Most problem domains in physics involve a certain amount of spatial content, but qualitative physics research has been restricted to problems with minimal spatial content. This has greatly restricted the utility of qualitative physics methods, because few ways have been developed of applying successful techniques to a more general range of problems. This weakness has been noted by a number of people working with qualitative reasoning systems [Fal87] [FNF87] [Jos87].
Recent qualitative reasoning literature describes systems which reason about the behaviour of mechanisms, rather than circuits. In these situations, shape and space information is important, and the development of qualitative spatial reasoning methods has been identified as a priority. The techniques described in this thesis are an alternative approach to spatial reasoning that is more general than that currently used for mechanism analysis.