Improving Software Systems By Flow Control Analysis

Abstract

Using agile methods during the implementation of the system that meets mission critical requirements can be a real challenge. The change in the system built of dozens or even hundreds of specialized devices with embedded software requires the cooperation of a large group of engineers. This article presents a solution that supports parallel work of groups of system analysts and software developers. Deployment of formal rules to the requirements written in natural language enables using formal analysis of artifacts being a bridge between software and system requirements. Formalism and textual form of requirements allowed the automatic generation of message flow graph for the (sub) system, called the “big-picture-model”. Flow diagram analysis helped to avoid a large number of defects whose repair cost in extreme cases could undermine the legitimacy of agile methods in projects of this scale. Retrospectively, a reduction of technical debt was observed. Continuous analysis of the “big picture model” improves the control of the quality parameters of the software architecture. The article also tries to explain why the commercial platform based on UML modeling language may not be sufficient in projects of this complexity.