The degree of complexity and randomness of aircraft tracking which can be safely managed is a fundamental question for a given airspace design. To answer this question and to provide optimum conflict resolution strategies, the nature of conflict needs to be thoroughly understood. This paper reviews traditional conflict models showing how these models may be integrated into a more unified model based on computational geometry. This mathematical approach is being investigated for the specification and management of future airspace, achieving a more solid architectural basis for airspace design. This approach has the potential to significantly reduce the subjectivity when attempting to model controller and pilot workload, an important requirement in future designs.