ABSTRACT

Traditionally, requirements inspection is performed at an early stage of the software development life cycle to reveal defects in a requirements specification. On the other hand, software testing is commonly done at a later stage to look for program faults after coding. Since their purposes are different, requirements inspection and testing are often treated as "separate" and "unrelated" tasks by software practitioners. In recent years, many researchers have proposed to apply testing techniques to requirements inspection at an initial phase of the SDLC. The idea is that generating test cases from a specification may uncover requirements defects well before programming starts. Thus, the possibility of inadvertently developing software based on an incorrect specification can be reduced. The benefits of such proposals are particularly prominent for large-scale projects where the specifications are complex and may easily contain many requirements defects, and the costs of repairing these defects at the late stages of the SDLC are typically hundreds of times greater than if the defects are corrected early. To support this proposal, we describe a method of applying testing techniques to requirements inspection, with a view to improving the quality of the specification before software design commences.

1. The work described in this paper was partially supported by grants from the Research Grants Council of Hong Kong (project nos. 517704 and 116604). 2. Centre for Software Analysis and Testing, Swinburne University of Technology, Hawthorn 3122, Australia. 3. School of Accounting and Finance, The Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong Kong. 4. Department of Computer Science, The University of Hong Kong, Pokfulam, Hong Kong. Email: 5. Department of Computer Science, City University of Hong Kong, Tat Chee Avenue, Hong Kong.

EVERY VISITOR COUNTS: