Context in knowledge sharing has not been sufficiently investigated. It is mentioned only passively or left to implicit interpretation by the actors engaged in the knowledge sharing exercise. Knowledge sharing is predicated upon the task with which one is involved and the constituents of a knowledge sharing environment--termed features in this work--that influence that task. This knowledge sharing environment consists of all factors of the situational milieu that influence the observed context of any entity. In this way, context and task are irrefutably intertwined. The elements of the environment bear temporally dynamic weights dependant on the task being conducted. In this research, we are investigating the role and determination of context in knowledge sharing, in general, and its application in peer-to-peer (P2P) network-mediated knowledge sharing, in particular.
We propose a novel representation of context for knowledge sharing activities and define a framework for context-aware knowledge sharing. Our novel context representation takes the form of a Task-Feature Relevance Matrix. The features consist of all concepts and attributes of the peer's domain as codified in a ontological peer ontology, as well as task descriptor features that distinguish the unique or variable attributes of a knowledge task. Cross-referenced with the potential tasks of an activity, we obtain a relevance factor for that task/feature pair which is then used to determine the affinity between peers of a P2P network in order to mediate knowledge sharing. We show how this framework can be seamlessly integrated into current ontology-based knowledge sharing solutions.