For one thing, to deliver on these promises Palladium implementations will have to be bug-free. "Bug-free" and Microsoft are not often associated concepts. But a bigger worry is that Palladium could be too good at what it's designed to do. Microsoft could use Palladium to control what software could run on your computer, a government could use it to control what you could see on the Internet, or a PC vendor could make it impossible for you to sell your used computer. People could send e-mail that only the intended recipient could read, a neat feature if you are Bill Gates about to be hauled into court, but not so good if you want to prove that someone is harassing you. Not exactly Independence Day concepts.

There needs to be better security for our computers - but the bulk of the security problems are caused by Microsoft applications, not by the operating system. Fix the flaws in the apps, Redmond, then try to develop security conscious apps. Security is important, but not important enough to give up all of my privacy.

Ladies and gentlemen of the jury, let me offer you Exhibit A - Microsoft Passport. Intended to be a single sign-on system that would manage and selectively express your identity online to Web sites and services, Passport has been found to have security holes big enough to drive a bus through, and its "benefits" are largely based on making it easier to buy stuff. Not exactly the profile of a strategic IT solution.

My guess is that unless Microsoft manages to pull some kind of systems engineering rabbit out of its virtual hat, Palladium will be buggy and hard to administer, and do more to secure the Microsoft stranglehold over OEMs and users than solve the problems of secure, reliable computing.