May 9, 2020
Today we’re going to cover the Atanasoff–Berry computer (ABC), the first real automatic electronic digital computer.
The Atanasoff-Berry Computer was the brainchild of John Vincent Atanasoff. He was a physics professor at Iowa State College at the time. And it’s like he was born to usher in the era of computers. His dad had emigrated to New York from Bulgaria, then a part of the Ottoman Empire, and moved to Florida after John was born. The fascination with electronics came early as his dad Ivan was an electrical engineer. And seeking to solve math problems with electronics - well, his mom Iva was a math teacher. He would get his bachelors from the University of Florida and go to Iowa State College to get his Masters. He’d end up at the University of Wisconsin to get his PhD before returning to Iowa State College to become a physics professor.
But there was a problem with teaching physics. The students in Atanasoff’s physics courses took weeks to calculate equations, getting in the way of learning bigger concepts. So in 1934 he started working on ideas. Ideas like using binary algebra to compute tasks. Using those logic circuits to add and subtracted. Controlling clocks, using a separate memory from compute tasks, and parallel processing. By 1937 he’d developed the concept of a computer. Apparently many of the concepts came to him while driving late at night in the winter early in 1938. You know, things like functions and using vacuum tubes.
He spent the next year working out the mechanical elements required to compute his logic designs and wrote a grant in early 1939 to get $5,330 of funding to build the machine. The Research Corporation of New York City funded the project and by 1939 he pulled in a graduate student named Clifford Berry to help him build the computer. He had been impressed by Berry when introduced by another professor who was from the electrical engineering department, Harold Anderson. They got started to build a computer capable of solving linear equations in the basement of the physics building. By October of 1939 they demonstrated a prototype that had 11 tubes and sent their work off to patent attorneys at the behest of the university.
One of the main contributions to computing was the concept of memory. Processing that data was done with vacuum tubes, 31 thyratrons, and a lot of wire. Separating processing from memory would mean taking an almost record player approach to storage.
They employed a pair of drums that had 1600 capacitors in them and rotated, like a record player. Those capacitors were stored in 32 bands of 50 and because the drum rotated once per second, they could add or subtract 30 numbers per second. Thus, 50 bits. The concept of storing a binary bit of data and using binary logic to convert that into more of a zero or one was the second contribution to computing that persists today.
The processing wasn’t a CPU as we’d think of it today but instead a number of logic gates that included inverters and input gates for two and three inputs. Each of these had an inverting vacuum tube amplifier and a resistor that defined the logical function. The device took input using decimals on standard IBM 80-column punched cards. It stored results in memory when further tasks were required and the logic operations couldn’t be handled in memory. Much as Atanasoff had done using a Monroe calculator hooked to an IBM tabulating machine when he was working on his dissertation. In many ways, the computer he was building was the next evolution from that just as ENIAC would be the next evolution after. Changing plugs or jumpers on the front panel was akin to programming the computer. Output was also decimal and provided using a display on the front panel.
The previous computers had been electro-mechanical. Gears and wires and coils that would look steampunk to many of us today. But in his paper Computing Machine For the Solution Of Large Systems of Linear Algebraic Equations (http://jva.cs.iastate.edu/img/Computing%20machine.pdf), Atanasoff had proposed a fully digital device, which they successfully tested in 1942. By then the computer had a mile of wire in it, weighed 700 pounds, had 280 vacuum tubes, and 31 thyratrons.
The head of the Iowa State College Statistics Department was happy to provide problems to get solved. And so George W. Snedecor became the first user of a computer to solve a real problem. We have been fighting for the users ever since. But then came World War II. Both Atanasoff and Berry got called away to World War II duties and the work on the computer was abandoned.
The first use of vacuum tubes to do digital computation was almost lost to history. But Mauchly, who built ENIAC would come later. ENIAC would build on many of the concepts and be programmable so many consider it to be the first real computer. But Atanasoff deserves credit for many of the concepts we still use today, albeit under the hood!
Most of the technology we have today didn’t exist at the time. They gave us what evolved into DRAM. And between them and ENIAC, was Konrad Zuse's Z3 and Colossus. So the ‘first computer” is a debatable topic.
With the pioneers off to help win the war, the computer would go into relative obscurity. At least, until the computer business started to get huge and people didn’t want to pay Mauchly and Eckert to use their patent for a computer. Mauchly certainly would have known about the ABC since he saw it in 1941 and actually spent four days with Atanasoff. And there are too many parallels between them to say that some concepts weren’t borrowed. But that shouldn’t take anything away from any of the people involved. Because of Atanasoff, the patents were voided and IBM and other companies saved millions in royalties. ABC would be designated an official IEEE Milestone in 1990, 5 years before Atanasoff passed away.
And so their contributions would be recognized eventually and those we can’t know about due to their decades in the defense industry are surely recognized by those who enable our freedoms in the US today. But not to the general public. But we thank them for their step in the evolution that got us where we are today. Just as I think you dear listener for tuning in to this episode of the history of computing podcast. We are so lucky to have you.