Evaluating web applications testability by combining metrics and analogies

This paper introduces an approach to describe a Web application through an Object-Oriented model and to study application testability using a quality model focused on the use of Object-Oriented metrics and software analogies analysis. The proposed approach uses traditional Web and Object-Oriented metrics to describe structural properties of Web applications and to analyze them. These metrics are useful to measure some important software attributes, such as complexity, coupling, size, cohesion, reliability, defects density, and so on. Furthermore, the presented quality model uses these Object-Oriented metrics to describe applications in order to predict some software quality factors (such as test effort, reliability, error proneness, and so on) through an instance-based classification system. The approach uses a classification system to study software analogies and to define a set of information then used as the basis for applications quality factors prediction and evaluation. The presented approaches are applied into the WAAT (Web Applications Analysis and Testing) project.