Certified By

This thesis develops techniques and methods that extend the art and craft of storytelling, and in particular enable the creation of mobile cinema. Stories are always constrained by the medium in which they are told and the mode by which they are delivered to an audience. This dissertation addresses the design of content, systems, and tools that facilitate the emerging type of computational audio-visual narrative that we call mobile cinema. Storytelling in this medium requires temporally and spatially encoded narrative segments that are delivered over a wireless channel to mobile devices such as PDAs and mobile phones. These devices belong to "the audience," individuals who are navigating physical space and interact with local circumstances in the environment. This thesis examines the underlying requirements for coherent mobile narrative and explores two particular challenges which must be solved in order to make a reliable and scalable stream of content for mobile cinema: technology uncertainty (the fact that what the mobile cinema system presents may not be what the creator intends) and participation uncertainty (the fact that what the audience does may not be what the creator expects). The exploration and analysis of these problems involved prototyping two versions of the MViews system for mobile cinema and three prototype cinematic narratives. Small user studies accompanied each production. The iterative process enabled the author to explore both aspects of uncertainty and to introduce innovations in four key areas to help address these uncertainties: practical location detection, authoring tools designed for mobile channels, responsive story presentation mechanisms, and creative story production strategies. Thesis supervisor: Glorianna Davenport