Collaborative, task-specific information delivery for agile processes

Knowledge management (KM) approaches for agile software processes need to be lightweight, and must be able to react quickly to new information needs arising for developers. In this paper, we describe how we extended the existing KM tool PRIME by collaborative filtering techniques, in order to pro-actively provide users with task-specific recommendations of potentially relevant information items. The resulting hybrid system thus allows for both the formal specification of "when to offer what information to whom", and heuristic-based information recommendation functionality. As the latter requires no knowledge engineering effort during operation, we argue that the resulting system supports a lightweight KM approach that is viable for agile teams.