Software Design for Six Sigma: A Roadmap for Excellence

Chapter 18. ROBUST DESIGN FOR SOFTWARE DEVELOPMENT

INTRODUCTION

In the context of this book, the terms "quality" and "robustness" can be used interchangeably. Robustness is an important dimension of software quality, and it is a hallmark of the software Design for Six Sigma (DFSS) process. The subject is not familiar in mainstream software professionals, despite the ample opportunity for application. This chapter will explore the application of the Taguchi robustness techniques in software DFSS, introducing concepts, developing basic knowledge, and formulating for application.^[219]

In general, robustness is defined as a design attribute that represents the reduction of the variation of the functional requirements (FRs) or design parameters (DPs) of a software and having them on target as defined by the customer (Taguchi, 1986), (Taguchi & Wu, 1986), (Phadke, 1989), (Taguchi et al., 1989), and (Taguchi et al., 1999).

Variability reduction has been the subject of robust design (Taguchi, 1986) through methods such as parameter design and tolerance design. The principal idea of robust design is that statistical testing of a product or a process should be carried out at the developmental stage, also referred to as the "offline stage." To make the software robust against the effects of variation sources in the development, production, and use environments the software entity is viewed from the point of view of quality and cost (Taguchi, 1986), (Taguchi & Wu, 1986), (Taguchi et al., 1989), ...