I've been using SpamAssassin for about a week now and the results have been excellent [here]. However, I have noticed a number of false positives. Specifically, I am subscribed to a couple of announcement mailing lists for Salsa dancing that are getting filtered. Of course, this makes sense. The SpamAssassin algorithms really know nothing about me and what messages really are spam, so it sees messages that look like spam and filters them out, not knowing that I really am interested in email advertising Salsa dancing in the Seattle area.

A better approach has been proposed using Bayesian statistical analysis [here]. Both SpamAssassin and the Bayesian analytical approach are content based, but whereas the SpamAssassin approach relies on clever programmers who are guessing at rules that will work, the Bayesian approach relies on statistical analysis of actual emails recieved by a user. Combine this with feedback from the user about spam messages that have made it past the filter and you will have a system that can truly beat spam.

Of course, when I say the system will beat spam, I don't mean that no spam will ever make it through. Some spam will, but the frequency will be so small that it will be more than tolerable.
12:24:11 PM    

A few weeks ago, Australia's HyShot program performed the first test launch of a scramjet engine [here]. Now, they are reporting that the test was a success [here].

"All indications are that supersonic combustion occurred. We’ll now be submitting the results to international peer review," said Allan Paull, of the center.

Scramjets have long been used in science fiction, now fiction is becoming reality.
11:56:40 AM