Integrating and Optimizing Object-Oriented Software Metric Tools

Abstract

Software Metrics provide critical information about the reliability and maintainability of the system and hence are used to improve the productivity and quality of software. As the mathematical nature of metrics calls for precise definitions of the same metric depending on the implementation language, it is suggested to express and define metrics using a language-independent Meta-model based on graphs. Several software metric tools are available, using different methods to assess metric-based software systems and hence project different results. The results are thus tool dependent and are in question for validation. Here an attempt is made to integrate five different object-oriented commercial and free metric tools, and their results have been optimized. An Empirical study has been done to calculate the metrics values using the same standard metrics for two software projects of different sizes. The results have been presented and discussed here to show the variations in results from different tools for the same metrics. To support the validation, a manual investigation has been done on a small java file to check the validity of tools for optimization of the different values obtained for the same metrics from different tools.