How Specific should Requirements Engineering be in the Context of Decision Information Systems
Résumé
A decision information system (DIS) is a specific component of information system specialized in the support of decision-making by means of online analytical processing tools. Data warehouses (DW), at the core of DIS, integrate and historize collections of data often designed starting from the operational components of the Information Systems (IS). Methods, techniques and tools widely used in the industry to engineer DIS focus, mostly, on designing DW models. However a few approaches were recently proposed to deal with the early phase of DIS development and maintenance, i.e. requirements engineering (RE). As for other specific kinds of systems, the expectation is that better value shall be delivered to DIS users (i.e. different sorts of decision maker such as executives, or managers, as well as control managers or operational) if more effort is spent on analyzing their requirements in early phase of DIS engineering projects. However, DIS is very specific kinds of systems. Therefore the question is how specific should RE approaches be when developing a DIS? To addresses this question, we undertook an in depth analysis of (a) the specificities of DIS systems that shall have an impact on RE approach (b) expectations from DIS project stakeholders, and (c) gap analysis of DIS specific RE approaches and traditional RE approaches. The analysis is structured around the three main axes of RE approaches, i.e. formalization, specification, and consistency. The paper shows that although on a fundamental level RE for DIS approaches does not show conceptual differences with traditional RE for IS approaches, some differences exist in the way of working.