出版社：University of Malaya * Faculty of Computer Science and Information Technology
摘要：The increase dependence on software aspects has led the society to emphasize the importance of software quality and metrics. At present there are two categories of quality metrics; dynamic and static. Although dynamic metrics can provide a clearer insight into the software quality issue; it is observed that static metrics are often used for such a purpose. This is due mainly to the technical difficulties associated with the collection of dynamic metrics. One of the known issues when dealing with dynamic metrics is the need for instrument code by inserting points for data collection. This is a very tedious and counterproductive task. AspectOriented Programming (AOP) is a promising technology that is currently used to add crosscutting concerns to the software applications.AOP can be easily used to transparently instrument the code at compiletime. This work proposes AOP as a technique that can be used for collecting software maintainability dynamic metrics data. Therefore, an AOPbased framework for collecting dynamic coupling metric has been designed, implemented; and evaluated. Evaluation results showed that the framework is a reasonable approach for collecting a maintainability dynamic metrics. This approach allowed for the total separation of the metric’s code (the metric aspect) from the application’s source code. Thus, the AOPbased framework provides an effective way for the transparent collection of a maintainability dynamic metrics data. The designed framework can be extended to fit other kinds of quality dynamic metrics.
