Abstract
Strategic Computing, a 10-year initiative to build faster and more intelligent systems, is ambitious, flawed by overscheduling perhaps and problems of definition, but basically sound.