Although visionaries think of breakthrough ideas long before the rest of us, they often think of them before there's technology ready to handle them. Would Multics or the Lisp Machine still have achieved little market success if today's VLSI technology had been ready when they were designed? Forty years ago, there were many things we wanted our computers to do (but few transistors with which to do them), and computer security was a niche concern. Today, computer security affects everybody, and we have transistors looking for useful tasks to perform, but the consequences of design decisions made in resource-constrained days remain. DARPA has launched several clean-slate programs to explore what technological visions might be realized if you didn't have to maintain compatibility with legacy architectures and OSs.