A practical approach to workstation cost-justification

Over the last couple of years several tools have emerged which help intending users of technical computing systems to estimate cost-effectiveness. Most of these are variants or derivatives of the popular BANG approach, but recent research on workstation performance has thrown useful light on the problem. Workstations are too new a phenomenon for the theory to be rigorously defensible, but the indications are strong enough for useful implications to be drawn. They cover not only the raw processing power of the workstation, but also the effects of display interfacing, operating system performance, and networking method. Some preliminary conclusions are drawn. Some of the background theory is only implied, as this paper is aimed at the likely user of such systems rather than computer or social scientists.