bogofilter-0.13.0 available
David Relson
relson at osagesoftware.com
Thu May 22 02:12:41 CEST 2003
At 07:58 PM 5/21/03, Jim Correia wrote:
>On Wednesday, May 21, 2003, at 06:47 PM, David Relson wrote:
>
>>RTFM ! Seriously, that's covered in the RELEASE.NOTES-0.13.
>
>Sorry for jumping the gun. I was posting from work, and hadn't downloaded
>TFM yet since I wasn't planning on building the new package until I was
>sitting in front of my home machine where bogofilter runs. I will read it now.
No problemo.
>>Since I am compulsive about correcting false positives and false
>>negatives, it's a workable strategy. I'm also interested in seeing
>>bogofilter get better as it learns more. If you're patient, my strategy
>>will work for you. If you're in a hurry, then retrain.
>
>I am compulsive about correcting false negatives (haven't had a false
>positive yet), but since I haven't automated this into a cron task it
>requires me to kick off a script after selecting the candidates in my mail
>client, so it happens when I am sitting on the machine with the bogofilter
>dbs (usually once or twice a day, but it may go several days if I am out
>of town - I know that Greg has warned about the short term drift that
>results from using -u in this scenario.)
I'm the laid back member of the team. Greg and Matthias help to keep me
honest. That said, Greg says not to use '-u' and I find it quite
useful. Using fisher tristate, I have my mail client check for
"X-Bogosity: Spam" and "X-Bogosity: Unsure" and sideline them right
away. Periodically I check those folders and take appropriate action so
bogofilter will train on unsures (and false positives and false negatives,
too).
The drift Greg writes about will take a while before it can cause
errors. If you have large wordlists, it will be a long while. For small
wordlists, it'll be a short while. The effect will be more false negatives
and/or positives. As long as you're willing to deal with the effects,
you'll be fine.
David
More information about the Bogofilter
mailing list