An Expert Module Design and Implementation that Evaluation the C&K Metric Suite in Terms of Object Oriented Software Quality
This work addresses the need for better interpretation of object-oriented metrics in software testing tools, but it is incremental as it builds on existing tools and metrics.
The paper tackled the limitation of existing software testing tools that only produce metric findings without interpreting their impact on software quality by developing an expert module that automatically reviews and interprets C&K metric suite results for object-oriented projects. The result is a module that supports white-box testing through automatic, manual, and cross-version comparison methods.
Object-oriented software has some features such as encapsulation, inheritance, cohesion, coupling and polymorphism. These features significantly affect the testing, maintenance and repair activities of object-oriented softwares. The successfully conclusion of these activities, depends on correct detect and correct interpretation of metrics will be evaluated. Also safely and rapidly evaluation of this metrics, is possible but by developing and using the software testing tools. Nowadays, many software testing tools that there are for object-oriented metrics. However, there are some limitations on the capabilities of this tools. Many of the this tools produces only metric findings, does not take any action in the point what this findings mean and how it affects the quality of the software. Therefore, diversifying and increasing capabilities of these tools is required. Based on the limitations of software testing tools mentioned in this study, is intended to provide the automation by expert system approach to object-oriented testing process. For this purpose, an expert module that review and interpret the metric findings, that providing information about the internal quality attributes of a project (.jar extension) developed according to object-oriented programming methodology have been developed. Metric suite that developed by Chidamber and Kemerer (C&K) have been used in this modul. This module, metric findings obtained by software testing tool named Metrics, has the evaluation feature by methods automatic, manuel and cross-version comparison and the interpreting the relationship between internal quality specifications. It has been expected that the betreffend module to be a support element to white-box tests been valid in software testing activities.