Fitzgerald article

0 7 4 0 7 4 5 9 / 9 9 / $ 1 0 . 0 0 © 1 9 9 9 I E E E M a y / J u n e 1 9 9 9 I E E E S o f t w a r e 3 7 iven the extent to which software underpins all our everyday activities, software development has become a critical issue for modern organizations, and indeed for all of society. The “software crisis,” a term first coined more than 30 years ago, continues unabated, however— software still takes too long to develop, costs too much, and does not work very well when eventually delivered. Some argue that this is because the software process in many organizations is undisciplined, chaotic, and completely unpredictable.1 In other organizational areas, increased process maturity has led to concomitant improvements in efficiency. Therefore, increasing software process maturity is an obvious and logical step in addressing the software crisis. The most comprehensive and effectively realized model of process maturity is the Capability Maturity Model from the Software Engineering Institute.1 While the CMM was initially applied to government and military software development, its use is now spreading to all industry sectors. However, a large-scale comprehensive study of organizations that have undergone CMM evaluation reveals that only two percent of organizations have achieved level 4 or 5; nearly 62 percent are at level 1.2 Also, the process of achieving level 3 takes four years on average,2 which shows that progression through CMM maturity levels is both time-consuming and difficult. Brian Fitzgerald, University College Cork, Ireland Tom O’Kane, Motorola A Longitudinal Study of Software Process Improvement Study ing a so f t ware pro cess improvement e f for t over t ime revea ls the fac tors asso c ia ted wi th i t s success. Th is case s tudy show s how Motoro la’s Ce l lu la r In f ras t ruc ture G roup progressed to CMM leve l 4 , and examines what i s needed to opt imize i t s so f t ware deve lopment pro cess.