Requirements analysis from multiple perspectives: experiences with conceptual modeling technology

The paper discusses experiences in several commercial requirements engineering (RE) projects in augmenting an informal, teamwork oriented method with formally based computer support for the analysis and interrelation of heterogeneous perspectives. For several years, the German consulting firm USU had been working with a methodology called PFR (Presence and Future Requirements analysis). PFR is targeted towards the rapid and focused requirements capture in a setting that alternates between team workshops and distributed interviews, in order to avoid the well known problem that requirements change while they are being analyzed. PFR creates a lot of visual, partially overlapping and conflicting requirements perspectives in a very short time. The successful evolution of an RE project depends on the rapid and accurate cross analysis of these viewpoints for possible incompleteness, inconsistency, and conflicts of perception or opinion. In the past few years, meta modeling has been promoted as a step towards solving such problems. To become robust against variations of notations used in different organizations or situations, you can further abstract the different modeling languages to an axiomatic meta meta model which defines the principle concepts of "what information systems are". The meta data management system ConceptBase jointly manages multiple levels of abstraction and handles queries and updates across these levels.