M. Dulipovici and P.N. Robillard (Canada)
Software engineering, software requirements, development process
This paper presents the results of measuring the completeness of software requirements implementation and of comparing multiple software products based on the same requirements specification, by using an original approach settled on a vector-based method. The application of this method is presented in a practical way, by analyzing 6 software products using two different development processes: UPEDU – a scaled-down version of the organized process RUP and XP – a member of the agile methodologies. An experts’ committee measured and analyzed the realization quality of each of 109 requirements of equal importance and judged them based on an evaluation vector system. Two groups of requirements were identified referring dynamic functions and static functions of the system. The observational study was conducted in an academic environment. An important contribution of this paper is the vector based comparison method that could be generalized and applied in any other setting. The findings of this empirical study point to a potential link that may exist between the quality of the implemented requirements and the process used by the development teams.
Important Links:
Go Back
