Design for Trustworthy Software: Tools, Techniques, and Methodology of Developing Robust Software

Design for Trustworthy Software (DFTS) requires innovative "outside-the-box" thinking, but Genrich Altschuller has shown that creativity can be learnedit is not an inborn characteristic limited to only a few people. His TRIZ theory of innovative problem solving has been developed from a careful analysis of more than two million engineering patents worldwide to discover the fundamental patterns of innovation. We report here on the adaptation of the highly structured and well-tested TRIZ methodology to the design of computer software, which has already shown excellent results for applications to parallel processing. The essence of the TRIZ methodology is to take a problem you cannot solve in your real-world problem space and map it to a similar generic problem in the TRIZ space. Then you follow the very explicit TRIZ methodology to find the generic solution to the generic problem in the TRIZ space. Finally, you map that solution back to a candidate solution for the specific problem in the designer's problem space. This chapter covers the relationship between and interoperability of TRIZ with QFD and Taguchi Methods. It also discusses the application of brainstorming and the Pugh concept selection method to clearly identify the smallest solutions that meet the customer's needs and that are also technically and economically feasible. The role of software patents in protecting the intellectual property inherent in software is discussed, with warnings that the patent law is too weak to protect software.