On Fri, 2003-08-29 at 21:18, David Clack wrote:
Everything is stable, I do find the spamassassin to be slow though.
Agreed, but then it's in Perl... I use spamprobe instead, and while I still see a pause between the sound for new mail and it arriving in my Inbox, it seems to perform faster. It's in C++, and I've no idea if it's letting through more spam or less, as I've not done that kind of testing. But I've got a spam filter (BrightMail) up line at work already, and then put spamprobe in place to filter what that misses. And then send those caught back up to BrightMail's system to have their rules changed to catch those, too. So what gets through is "new" mostly anyway... I say "mostly" as I'm not sure what difference me sending my spams back really makes - it seems to cause a reduction in my spam, so I *think* it's worth it. And I see a 90-95% success rate on catching spam, with only 2 false positives so far to date (both from companies trying to sell me stuff, but I'd subscribed to them and not white-listed them, as I can't see how with spamprobe - it's a pure and simple Bayesian filter.) Ta, -- o o mailto:Moredhel earthling net /v\ark R. Bowyer http://www.bowyer.screaming.net Mark Bowyer Sun COM `-' --------------------------------------- /"\ ...fingerprint = 7924 9E9E 7B91 225E B065 \ / ASCII Ribbon Campaign 1024D/15140DC1 39D0 551D ABE6 1514 0DC1 X Against HTML Mail -------------------------------------------- / \ If you're a blacksmith, probably the proudest day of your life is when you get your first anvil. How innocent you are, little blacksmith.
Attachment:
signature.asc
Description: This is a digitally signed message part