An Evolutionary Perspective of Software Engineering Research Through Co-Word Analysis

Abstract : This study applies various tools, techniques, and methods that the Software Engineering Institute is evaluating for analyzing information being produced at a very rapid rate in the discipline--both in practice and in research. The focus here is on mapping the evolution of the research literature as a means to characterize software engineering and distinguish it from other disciplines. Software engineering is a term often used to describe programming-in-the-large activities. Yet, any precise empirical characterization of its conceptual contours and their evolution is lacking. In this study, a large number of publications from 1982-1994 are analyzed to determine themes and trends in software engineering. The method used to analyze the publications was co-word analysis. This methodology identifies associations among publication descriptors (indexing terms) from the Computing Classification System and produces networks of terms that reveal patterns of associations. The results suggest that certain research themes in software engineering remain constant, but with changing thrusts. Other themes mature and then diminish as major research topics, while still others seem transient or immature. Certain themes are emerging as predominate for the most recent time period covered (1991-1994): object-oriented methods and user interfaces are identifiable as central themes.