Selecting a Military Computer Architecture

Abstract

Selecting or even designing a computer architecture is at best as much a black art as a science. The problem is particularly difficult when one tries to choose an architecture meant to serve a very broad range of users whose present and future requirements are poorly understood. This article describes the activities of a joint Army/Navy Selection Committee, which was charged with the task of selecting a single computer architecture to be used as the basis for a new military computer family. The range of applications for military computers is broad and ill defined; how the committee tried to cope with this problem should be of interest to others who are interested in measuring, designing, or selecting computer architectures.