Towards Efficient Software Component Evaluation: An Examination of Component Selection and Certification

Abstract

When software systems incorporate existing software components, there is a need to evaluate these components. Component evaluation is of two kinds according to literature: component certification is performed by an independent actor to provide a trustworthy assessment of the component¿s properties in general, and component selection is performed by a system development organization. While this principle is in general understood, in practice the certification process is neither established nor well defined. This paper outlines the relationship between the evaluations performed during certification and selection. We start from the current state of practice and research and (a) propose a component-based life cycle for COTS-based development and software product line development, (b) identify a number of differences in process characteristics between the two types of evaluation, and (c) classify concrete quality properties based on their suitability to be evaluated during certification (when there is no system context) and/or during system development.