Interactive Walkthroughs of Large Geometric Datasets A survey of the principles involved in designing interactive walkthroughs of large geometric datasets. The course theme was based on two questions:   1. How can we make effective walkthrough illusions?   2. What are their applications?   The course began with an introduction to interactive walkthroughs and their applications. It discussed taxonomy and representations of geometric datasets. Then it summarized four rendering acceleration techniques: use of image-based representations, model simplification, visibility culling, and dynamic tessellation of spline models. Each speaker surveyed a number of algorithms for each technique and evaluated them with respect to system integration. Next, two speakers presented algorithms for real-time collision detection and managing large geometric datasets for interactive walkthroughs. Finally, the speakers discussed their experiences in developing walkthrough systems for more than a decade and their application of these systems to architectural and CAD datasets.   Organizer Dinesh Manocha University of North Carolina at Chapel Hill   Lecturers Daniel Aliaga Lucent Technologies Bell Laboratories   Frederick P. Brooks Ming C. Lin Dinesh Manocha Andrew Wilson University of North Carolina at Chapel Hill   Jonathan Cohen Subodh Kumar Johns Hopkins University   Hansong Zhang SGI