Writing with collaborative hypertext: Analysis and modeling

This work explores a novel approach to the analysis and modeling of computer supported collaborative writing. The aim of this dynamic modeling approach is to capture the structure of user behavior as they interact with a computer system. Markov chains models are derived from the empirical data on the use of the system. These models are compared so as to highlight how the dynamic structure of collaborative writing is affected by factors such as the level of users' experience with the system and workspaces for different writing projects. A workspace is associated with a collaborative writing project in terms of both its structure and content. The study shows that the patterns of interactive behaviors of the users were clearly influenced by the system design. Users' experience and workspaces both significantly affected the structure of interactive behaviors in collaborative writing. Users intensively used the system for exploration, organization, and composition tasks in collaborative writing, whereas the use of collaboration facilities of the system was transient in nature. The descriptive‐prescriptive nature of the approach has the advantage of reducing the complexity of computer supported collaborative writing down to a more manageable level at which investigators can concentrate more directly on the essence of the dynamics. The experience from the study suggests that this approach is worth pursuing, and it may be fruitfully generalized by integrating with other process modeling techniques and the mainstream practice of software engineering. © 1997 John Wiley & Sons, Inc.