In article <1991Apr3.193547.15558@milton.u.washington.edu>, mitchell@mitre.org ( Richard B. Mitchell) writes: > I have thought > about dividing the world into subworlds and only "polling" those objects > that are in the same subworld as the hand. This seems ok until we > want ot move an object from one point in space to another. I have > also considered using some sort of quad trees, but this does not seem > like the right approach either. If any one has any ideas or pointers > to solutions, I would be grateful. > > ------------------------------------------------------------------------------ - > Richard B. Mitchell mitchell@linus.mitre.or g You should find Hanan Samet's recent pair of books on spatial data structures quite useful. I think these grew out of the article "The Quadtree and Related Data Structures" in Computing Surveys, v16,2 June 84 p187. My favourites are k-d trees, especially netted k-d trees which were first described by J.L. Bentley: "Multidimensional Binary Search Trees used for Associative Searching", CACM 18(9), Sept 75, pp. 509-517. An application to a geographical information system was given in the paper Matsuyama, Hao and Nagao, "A File Organisation for Geographic Information Systems based on Spatial Proximity", Computer Vision, Graphics and Image Processing 26(3), June 84, pp 303-318. Other more regular structures like octrees have been used quite a lot for ray tracing: examples will be found in Andrew Glassner's book, called something like "An Introduction to Ray Tracing". Regards, Matthew Chalmers