Customer Reviews

Software Error Analysis

5 star:		(1)
4 star:		(0)
3 star:		(0)
2 star:		(0)
1 star:		(0)

 

 

This gem is a straightforward approach to error analysis that is presented in a life cycle structure. It starts with an introduction that discusses the cost benefits of early error detection and provides guidelines for selecting error analysis techniques.

The error life cycle is covered in chapters 3 through 5 and is as follows: techniques for detecting errors, removal of errors and collection and analysis of error data.

Techniques for detecting errors cover detection strategies in each phase of a standard development life cycle. The phases that the authors focus on are requirements, design, implementation, test, installation and operations/maintenance. While this aligns to the older waterfall approach, the detection strategies can be easily refactored to fit spiral or iterative development. As such, this section is not as out of date as it seems - I saw how to easily fit this to the Rational Unified Process. What I found valuable in this chapter was the section on benefits of classes of error detection techniques. This is a solid set of guidelines with which to develop a tailored detection strategy that can fit nicely into a number of areas. Good candidates include: software engineering process group, SQA function, or program management office function that oversees quality from a high level. Moreover, this chapter is augmented by Appendix A, which provides 32 techniques ranging from algorithm analysis to walkthroughs. Appendix B extends this information by cross-referencing these techniques to standards documents that include ISO 9000, IEEE, ANSI and a number of others. These two appendices alone are worth the price of the book.

Error removal covers identification, investigation and resolution approaches. This chapter gives a methodical approach to this function and, like the preceding chapter, can be tailored and incorporated into an organization-wide strategy.

My favorite chapter covers data collection and analysis. What made this chapter valuable is the step-by-step approach the authors give to applying SPC (statistical process controls) to error analysis. This chapter is important to organizations that are striving for CMM Levels 4 or 5, and shows how easy it is to implement SPC and the value such an approach brings.

Both authors work for NIST (National Institute of Standards and Technology). I have read much of Ms. Wallace's other work and have come to regard her as a world-class expert in error analysis and software assurance techniques. This book reflects both author's wide experience and expertise and is characterized by clear writing and ample use of graphics to illustrate the SPC portions of the book. I personally think it should be in the library of every serious SQA practitioner, and should be used in the development of software engineering process improvement and quality strategies. It earns a solid five stars.