Training the New Spam Filter

I recently upgraded the framework that drives the developer.* blogs and discussion site, and it looks like the spam filter has become more sensitive as a result. Hopefully this is just a matter of training it (it's a Bayesian filter), though it might be a matter of my tweaking the settings. If a comment or post of yours is marked as spam, don't worry, I'll catch it and approve it. If more than a day goes by without seeing it, please let me know. In the meantime, I apologize for the inconvenience.

While we're on the topic, you may have also noticed that the captcha module was upgraded also. I don't like the new captcha pictures as much as I like the old ones, but it seems to be working OK. I have seen a couple captchas where one or more of the letters is off the edge of the image, and therefore unreadable. If you encounter an impossible-to-solve captcha, copy the contents of your post to the clipboard for safety, submit a wrong answer for the captcha, and the next one that comes back will probably be readable. If you have continual problems, please let me know.

I hate having these inconveniences in place, but without them the zombies would be busting down the door. It's a constant battle.

Thanks,
Dan