Beginnings of a revolution - The Computing Universe: A Journey through a Revolution - page 13

composed. The conclusion was that the machine had failed to deliver any
significant benefits or cost savings. A sad postscript to this story is that
both Scheutz and his son ended their lives bankrupt, a condition at least
partly caused by their overenthusiasm for their calculating engines.
There is a positive postscript to this story. Babbage's son, Henry,
inherited most of the unused parts for the Difference Engine that had
been manufactured by Clements. Although many of the parts went for
scrap, Henry saved enough parts to assemble six small demonstration
machines. He sent these to several universities including Cambridge,
University College London, and Manchester in the United Kingdom and
Harvard in the United States. In the late 1930s, Howard Aiken, who with
IBM pioneered the development of the Harvard Mark I, one of the early
electromechanical computers, discovered the small demonstration engine
sent to Harvard. He later said that he “felt that Babbage was addressing
him personally from the past.” 12
It was not until 1991 that Doron Swade and a team at the Science Museum in the United Kingdom
unveiled a working model of a full-scale Difference Engine built following Babbage's designs ( Fig. 1.16 ). This
demonstration showed conclusively that such machines could have been built in Babbage's day, albeit with
a huge amount of engineering effort (see Fig. 1.17 ).

Fig. 1.16 Postage stamp issued to mark
the bicentenary of Babbage's birth.

The Analytical Engine

Babbage was not a master of tact. Instead of finishing his Difference Engine, he unwisely suggested that the
government abandon work on the still incomplete original machine and build a much more powerful and
versatile machine he called the Analytical Engine. The Difference Engine was essentially a special-purpose
calculator, and Babbage had realized that he could design a much more general-purpose machine capable
of performing any arithmetical or logic operation. He conceived the idea for this vastly more powerful and
flexible machine between 1834 and 1836 and kept tinkering and improving his design until the day he died.
This Analytical Engine was never built and was therefore only a thought experiment. Nevertheless, its design

Fig. 1.17 In 1991, to mark the bicentenary of Babbage's birth, Doron Swade and his colleagues at the
London Science Museum unveiled the Difference Engine II, a working model constructed according to
Babbage's original designs. The computing historians and engineers went to great lengths to preserve
the authenticity, using the original drawings, materials, and precision of manufacturing available
in Babbage's time. They considered this work as the continuation of Babbage's project, but almost
150 years later. Difference Engine II contains about eight thousand cogs and weighs about 4.5 tons.
It is operated by a crank handle, as can be seen in the accompanying photo of Swade cranking the
machine. The design included numerous failsafe features, such as mechanical parity checking that
prevented errors occurring even when some cogs get deranged due to vibrations. The cogwheels were
manufactured so that they could fracture in a controlled way; this is an equivalent of a mechanical
fuse. It must be said, however, that the machine was overdimensioned. It can calculate to an accuracy
of forty-four binary digits. This looks excessive, especially if we consider that many of the machines
we use today calculate accurately only to thirty-two binary digits. The machine can also calculate
up to seventh-order polynomials, today we usually use third-order polynomials. It is not clear why
Babbage thought he needed this level of accuracy; a simpler machine would have saved him many
cogs and certainly would have made the construction much easier. Babbage saw the computer as an
integral part of the calculation process, which he envisaged as a factory that produces numbers.