Crisis in real-time software

Abstract

The shortfall in the ability to produce credible real-time systems and software has become critical. The number of major civilian and military systems that have suffered expensive delays, failures, and functionality shortcomings is already large, and it is growing. Without a dramatically better scientific basis for designing and constructing real-time systems, massive expense will be incurred to build ill-functioning real-time systems. It is argued that there is an absolute need to invest in basic research on proof techniques, program semantics, scheduling theory, dependability, state-space modeling, and other approaches that offer the promise of being able to create a scientific basis for real-time system specification and construction. Currently known techniques appear to be incapable of dealing effectively with the state-space explosion problem. New basic insights and breakthroughs are essential. >