Mason: An Open Development Contextual Sensing Framework Enabling Reactive Applications

Abstract

Mobile devices continue to push the limits of contextually aware application intelligence. However, due to the complexity of contextual processing and programming, a centralized system that handles all mobile context processing is difficult to realize. The problem of defining such a contextual reasoning unit that uses an all-encompassing contextual ontology for all possible uses of context is not feasible nor useful. Furthermore, implementing custom contextual logic ad hoc per application is difficult due to the complexity of sensor monitoring and contextual reasoning and may be redundant across applications. In this work we propose an openly developed dynamic ontology formation that allows developers to contribute logical pieces to a greater network of contextual reasoning for use by application developers. Specifically, we introduce Mason, a framework for supporting modular contextual reasoning development by handling low-level sensor routing and abstracting data sources as composable and functionally reactive data streams. This provisions for high levels of abstraction for contextual logic developers that contribute to the framework as well as application developers that use it. We demonstrate the simplicity of developing with Mason and show, through an audit of open source applications, the increased contextual functionality offered, better enabling the next generation of contextually reactive applications.