Previous: A Brief History (Part 2)

Next: The Trickle-Down Theory of Engineering

A Brief History of Engineering (Part 3)

Throughout the first Industrial Revolution, engineering remained largely independent of science. The key inventions of the period, such as Edison's electric light, were made by men without scientific training. The first new technology of the twentieth century -- the aeroplane -- was developed by a pair of bicycle mechanics. The second great technological advance -- the technique of mass production introduced by Henry Ford for the model T -- also owed nothing to science. The major event in science at this time was Einstein's working out of the Special and General theories of relativity. These theories led to no new products at the time, and, as far as I am aware, they remain irrelevant to engineering practice: that is, there is no product on the market that requires a knowledge of relativity theory for its manufacture.

Nevertheless, the twentieth century did see an increase in the use engineers made of science. The science they used was that of the nineteenth century -- applied chemistry for the production of synthetic dyes, plastics and synthetic fabrics. And, in the service of the military, for the production of mustard gas, phosgene, and Zyklon-B. At the same time, the demand for engineers increased. Between 1890 and 1938 the number of engineers in the United States increased from 25,000 to 250,000. The traditional system of apprenticeship was not adequate to meet this demand, so academic programs in engineering started in many universities. The fact that these programs were at universities created a strong pressure to emphasise those aspects of engineering it was easiest to teach, to test, and to write research papers on -- that is, the formal theories of science. This created a large group of engineers familiar with contemporary science.

One critical invention of the early twentieth century was the vacuum-tube diode, though its importance is clear only in hindsight. Because the tube acts as an amplifier, it enables the division between transmission of energy and transmission of information to be realised. A signal so weak that it can be sent around the world for a few cents can be amplified to control a powerful machine. The weak feedback from a machine running slightly out of balance can be amplified till it has the power to put the machine back in balance.

The single event that really tipped the balance towards science-based engineering, however, was the Second World War. I have earlier referred to the attention-getting results of the Manhattan project. But the atomic bomb was only one of the new science-based technologies to come out of the war. Another problem that the Allies faced was guiding anti-aircraft fire; it was necessary to know the target's current position and to estimate its future position, apply corrections to the gun controls, and set the fuse so that it would detonate the projectile at the right instant. The mathematician Norbert Wiener dealt with the general problem of predicting the plane's trajectory based on incomplete information; in so doing, he came up with a mathematical theory of great generality, which forms the basis of modern statistical communications theory.

Meanwhile in England another mathematician, Alan Turing, was developing early versions of the computer to crack the secret code, Enigma, used for German messages. And in Germany, Werner von Braun was designing the V-weapons -- the V-1, which would evolve into the cruise missile, the V-2, which von Braun himself would develop into the rockets that put Armstrong on the Moon, and the V-3, an extraordinarily long gun with rocket-assisted projectiles, whose closest modern descendant may be the `SuperGun' that Canadian engineer Gerald Bull was building for Iraq.

What was common to these problems -- to detonate uranium; to hit a rapidly manoeuvring target; to crack a code -- was their novelty and their urgency. The usual engineering strategies of generalizing from similar problems could not be applied, for there were no similar problems. Nor was there time to evolve a solution by gaining experience by incremental changes to a series of prototypes. Only science could provide results fast enough. These technologies are still the ones that dominate our society, and the scientifically based style of engineering still dominates our idea of what an engineer is.

The twentieth century saw the bloodiest wars in human history. It also saw the greatest progress in engineering. This is not a coincidence: warfare is a uniquely effective stimulus to engineering development. The latter half of the twentieth century consisted largely in the exploitation of inventions made during the second world war: the missile, the computer, the nuclear reactor.

Without the stimulus of warfare, engineering tends to degenerate into sterile stylistic changes -- the aerodynamically useless tailfins on automobiles of the fifties, the rapid succession of bug-ridden operating systems of the nineties. Engineering changes of this type no more represent progress than do the annual changes in Paris fashions. These sterile changes are made for the same reason as the Paris designers change fashion styles -- people will buy something new, just because it's new.

So we must distinguish progress from mere change. As a first approximation, we can say that progress is change that makes a difference. For a tank crew, for example, the difference between conventional tank armour and Chobham armour is the difference between life and death. So warfare provides an objective criterion of progress: the side with more advanced engineering will prevail in combat.

Most of the latter half of the twentieth century was dominated by the `Cold War' between East and West, and the accompanying arms race. Did this produce real engineering progress? On balance, I would say that it did not. Without the touchstone of actual combat, military developments were driven by forces similar to those that drive Paris fashions: the need to produce something that can be sold, rather than the need to produce something that works. A few examples: the development of larger and larger H-bombs during the fifties and sixties, culminating in the 63-megaton bomb exploded by the Soviet Union over Novaya Zemlya in 1963. Now, a 63-megaton bomb is ideally suited to one kind of target: a target encased in concrete a mile thick. There are no such targets. It took several decades to realise this, but in the seventies and eighties, both the US and the USSR scaled down to bombs of a more reasonable size -- about the size of the Hiroshima bomb. Notice also that the numbers of nuclear weapons on both sides grew beyond any rational military need; Defence Secretary McNamara calculated in the sixties that 400 megatons delivered on target would destroy 80% of Soviet industrial capacity. By 1967, the United States had 30,000 warheads aimed at the Soviet Union, with a total megatonnage of about 10,000 MT.

A second example would be the enormously expensive `Star Wars' program. Spending on this started under Reagan in the 80's; it has by now consumed many billions of dollars. Yet, although its goal has been scaled back from defending against a Soviet strike of thousands of warheads to defending against a North Korean strike of a couple of warheads, it has yet to successfully shoot down a single missile. (The recent `successful' test of a missile interceptor relied upon a homing device embedded in the target missile.)

One symptom of the decadence of cold war military engineering is the increase in cost and complexity. In 1980, bombers cost 200 times what they did in WW2; fighters cost 100 times more. A WW-2 Gato-class submarine cost $5,500/ton, while the Trident submarine costs $1.6 million dollars per ton.

The failure of cold war technology to deliver real engineering progress was dramatically illustrated by the difficulties the US experienced in Vietnam and in Iran: when Jimmy Carter sent eight RH-53D helicopters to rescue the hostages held in the Iranian embassy, three out of the eight broke down during the mission. That wasn't bad luck; that was atypically good performance for the RH-53D -- they usually break down more frequently than that. The USSR had similar experiences in Afghanistan.

It is tempting to conclude with a description of the current state of engineering, together with a classification of engineering into civil, electrical and mechanical branches. But this is potentially misleading. This classification is not parallel to the division of science into physics, chemistry, etc. The divisions of science, though not eternal, are relatively long-lasting. But the divisions of engineering reflect a particular time and a particular culture. Moreover, there is a time-lag in these labels, due to institutional inertia. Thus the activities of tomorrow's engineers do not yet have names. (We will review some of these activities in the coming weeks.) Remember also that the engineer solves problems, while the scientist seeks knowledge. `Physics', `chemistry', name areas of knowledge; `civil' and `mechanical' engineering name classes of problem, but we should not limit in advance the places we look for solutions. For example, many of the functions once filled by mechanisms are now filled by electronically controlled actuators; this is the new area of mechatronics. One of the reasons we have the `systems engineering' option is to create people who can move across these traditional boundaries.

Next: The Trickle-Down Theory of Engineering