3 ms·
Flowers was one of the first people to push data around with tubes. There was also an IBM group working on that before WWII, and by 1943 they had the "Vacuum Tu
by Animats 3y ago
Flowers was one of the first people to push data around with tubes. There was also an IBM group working on that before WWII, and by 1943 they had the "Vacuum Tube Multiplier" working. That eventually became the IBM 603 in 1946.
I've seen one in an IBM historical display; it looks like a very large hard-sided suitcase made of perforated black metal. The 603 was a limited product built to answer the question of whether electronics could be deployed to field locations and serviced by IBM maintainers. Only 100 were built. They solved the problems of making tubes work in the field, so the IBM 604, the production version, was built in 1948. This was the beginning of IBM's line of cost-effective electronic business machines, ending with the IBM 650, which was a full scale computer affordable by businesses. Knuth learned to program on an IBM 650.
Much of early computing was about getting memory that worked, tubes that worked, capacitors that worked, connectors that worked, and making the things maintainable. All that didn't really come together until the IBM 1401, which was the first mass-produced computer that Just Worked.
It wasn't an architecture problem. The general ideas there had been worked out in the 1930s by Eckert at Columbia University, who came up with some insane mods to IBM tabulating machines to use them for scientific purposes.
This kludge, in 1934, was the first automatic number-crunching computing machine.[1] Eckert went on to become a founder of UNIVAC.
Understand, all this was being done in an era where a reliable AM radio was hard.
[1] http://www.columbia.edu/cu/computinghistory/switch.html http://www.columbia.edu/cu/computinghistory/switch.html