Foundational Issues in Software Reuse and Reliability

Abstract. This paper examines the question of whether increased software reuse can improve software reliability. One conclusion is that the assumed positive relationship between them depends on three different hypotheses, which this paper clarifies and makes testable. The intuitive concept of reusable and equivalent one-use components is clarified through the identification of attributes for comparing them, and showing how these attributes affect reliability. By applying measurement theory and using simple programs, the paper shows how various reliability models can be applied to collect data for one-use and reusable programs