An Empirical Investigation of Models Versus Prototypes in Systems Development

摘要

It is generally accepted that models are useful in systems development and a variety of modeling tools and techniques are taught in information technology academic programs and used by software professionals. However, the recent emergence of agile systems development approaches, including extreme programming, SCRUM, and pair programming is challenging the value of modeling. Surprisingly, other than anecdotal evidence, there is little empirical data to support or refute the conventional wisdom that models are useful in systems development. This paper reports the results of a pilot study that examined the preference for diagrammatic models and a working prototype among student subjects, who participated as potential end users of a stock trading system. Surprisingly, the results show that non-technical users found the models to be more useful than the prototype. Implications of the study are discussed, and planned modifications for a follow-up experiment that addresses the limitations of the pilot study are presented.