Structured and unobtrusive observation of anonymous users and their context for requirements elicitation

Today, people find themselves surrounded by IT systems in their everyday life. Often they are not even aware that they are interacting with an IT system. More and more of these systems are context adaptive. Requirements to such systems may change for various reasons: The context may fundamentally change when other systems are introduced. New trends and fashions may evolve. Operators need to react quickly to such changes if they want to keep their systems competitive. Traditional approaches to requirements elicitation start to fail in this situation: context adaptive systems serve many users with different profiles. In addition, users may be reluctant to participate in improving it. Thus, it is hard to establish a representative model of requirements. Furthermore, it is hard to capture the context of requirements by subsequent interviews. In this paper we present a systematical approach for requirements elicitation based on observing anonymous users. The interaction of users with the system is observed in the normal working context. Observation is based on assumptions on how interaction should take place. Deviations from these assumptions point to new requirements. Observing a large number of users leads to a quantitative map of requirements in context. Preliminary evaluation shows that the approach is promising. It allows efficient observation of many stakeholders and the derivation of new requirements.