Trustworthy Systems for Today and Tomorrow

Professor Larry Bernstein

Stevens Institute of Technology

Abstract

Software that is reliable, secure and safe is deemed trustworthy. Recent experiments show that software with faults can be run without causing failures. Bounding the execution space of the software enhances reliability. I propose a new approach to software engineering based on reliability analysis rater than focusing only on staff and schedule estimation. Quantitative tradeoffs among tool investment, staffing and design simplification are possible. It is time that we constrain the design of software to make it possible to analyze its reliability.

Bio