On this day in 1956 direct keyboard input was introduced on MIT’s Whirlwind, a vacuum tube computer developed for the U.S. Navy. The direct keyboard input method was revolutionary at the time since most programmers used punched cards, dials and switches to offer instructions to computers.
The Whirlwind was also among the first digital electronic computers that operated in real-time for output, able to calculate in parallel rather than serial, and use magnetic core memory.
When I first got into the mainframe computer industry I had the opportunity to work with mini-computers that still used magnetic core memory. Many mainframes used in the banking industry at that time were maxed out at 64K of memory, yes that is K, not M, must less G. It’s not a typo. And these computers ran large Datacom networks, high-speed check sorters and a robust operating system simultaneously within that 64K. The operator console was a teletype machine which I imagine was the “keyboard” mentioned in this article. I also worked with mainframes that still used diode logic, the precursor to transistor logic. Each logic gate consisted of diodes, capacitors and resistors. The components were mounted on a plastic “block” which plugged into a backplane. One logic gate per plastic block. And there were still tape drives in use that used vacuum tubes for the preamp off the read head. Transistors were still too noisy at those low signal levels.
Then there was head-per-track disk drives. One huge platter about four feet in diameter, covered with hundreds of read-write heads, one head for every track on the disk. Heads were loaded (flown a couple of thousandths of an inch) from the platter surface. There was an air compressor and positive air pressure was used to “load” the heads but the air coming off of the platter itself kept the head at the proper distance from the platter. Of course the same is true for how heads are flown on disk drives today. The platter was extremely heavy with one very large motor driving it. Each time you spun the drive up the motor was under such a heavy load that the winding got so hot they smoked. If the drive was shut down for repairs, preventive maintenance or power failure the platter would spin for about 10 minutes. Those had to be tough motors but I never saw one go bad. Then there were the drive belts from the motor to the platter hub. It was a leather-link belt. It consisted of individual pieces of leather about 1/8 inch thick, ½ inch wide and a little over an inch long. There was a metal pin with a head on it attached to one end of the link. The other end of the link had a slot cut into it lengthwise with a hole on the inside edge of the slot. These links fastened one to the other, pin to slot, until you reached the desired length. Obviously these belts would stretch so on a regular PM schedule we had to shut the drives down, check and adjust the belt tension which always had to be tightened. At some point there would be no more slack to adjust the belt tighter so you loosened the belt, removed a link, and adjusted for proper tension again. Never saw one of these belts break. Never had to replace a belt in its entirety. Just remove a link every now and then. Old school but it worked so well.
And it never ceased to amaze me how many programmers continued to use punch cards long after they had other means available to write and load their software. Huge decks of cards loaded into high-speed card readers used to load a program that was already stored on disk. “It doesn’t make any sense Admiral,” from the line in the movie “Midway”. And a change to the program meant punching a whole new deck of cards or if there was only a minor change punching the new cards, searching through the deck and pulling and replacing individual cards as needed. And woe be to the person who dropped a deck of cards.
How far we have come. How old I feel.
for now we can type on the glass screen