Re: Too many false positives...goodbye, razor

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 pos that's much higher in CF than any previous false 
I've had that couldn't be attributed to bugs in hash generation of my 
previous version of razor. It hit at 34. It was a Compuware newsletter, and 
I know for a fact that I'm a registered compuware mailing list user, but 
they put out emails only once a year or so.

Since they mail out so rarely, it's very likely that they are on someone's 
old-email address "spamtrap" and nobody would notice.


I got these hashes on it (I've since issued a revoke for the message, but 
it's still listed at the same cf at the moment)

Feb 11 18:21:40.445270 check[4121]: [ 2]  Razor-Agents v2.22 starting 
razor-check -d
<snip>
Feb 11 18:21:40.833965 check[4121]: [ 3] mail 1.0 e=2 
sig=Q8TKSmJayvWAgL8RMQE7DMDAkecA: Is spam: cf 34 >= min_cf 1
Feb 11 18:21:40.834322 check[4121]: [ 3] mail 1.0 e=4 
sig=n8x1l7XMHlcUQZ--iQRjM8l8-i8A: Is spam: cf 34 >= min_cf 1



-------------------------------------------------------
This SF.NET email is sponsored by:
SourceForge Enterprise Edition + IBM + LinuxWorld = Something 2 See!
http://www.vasoftware.com