Pervasive smart devices link embedded concerns to information technology in a single, resource-constrained system. Both domains have quite opposite computation characteristics - reactive versus transformational. Due to C's dominance in industry, the reactive part is usually based on conventional sequential programming which lacks domain-specific support making the solution hard to program, comprehend and maintain. Synchronous languages might be a promising solution in order to facilitate software engineering and improve software quality. However, to date, they are restricted to very specific industrial niches; real-world deployments and evaluations are rarely reported in literature.

This thesis conducts a case study that examines the feasibility and suitability of the synchronous approach based on a real-life smart device. Focusing on its reactive concerns, it elaborates the engineering challenges and quality issues of the existing production code. By taking advantage of the synchronous language Céu, it outlines a synchronous reimplementation, thereby illustrating the deployment of synchronous programming and how to reconcile it with the transformational part of the system. Architectural considerations and best practices are provided for developers in order to effectively apply the synchronous language concepts. Furthermore, it shows the applicability of established object-oriented software design patterns and how to implement reproducible unit tests for reactive code.

Several qualitative discussions treat the software engineering and quality benefits gained by the synchronous reimplementation compared to the existing production code. A code analysis uses the separation of concerns, the scattering of interfaces and the code size as performance indicators in order to quantitatively substantiate the results. A user study confirms that reactive behavior is easier to implement and comprehend using the synchronous approach.

This thesis represents a proof of concept which demonstrates the feasibility and suitability of synchronous programming in resource-constrained, real-life industrial embedded applications that are exposed to reactive and transformational concerns likewise. By using synchronous programming, we were able to recover fundamental software engineering principles while, at the same time, fulfill the strong resource limitations - a combination that is known to be hard to achieve. Finally, we believe that our work generally suggests a practicable way of improving embedded software quality in industrial applications.