1171 Pair Modeling Pankaj Kamthan Concordia University, Canada P IntroductIon We model software for a variety of reasons: to assess the viability of or plan software systems to be built, to optimize use of (minimize, or ideally, avoid waste of) resources in response to inevitable changes in business, social, or technological environments, or simply to un- derstand existing software systems. Indeed, as indicated by the model-driven approach to software development (Beydeda, Book, & Gruhn, 2005), models are becom- ing first-class members of organizations and software process environments that embrace them. The examples of collaboration in implementation of software are well-known (Nosek, 1998; Williams & Kessler, 2003). As models get large and complex, the need for creating them collaboratively in a systematic manner arises, and we propose pair modeling as an and its application to education. Next, challenges and directions for future research are outlined. Finally, concluding remarks are given. Background Over the last decade, there has been an increasing reli- ance of activities and deliverables in the later phases in the software life cycle on model artifacts that are created earlier. In particular, modeling has begun to play an increasingly central role in adaptive software process environments such as extreme programming (XP) (Beck & Andres, 2005) and the unified process (UP) (Jacobson, Booch, & Rumbaugh, 1999). The in- creasing dependency on models raises the question of how such models can cooperatively be created within