Martin Hutchinson’s blog (often on matters of economics) is reliably an erudite gem – and he outdoes himself in today’s column. His thesis is captured in the first paragraph:
“In 1830, Britain was completely economically dominant, the only significant industrial power. Since that time, it has steadily lost its industrial dominance and that relative decline, while to an extent inevitable, was hurried along by a grossly malign approach to policy which persists today. Unless it is reversed, further accelerating decline is ahead.”
Unfortunately, the signs of that “English Disease” are clearly visible today far from Albion and closer to home. The underlying problem is that there is always a handful of insulated individuals who think they know best. When the political situation gives such individuals free reign, negative consequences are inevitable.
Particularly interesting is Hutchinson’s example of Colossus – the British-built world’s first electronic computer which played an important role in decoding German communications in World War II, despite having been rejected by the code-breaking geniuses at Bletchley Park. If a different path had been followed, England could have led the post-war electronics revolution, and perhaps have avoided its economic decline. It is definitely worth reading the whole article.
Here is money paragraph about Britain and computing.
Manchester University worked on developing a computer, complete with Turing, but Turing was no engineer, spending his time instead on “Big Brain” problems such as the nature of Artificial Intelligence, so progress was slow and the first machine did not appear until 1950, well after the Americans. The British computer industry was late and feeble.
Oh dear. During World War II there were two major electronic computer projects: Colossus in Britain (for codebreaking) and ENIAC, developed primarily to produce artillery firing tables for the U.S. Army and not completed until after the end of the war, in December 1945. Both of these machines used vacuum tubes as logic elements, but both were special purpose machines, and neither had a stored program architecture, central to almost all general purpose computers.
After the war, the British Colossus group from Bletchley largely migrated to the University of Manchester and the ENIAC designers to the University of Pennsylvania where each independently pursued computer development, largely as an academic effort with a modest budget. There was no large government or commercial effort to develop computers in either country. In 1948 the Manchester group tested the Manchester Baby, the first stored program computer, and in 1949 completed the Manchester Mark 1, the first general purpose stored program computer, which went into service providing computing for the university. Also in 1949, Cambridge University began operation of EDSAC, the world’s second stored program computer, based upon the design by John von Neumann, which provided computing resources for Cambridge.
In the U.S., the ENIAC group at the University of Pennsylvania designed BINAC, a binary stored program computer for Northrop Aircraft, but the principal designers, Eckert and Mauchly, quit to start their own company and Northrop said the computer never worked properly after having been delivered. It was not until 1950 that the first operational stored program went into operation in the U.S., SEAC at the U.S. National Bureau of Standards, who designed and built it in house.
In 1951, British electronics firm Ferranti Ltd. introduced the Ferranti Mark 1, a commercial version of the Manchester Mark 1. This was the first commercially available computer. A total of 9 were sold.
Also in 1951, the Eckert-Mauchly Computer Corporation (ex-ENIAC designers) in Pennsylvania introduced UNIVAC I, a computer oriented to commercial applications, with the first delivery of a system to the U.S. Census Bureau that year. This was the first “mass produced” computer, with a total of 46 eventually sold.
The first business to embrace automation was J. Lyons and Co. of Britain, which introduced its first LEO (Lyons Electronic Office) computer in 1951. The first machine was based upon the Cambridge EDSAC. LEO computers remained in use until 1981. See Georgina Ferry’s 2003 book, A Computer Called LEO for the story of successful wide-scale automation in 1950s Britain.
Thus, Britain and the U.S. were essentially peers in both the technological and commercial development of computing from the end of World War II through the early 1950s. It was only after the entry of IBM into the electronic computing market, marked by the IBM 701 in 1952, that “Big Blue’s” financial, research, and marketing muscle began to pull ahead of the smaller operations on both sides of the Atlantic. In 1954, IBM introduced the IBM 650, which went on to sell almost 2000 systems worldwide until it was discontinued in 1962, becoming the first true mass market computer. By that time, everybody else in the worldwide computing market was playing catch-up. IBM’s success came from its resources and dominance in the market for electromechanical punched card tabulating equipment, and there is nothing that government policy in Britain or any other country could have done to compete with it.
Great story. I have noticed that when the great computer hardware companies quit making hardware and switch to services and licensing their products that this begins a downward spiral to bankruptcy. I’m referring to DEC → Compaq → HP. Remember CDC and Cray computers.
There are many money paragraphs in Mr. Hutchinson’s article, which is why it is worth reading the whole thing.
The part I found relevant to government policy was the rejection of Colossus by the insider big brains at Bletchley Park – which would have been the end of the story, had it not been for some relatively low-level guy at the Post Office (!) and a low-caste Cockney punter genius. That is where government policy could have made a difference to the development of an English computing industry.
But that was not the path followed. Thanks for the detailed history of the actual path, from which the take-away seems to be that IBM, being in the commercial world, was able to outperform both the English & American governments. Today we might look at the performance of SpaceX and NASA as an analog. Maybe there is a broader lesson hidden in there about the most effective way to run a society?
This is also a glib and ignorant take on what really happened at Bletchley Park. For the actual experience of those who were there, then, see Jack Copeland’s 2006 book, Colossus, which is a collection of chapters written by those with detailed knowledge of the process of cryptoanalysis, the technologies developed at Bletchley Park, and the management of the development and operational aspects of codebreaking. Many of the contributors were at Bletchley Park at the time, and not permitted to discuss their work until June, 2000, when most of the work was declassified. Fortunately, the team there were mostly young and still around to recount their experiences decades later.
The secrets of their work on codebreaking did not constrain them in any way in applying the electronic circuits and techniques to computers after the war, when many of them migrated to the University of Manchester and developed the pioneering computers built there.
We will simply have to agree to disagree on your interpretation.
The clear point was that the English government had the seed of a huge new industry in its hands at the end of WWII, and failed to help that seed grow. There is a very clear parallel to the situation for the US government in the early 1990s with the demonstration of launch-and-return reusable rocket technology in the DC-X rocket. The big brains in the US, just like the big brains in England, failed to seize the opportunity. In England’s case, their failure may have had a lot to do with their class structure. But what was the excuse for the US in rocket technology?