I've been poking at the problem that I've talked about in the following threads. https://groups.google.com/forum/#!searchin/haskell-cafe/michael$20litchard|sort:date/haskell-cafe/n0Tc29UUgoQ/iitt3z3PCwAJ https://groups.google.com/forum/#!searchin/haskell-cafe/michael$20litchard|sort:date/haskell-cafe/qD2kaZ9qpEA/jTDAp8KoCgAJ And my misguided conclusions here https://groups.google.com/forum/#!topic/haskell-cafe/PMtYhVQ5nNQ I'm trying to write a clone in haskell of the space-system implemented in http://swmud.org. The biggest error in my thinking so far is assuming I could do without spatial extent. Nope, these objects in space will have to have spatial extent. So no octree for me. The advice and comments from the first two threads prompted me to investigate R-trees. I could only find specifics about how to describe 2-D. Until I found this paper on layered R-Trees. http://www.isprs.org/proceedings/XXXIII/congress/part4/1216_XXXIII-part4.pdf This looks like what I want. Here's my re-formulation of the criteria: (1) Objects in space will be either ships that can move or stationary things like non-moving game-controlled ships, space-stations , moons and planets. (2) Objects will have spatial extent. Will use bounding volumes to help determine collisions. (3) Space is non-continuous (4) Movement happens by setting a destination vector and a speed. There's no steering exactly, but you can change destination while moving, slow down or stop suddenly. (5) Represent this space with a layered R-tree, This tree will at most have say, 200-300 objects in it, most of which could be moving around in this space. It's the number of objects that have me wondering if layered r-tree might be too heavy-weight. If so, Any links to R-tree variants that include bounding volumes? I could only find ones with bounding boxes.