The Utility Space Of Requirements Engineering

This paper exploits the micro-economic concept of expected utility to characterise the costs and beneets of diierent levels of expenditure on requirements engineering. It is argued that diierent groups associate diierent levels of expected utility with particular development techniques. For instance, the advocates of formal methods argue that the high costs of mathematical speciications ooer substantial beneets during the later stages of development. For other groups within the requirements engineering community, the high costs of formal speciication techniques are perceived to ooer few rewards. This paper argues that utility curves can be exploited to characterise these diierent viewpoints. The intention is not to map out all of the complex factors that aaect human decision making but to focus on the interaction between issues of cost and probability in commercial project management. The resulting models are then used to identify an agenda for progress in the eld of requirements engineering. Quality assessment procedures must be developed to demonstrate the beneets of particular approaches. Cost eeective methods must be developed to encourage the application of advanced techniques. Finally, it is argued that the application of high-cost approaches must be targeted so that they produce the greatest expected utility.