ComPARE: A Generic Qua for Component-Bas

Component-based technology is gaining popularity in modern software development. This approach helps reduce development cost and time-to-market, as well as improve maintainability and reliability. One of the key problems in component-based software development is finding a way to certify the quality of individual components and that of the integrated component-based software systems. There are several different techniques which have been developed to describe the predictive relationship between software metrics and the reliability of the software components. In this paper, we propose a generic quality assessment environment for software components: ComPARE. ComPARE collects various metrics from candidate components including process metrics, static code metrics and dynamic metrics. Also it integrates different models to predict software quality and reliability, and compares the result of different models. With ComPARE, user can select and define their own prediction models and validate these models against the failure data collected in real life. The benchmark models can be established after validation for future use. Finally, prediction results can be visualized and hidden problems can be identified in the source code in the ComPARE environment. Keyword: Quality assessment tool, componentbased software, classification tree model, case-based reasoning, Bayesian Belief Network.