A Controlled Experiment on the Understandability of Different Requirements Specification Styles

In this paper, we report on a controlled experiment, in which we compared two different requirements specification styles. Following the traditional black-box style, a system is described by its externally visible behavior, any design detail is omitted from the requirements. Following the white-box style, which was popularized by object-oriented analysis, a system is described by the behavior of its constituent entities, e.g., objects. In the experiment, we compared the understandability of two requirements specifications of the same system each written in a different style. The appropriate choice of a specification style depends on several factors including the project characteristics, the nature of the requirements at hand, and the intended readers. In this paper, we focus on the last factor, and investigate understandability from the viewpoint of a customer. The results of the experiment indicate that it is easier to understand black-box requirements specifications from a customer point of view. Questions about particular functions and particular behavior of the specified system were answered by the participants faster and more correct. This result suggests using the black-box specification style when communication with customers is important. 1