“Is this the first step in an accounting revolution, or merely an interesting and expensive experiment?” asked The Economist in an article devoted to the world’s first business computer, nearly 50 years ago. The machine, the Lyons Electronic Office (LEO), was built by Lyons, a British catering company. On November 17th 1951, it ran a program to evaluate the costs, prices and margins for that week’s output of bread, cakes and pies, and ran the same program each week thereafter. In February 1954 LEO took on the weekly calculation of the company’s payroll, prompting an article in these pages.
Other computers had been used to run one-off calculations for businesses, and many firms used mechanical or electrical calculators. But LEO was the first dedicated business machine to operate on the “stored program” principle, meaning that it could be quickly reconfigured to perform different tasks by loading a new program. It occupied 5,000 square feet of floorspace, contained 6,000 thermionic valves, and its mercury-delay-line memory could hold 2,048 instructions. LEO was built by a team led by John Pinkerton, and its design was based on Cambridge University’s EDSAC computer.
In 1954 Lyons spun off a separate company, LEO Computers, to build machines for other firms. Today the company is nowhere to be seen. British firms, unlike their American counterparts, were sceptical about business computing; American manufacturers soon took the lead. LEO Computers vanished in a series of disastrous mergers. “I know of no computer merger anywhere where there has been added value from the merger of competing forces of engineers, marketers and programmers,” David Caminer, LEO’s software guru, said last week at a conference to mark the 50th anniversary.
The troubled HP-Compaq merger suggests that the industry has yet to learn this lesson. And one big question remains unanswered. “Might computers not have a valuable contribution to make in improving business efficiency?” asked our 1954 article on LEO. The jury is still out on that one.