The quality delivered by existing Web applications is often poor. A consequence of this situation is a strong demand for techniques and tools that address the problem of the Web application quality. Lots of approaches are currently available, often coming with prototype or commercial tools implementing them. However, no attempt has been made so far to validate their effectiveness.

In this paper, we consider the available techniques for Web testing and we propose a classification into three major groups. We deal with the problem of defining the Webspecific faults. Our approach is an empirical investigation of the reported faults, abstracted into a fault model. Then, we evaluate the available techniques against the fault model, in terms of the fault categories directly addressed by them. Finally, we sketch a roadmap for the future empirical research.