Proactively Managing the Evolution of Embedded System Requirements

Without distinguishing unstable from stable features and anticipating likely new features, building evolvability into software systems can be time-consuming and, above all, ineffective. This paper introduces a method based on a software evolution model whose goal is to help requirements engineers and product managers identify the unstable features of an embedded system and anticipate its potential adaptation needs in the future, with the aim of planning for changes beforehand. The core characteristic of this method is its support for systematic reasoning on requirements volatility and planning for changes. An application of the method in the domain of assisted living is described as a proof-of-concept, providing both a preliminary validation of the proposed solution and a useful example of its use.