A model-driven approach for quality of context in pervasive systems

Handling context is a crucial activity in context-aware systems. In building such systems, the creation of models helps developers to understand and reason on the context information. The quality of context information is required to achieve these systems behavior according to the requirements. Therefore, context models should not only represent the context information but also the quality of context (QoC). MLContext is a textual domain-specific language (DSL) designed to model context, which has been used to automatically generate software artifacts related to the context management for some context-aware frameworks. In this article we present an MLContext extension for modeling QoC. The QoC added features include constructs to express context situations (i.e. constraints on context information), quality requirements and quality levels. The new constructs have been mapped to code for two frameworks supporting QoC (COSMOS and SAMURAI). These mappings have been validated by means of a case study.