Calculating Machines Douglas R. Hartree
A review by Frederick A. Ware
This book is volume six in the Charles Babbage Institute Reprint Series for the History of Computing. Douglas Hartree was primarily a physicist, his most significant contribution being the Hartree-Fock self-consistent approximation technique for generating molecular wave functions. Later in his career, he became one of the leading computer researchers in Britain during the late 1940s when Britain briefly overtook the United States in this new field.
This was a tremendously enjoyable book. It goes into significant depth in a number of areas. The reader is impressed and surprised by how far computer hardware and software development had moved in the three years since the unveiling of Eniac.
This is a valuable historical reference, as well, touching upon the primary classes of computing machinery during what is arguably the most critical two decades in the realization of the computer. In 1930 the only computing instruments were the mechanical desk calculator and slide rule. By 1950 there were several dozen stored program machines doing useful work scattered across two continents.
This book covers the development of calculating instruments and machinery during the 1930s and the 1940s. Briefly, the nine chapters are:
 The Differential Analyser (those Brits don't know how to spell) - The Bush Differential Analyser is described extensively. The core of this mechanical analogue computer is the two disk integrator. Shaft rotations are used to represent numerical values, and ordinary differential equations may be solved numerically with the instrument. There is also some discussion of how mechanical analogue computers were quickly replaced by analogue electronic computers with the development of the vacuum tube operational amplifier.
 The Differential Analyser and Partial Differential Equations - Partial differential equations are turned into ordinary differential equations be evaluating them with one of the independent kept to a limited range. The equations are solved repeatedly to give a family of numerical solutions.
Some Other Instruments - Analogue instrumentation capable of Fourier decomposition and Fourier synthesis is described.
 Introduction to Large Automatic Digital Computers - The structure of a digital computer is described - main store, processing unit, control unit, and an input/output unit. Also the basic classes of commands used to program the machine are described: load/store transfers, arithmetic/logical operations, sequencing operations (branch and skip), and input/output operations. Computer science 101, but it was pretty new in 1949.
 Charles Babbage and the Analytical Engine - A description of Babbages computer demonstrating how it mapped into the digital computer of chapter 5.
 The First Stage of Development - A description of the first computers (really just souped-up calculaters with some kind of automatic sequencing capability). This included the Harvard Mark I mechanical digital computer, the Bell Telephone relay digital computer, and the Penn State vacuum tube digital computer. Of the three, the Eniac was the most significant because of its blinding speed - the electronic components gave it a 1000x performance advantage over the other technologies. The key contribution of Mauchley and Eckert was to prove that large numbers of vacuum tubes could be operated reliably. The architecture of the machine was not significant, with all subsequent vacuum tube machines utilizing the familiar stored program architecture.
 Projects and Prospects - The development of large, fast main storage is the critical problem to be addressed in the late 1940s. The two principle alternatives to vacuum tube flipflops are mercury delay lines and electrostatic storage on a CRT . Both are volatile and require refreshing techniques. Both go on to be used in a number of computers in the next five years until ferrite core memory is developed.
 High Speed Automatic Digital Machines and Numerical Analysis - The first attempts to develop numerical algorithms that are stable and efficient. The problems are the same as those attacked by the earlier analogue machines, but the problems of roundoff and quantization must be addressed for the first time.