Tackling visual and conceptual complexity of problem-oriented modeling of requirements

Abstract

In the contemporary age of cyber-physical systems (CPS), where software governs the control, coordination, and management of physical objects to tackle real-world problems, engineering requirements for such systems present considerable challenges. Consequently, it is not surprising that the requirements engineering (RE) community, encompassing both academia and industry, has turned to established RE methodologies and applied them within the context of CPS; notably, Jackson’s Problem Frames (PF) approach deploys problem diagrams for modeling CPS. However, problem diagrams of realistic CPS often present visual and conceptual complexities, which must be addressed before PF can be usefully applied in practice. In this paper, the above problem is addressed in two steps: firstly, preliminary findings are presented by deriving from the application of eye-tracking software in the assessment of a technique designed to mitigate the visual complexity inherent in the CARE (Computer-Aided Requirements Engineering) tool developed for PF. An auto-layout technique is developed for detecting and resolving overlaps, aimed at enhancing its usability from the perspectives of cognition, psychology, and user studies; secondly, the technique of a complexity matrix is applied for calculating and evaluating the conceptual complexity inherent in problem diagrams. Our eye-tracking results show the successful implementation of overlap detection and resolution. The case studies in the second step also demonstrate the applicability and effectiveness of the complexity matrix calculations.