This paper describes a demonstration of an application called MoodChat. This application is built to showcase the Grapevine context sharing middleware. Grapevine provides a very lightweight but potentially lossy way to efficiently share context information among a group of co-located pervasive computing devices using just device-to-device connections to exchange small bits of context information. Further, Grapevine embodies abstractions that allow flexible definition of contextdefined groups,

more »

... e., clusters of co-located users that are similar in some way based on their context values. MoodChat uses Grapevine to share context about device users' personal states (i.e., mood), including an individual's own assessment of his mood but also values sensed about the device or environment that might also be indicative of the user's mood (e.g., the level of ambient sound, the number and length of sent or received text messages, etc.). Co-located users with the same assessed level of mood are connected into groups, and users within the same group can chat with one another. The demonstration will allow the visitors to interact in MoodChat as one of the users, showing how contextdefined groups can enable a wide range of flexible pervasive computing applications.