I decided to look into how well a much more starkly limited model space would work for a Bayesian spam filter.
After all, whose spam filter does not groan with ads for suspiciously cheap "Viagra"?
If most legitimate senders adopted such a system( and there is little reason why they would not), spam filters could be made to be much more aggressive in dealing with the remaining mail, thus stopping the vast majority of junk messages.
Ultimately, you want to make sure that your e-mail system and the spam filtering solution have the right information and the right quality of information to enable them to filter the spam effectively.
Using a filtering system on your server or client also means that you can take advantage of simpler rules to dispose of spam that gets through the spam filters.
