You're Hired
He mentioned to the woman showing him the machine that he was looking for a job, and that he was a graduate student in mathematics at Columbia. At that, she said she would take him straight up to see Robert (Rex) Seeber, co-inventor of the SSEC. John Backus protested. I wasn't wearing a tie, I had a hole in the sleeve of my jacket, and I didn't know anything really about computers, he recalled. No, no, not a problem, the woman insisted, and she ushered him up to see Seeber. After a brief greeting, Seeber proceeded to ask a series of questions that Backus described as brain teasers; such as how to handle the align-ments and additions when using a 10-digit calculator to multiply a 20-digit number.John Backus recalled it as an informal oral examination, with no recorded score. Seeber hired him on the spot. As what? As a programmer, he replied, shrugging. That was the way it was done in those days.
IBN Takes a New Direction
John Backus joined IBM during a period of rapid transition for the company, the industry, and the technology of computing. For decades, manufacturers like IBM, Remington Rand, Burroughs, and NCR had thrived mostly by producing accounting machines for business. These calculators helped managers track payrolls, inventories, and sales as large companies proliferated enterprises that evolved in response to the economies of scale made possible by the rise of mass production, modern rail and auto transportation, and the growth of a national telephone system. But World War II had pushed the technology and the calculator makers beyond electromechanical machines for business and toward high-speed electronics for the aerospace and defense markets.
SSEC - Super Calculator
The SSEC, which IBM called the Super Calculator, reflected those trends. It was essentially an IBM science project, a one-of-a-kind machine designed and built to let IBM’s researchers push the limits of electronic calculators and gain experience.The SSEC was not a stored-program computer, but it was the state of the art when it was completed in 1948, probably the most powerful computa-tional machine at the time. The very name computer, however, was regarded with concern by Thomas Watson Sr. The term was still often used to refer to human clerks doing calculations and, Watson worried, might fan popular fears that the new technology would mean lost jobs. The elder Watson's reluctance to use the term computer was understandable. Public anxiety about computers causing unemployment continued for years. In 1957, a popular movie Desk Set, starring Spencer Tracy and Katherine Hepburn, tapped that nerve with a romantic comedy about the arrival of a computer in a big company. Bunny Watson (Hepburn) is convinced that workers in her department will be replaced by the EMERAC computer and that Richard Sumner (Tracy) is there not only as Emmy minder, but also to slash the work force. Naturally, Bunny's worries prove unfounded, and the movie ends as movies always did in the 1950s.
John Backus got his introduction to programming on the SSEC. The programs he worked on were large scientific calculations, like a program to calculate the position of the moon and nearby planets at any time over years, which required endless crunching of coefficients. It was pure science, Backus recalled.
Programming Not An Easy Task
The research may have been lofty, but the programming was the equivalent of trench warfare. For like the wartime ENIAC, the SSEC had to be reconfigured for each assignment it was given. After figuring how to set up the problem mathematically, the researchers then had to put it onto the machine. This involved devising elaborate flow charts of how the calculations should be funneled through the machine in general terms. Next came the arduous chore of mapping out the calculating steps, instruction by instruction, intricately on pre-printed sheets of paper. Then, the machine had to be set up by hand for each batch of calculations, which switches to flip and which wires to plug into which circuits, to get the Super Calculator flickering and clacking again.On the SSEC, a large program could take months to map out, and then run on the machine for six months. It would grind to a halt every three minutes on average, requiring further ministrations from the programmers. “As a programmer, you had to be there the whole time,” Backus said. When problems surfaced, the clues were to be deciphered by reading the binary coded numbers off the machine’s thousands of lights.
