Traces Using Aspect Oriented Programming and Interactive Agent-Based Architecture for Early Usability Evaluation: Basic Principles and Comparison

Early evaluation of interactive systems is currently the subject of numerous researches. Some of them aim at explicitly coupling design and evaluation by various software mechanisms. We describe in this paper two approaches of early evaluation exploiting new technologies and paradigms. The first approach is based on aspect oriented programming; the second one proposes an explicit coupling between agent­oriented architecture and evaluation agents. These two approaches are globally compared in this paper. 1 Introduction Interactive systems evaluation is a very rich research and application domain since the seventies. To evaluate how people use the interactive applications, many techniques and methods can be applied [18,19,22]. Some of them are widely used in companies and universities. New methods or variants of methods appear and are tested progressively, according with new needs and specificities coming from the emergence of information and communication sciences and technologies. Among them, automatic and semi­automatic methods and tools are considered as promising (see [16,17]). This paper concerns early usability evaluation, which is also currently particularly studied in the HCI literature. We present in this paper two complementary approaches contributing to early evaluation during the first stages of a project. It is a question of explicitly coupling design and evaluation by exploiting innovative technologies and paradigms. This coupling aims to produce traces concerning how is used the interactive applications to evaluate their utility and utilisability, for example by comparing prescribed and effective tasks. The first approach exploits the paradigm of aspect oriented programming to integrate mechanisms of trace in interactive applications. The