some examples? 5% sounds way too
> -Original Message-
> From: [EMAIL PROTECTED]
> [mailto:[EMAIL PROTECTED] On Behalf Of
> [EMAIL PROTECTED]
>
> We also discovered a bug in the backend implementation of Whiplash
> that affects a small percentage of Whiplash signatures and can
> potentially cause FPs. Again, example FP
, July 26, 2004 12:26 AM
To: [EMAIL PROTECTED]
Subject: Re: [Razor-users] Too many false positives
I've been poking around in the razor source code, and it appears that
the identification of domains has a serious bug.
http://www.greenpeace.org.nz/ is shortened to org.nz
http://www.scoop.co.n
Andrew,
Can you give me some examples? 5% sounds way too high.
We also discovered a bug in the backend implementation of Whiplash
that affects a small percentage of Whiplash signatures and can
potentially cause FPs. Again, example FPs would be helpful to
determine if you are affected by this
On Mon, Jul 26, 2004 at 07:26:21PM +1200, Andrew McNaughton wrote:
> Does anyone know where to find an existing table or lookup system giving
> the level at which various domains are made publically available? Is
> there any sort of mechanism which could avoid having to maintain a list
> like t
On Mon, 2004-07-26 at 08:26, Andrew McNaughton wrote:
> I'm thinking this needs to be handled with a hash lookup on the top level
> domain which returns the levle at which the domain is to be treated.
That won't work, because some TLDs accept registrations on both the
second and third level. And
I've been poking around in the razor source code, and it appears that the
identification of domains has a serious bug.
http://www.greenpeace.org.nz/ is shortened to org.nz
http://www.scoop.co.nz/ is shortened to co.nz
Here's the problem code:
# See if it's a non country domain. If so
On Wed, Feb 12, 2003 at 01:26:05PM -0500, Matt Kettler wrote:
> Nice, those numbers are quite informative Theo. Makes me think that I was
> right to set my min_cf to 11, effectively eliminating the noise present in
> range_01_10, which seems to have a pretty bad S/O.
>
> It also makes me think t
On Tuesday, February 11, 2003, at 10:47 PM, Jot Powers wrote:
inadequate job of compensating for them. I'm seeing far too many
false positives, including Merriam-Webster's "Word of the Day", eBay
outbid notification, and E*TRADE dividend notices...all of which are
important personal messages or
Nice, those numbers are quite informative Theo. Makes me think that I was
right to set my min_cf to 11, effectively eliminating the noise present in
range_01_10, which seems to have a pretty bad S/O.
It also makes me think that upping my min_cf past 11 is also well worth it.
Thanks for posting
> inadequate job of compensating for them. I'm seeing far too many
> false positives, including Merriam-Webster's "Word of the Day", eBay
> outbid notification, and E*TRADE dividend notices...all of which are
> important personal messages or opt-in mailing lists. False positives
The key is, are
On Tue, Feb 11, 2003 at 06:30:46PM -0500, Matt Kettler wrote:
> In general it seems to be rare that it FPs when the cf is very high, but
> there's a LOT of flaky matches in the lower end.
Well, based on ~90k messages (wow, that's not many...) that are from a
recent spamassassin mass-check run:
O
I'll totaly agree that razor does need to do a bit more to comp-out the
FPs.. I'm considering upping my production min_cf again (I'm currently at 11).
In general it seems to be rare that it FPs when the cf is very high, but
there's a LOT of flaky matches in the lower end.
Today I got a false po
13 matches
Mail list logo
