Some observations on the hardware development process

To begin our study of hardware development, we were faced with a dilemma. We could study the process or processes and learn something about individual development projects, but since each development was unique, how could we generalize? Furthermore, since we wanted to go beyond being merely descriptive in our analyses and develop normative conclusions, we were faced with the problem of evaluation. Since there are no universal criteria, how can you produce relative evaluations of unique activities? We solved both of these problems by seeking out instances in which the development activity was not unique. We took advantage of such formally sanctioned competitions by studying the National Aeronautics and Space Administration commissioned two or more contractors to develop prototypes of products, that N.A.S.A. would later procure. This enabled us then to obtain performance evaluations from the customer at a subsystem level. This is the level at which one or a very small group of engineers would be working, and thereby enabled us to relate observations of individual engineers to their performance outcome. We discuss some of the results and consider how this work may have some relevance to software development