Bayesian Filtering is a great method for fighting spam. Unlike rule based filtering which spammers can easily adapt to with simple modifications, Bayesian adapts with the spammers changes, making it much more difficult for them to defeat the filtering. As a result it’s used in server side mail filtering as well as client side filtering in various products including Mozilla Thunderbird, SpamAssassin, and SpamBayes. Despite this level of “intelligence” it’s not foolproof. Like anything that analyzes unsanitized input, its vulnerable to poisoning. To be fair, there is a debate on if it exists or not. I personally believe it does exist.
Thunderbird 2.0b1 is out, I updated a few days ago. I really love the new tagging functionality. Being able to create your own tags makes organizing mail about 100X easier. The presets of 1.5 just weren’t enough. As far as the UI goes, I was initially not to fond of the earth tone coloring, but I think the new icons are starting to grow on me. There is also a new phishing detection (similar to Firefox). To test it, I looked in my spam folder for a few phishing emails to test the new filter against. So far so good.
The only downsides thus far is bayes spam filtering is not performing as good as it did on 1.5. I reset things, hopefully after a few days of learning it will resolve itself. Or perhaps it’s a lingering regression in 2.0. It is after all still in beta. The other is the new mail notification doesn’t seem to open mail if you click on it. I was hoping it would open email when clicked. Perhaps it’s just not obvious where to click. The appearance and effect seems to be much better now.
It’s hard to write even a mini-review of beta software, since it is just beta and things are incomplete or subject to change. I plan to write more on it closer to the 2.0 release. Despite it’s lower profile development (compared to Firefox), and more subtle changes) it’s really evolving. The changes made really do make it a much better experience.