Experimental Evaluation of the Cleanroom Software Development Method

Abstract

The field of software engineering is evolving as various new techniques, languages, paradigms, process models, methodologies, metrics etc., are constantly being developed, used and improved. In the past, several empirical studies were conducted to experiment with various new ideas in software engineering in order to help us better understand, evaluate, predict, control and improve these ideas. Basiii et al. review several of these studies and present a general framework for analyzing these experiments[2.5]. They emphasize the need of such experiments for advancement of software engineering discipline. They recommend that the experimental planning process should include a series of experiments for exploration, verification, and application of new ideas. They also implied that the results of empirical experiments in software engineering should be verified by a series of other related experiments. Cleanroom methodology for software development is a relatively new concept. Currently researchers are attempting to study the benefits and implications of using the Cleanroom methodology. One such attempt was made at the University of Maryland (UM) a few years ago which empirically characterized various aspects of Cleanroom methodologyj211. The fact that this study was empirical motivates the need for verifying its results. Also, the fact that the methodology is new motivates the need for further experimentation. Therefore, this experiment has been replicated recently at Miami University (MU). The goal of this project is to analyze the data collected in this experiment to verify the results of the previous study.