Computer Memories (continued)
I advised him IBM was many years too late to invent that. He was crushed to find they weren¹t the first Just to complete the thought, other than the SWAC, most but maybe not all computers used the octal, or base 8 system, and used 0-7 as all the symbols needed for groupings of 3 binary bits. And I had to disabuse him of the idea that IBM created a lot of other things also. For example, the IBM 360 could address each and every character of memory, or groups of binary bits, as used in floating point calculations. The IBM man was disappointed to hear that the Control Data 3200 could do that also. Each machine did that in a completely different manner. At Control Data years later, there was a computer that was programmed using the quartic number system, for groupings of 2 binary bits. We counted 0,1,2,3, for the two bits, then on and on, if I do indeed remember that correctly, from 40 years ago.
IBM made great machines and were responsible for most of the good things that happened in the early years of computers. However, they usually did not bring out a machine with the very latest idea, they had to build machines that would work, and work, and work. IBM machines were often a little behind in the state of the art, but their equipment always worked, well.
FROM IBM CARDS TO TELETYPES AT ROCKETDYNE
At North American Aviation I worked at Rocketdyne, where we built the engines for the Atlas, Thor, and Saturn missiles. A production control system was created with a 305 RAMAC as the central computer. The disk file contained a record of each engine part or assembly that was on the factory floor. As an operation was completed, a card was placed in a card reader on the factory floor, and the information was sent to the IBM Room, then punched in a card that was then fed into the IBM 305 RAMAC System.
There were something like 25 disks, each about 2 feet in diameter. I am almost positive we had 10,000, 1,000 character sectors on the disk, a total of 10,000,000 characters. There were two "read/write" heads on sticks that could be watched as they moved in and out and up and down an oily pole, looking for data.
In response to questions about where a part was in the factory, we needed to send information from the RAMAC to Teletype machines located near the rocket-engine production facilities, in several buildings. I went to an engineering company of some kind, and helped them design a box about the size of a small refrigerator that could translate the holes in an IBM card, and send that information over a telegraph line to Teletype machines.
We would receive a request by telephone, record it on a wire recorder, then a key-punch operator would punch a card that would be read into the RAMAC, and the required information would be punched into cards. Those cards would be placed in a modified key-punch machine, which would send the information through the translate machine to the teletype network. People came from all over the world to see what we were doing; there was no such thing as on-line terminals in those days.
Each morning we sent a test message over the computer-to-teletype network at Rocketdyne. We tried to make that message useful, interesting, or perhaps argumentative. One that I thought-up and remember is: "People complain about the idle-rich, and the idle-poor. The one because they are idle, the other because they are rich."
During this time I attended a convention in Los Angeles on Aerospace production control systems using IBM equipment. I was surprised, more than a little disappointed, and partially proud that when the representatives from Chance Vought presented their project, the printed reports, procedures, punch cards, and flow charts were still almost exactly as I had designed them several years earlier. All that had changed was the color of some paper items. It may have been good when invented, but years had passed, and improvements were overdue. My work at that time at Rocketdyne was well advanced over that earlier procedure.
A CONFLICT WITH THE ALL STAR GAME
Back in the mid-50's modified key-punches could read an 80 column IBM card, then send that information over telephone lines to another modified key-punch. It seems the data rate was 240 bits per second, but maybe it was still 120, or perhaps it had increased to 480 by now. Don't remember that detail.
I was trying to send some very important rocket engine data from Rocketdyne in Canoga Park, CA, to a plant in Neosho, Missouri. Data would be sent for a few minutes, then we would be off the air. I would make a call to the telephone company and complain, then a few minutes later, we would be back on the air again, then in a few minutes, off.
Finally I contacted the telephone company in Missouri and discovered that we were trying to use the same telephone line that the radio network was using to broadcast the Baseball All Star Game. When they were on the air, we were off. After some complaint, we would get the telephone line until they complained and took over. What a mess, but there just weren't all that many telephone lines that were good enough to transmit data, even as crude as our equipment was in those days.
MOVE TO ARIZONA
I went to work for CEIR in Ft. Huachuca, Arizona in May 1961. The company's slogan was something like "Intellect, Plus the Amplifiers of Intellect," and the amplifier was a computer.
They had ordered one of IBM's still nonexistent, but planned-to-be-huge computers called STRETCH. It would fill a building, but was only a fraction as powerful as today's desktop computers.
I remember that when I worked at The RAND Corp., IBM engineers talked to each person who was familiar with the existing computers, trying to get ideas for the STRETCH. I can't imagine I gave any words of wisdom, but on the other hand, maybe I did!
I was to work on a computerized simulated War-Game at Fort Huachuca until the STRETCH computer arrived in Los Angeles. It never arrived.
COMPUTERIZED WAR GAME
Fort Huachuca, next door to the Mexican Border and to Sierra Vista, Arizona, south of Tucson, was an old US Army base, most recently used during World War II as a training base for Black soldiers. It had been closed for years, until the Signal Corp decided they needed an isolated place to test communications equipment, and if this was anything, it was isolated. About the largest building in town was the ruin of the nightclub that had been operated by the fighter Joe Louis, during the war.
By the time we moved to Sierra Vista to work for CEIR, the Army had built new buildings on the Fort, and had installed an IBM 709 Computer in the headquarters building. It had the equivalent of 196,000 characters of memory.
To be a little more exact, the memory wasn't organized in characters. There were actually 32,768 "words" of 36 binary bits each, plus a couple other bits that shall remain anonymous. Since in those days we used 6 bits to make an alphanumerical character (an IBM term, I think), it could hold about 196,000 characters. That was considered a lot, in those days. This memory cost many thousands of dollars rent each month. Each day the engineers spent hours just keeping it all working.
There were also four buffered input/output channels, and maybe 15 or 20 magnetic tapes. It was about the biggest computer available on the general market, in those days.
A few civilians and several soldiers prepared the programs, mostly in FORTRAN. In the 1960s, not many people knew what a computer was, or how to use it, so these soldiers were really lucky to be learning a trade that would be so useful. A minor rebellion was barely sidestepped when a Private -- who was discharged after his enlistment term -- was offered a job with our company at a higher salary than some of the Army officers he had recently reported to.
The US Army Signal Corp needed to test new and different kinds of communication equipment under wartime conditions, but did not want to start a war just to find a place to test.
By the time I got there, the project had been underway for a year or two already, and had a year or two to go. A dozen people, including several PhDs in math, retired Army officers, and others who thought they knew how to make this work, made up the staff. Here again, I was the only civilian without two or three college degrees, but since I understood computers and how they worked and what they did, and had worked with computer simulation at RAND, I was useful!
We mathematically described a friendly "Blue" army, and an enemy "Red" army. We described the headquarters, the artillery, the infantry, and how they worked with each other. Then we described the terrain where the war was to be fought - the hills, rivers, roads, trees, fields, etc.
Then we described the Army's communication equipment. How far could it broadcast, the antennae involved, how much wire was needed, how long it took to set up, could it be jammed by the enemy? In addition, equipment of any kind fails to work now and then, and that had to be included in the calculations.
Each of the two armies needed to communicate with their different units of soldiers, and we needed to describe what would happen if a message was received, and what wouldn't happen if it wasn't. If the artillery didn't get the message to fire the guns, and if they were not told where to point them, the war would not be won.
If the message got to the artillery and the guns were fired, we then computed what damage had been done, and how many people had been killed or wounded, how much equipment had been damaged or destroyed. If enough of the "Red" soldiers were casualties, then the "Blue" Army could advance, and that meant the communication equipment must have worked OK. Of course, during the battle, some communication equipment was destroyed, some was damaged, and some just didn't work.
It was mathematically possible to describe the difference in style of command between one commander and another. For example, General George Patton worked in one manner, while the British General Bernard Montgomery had a completely different style.
We had two retired Generals and two retired Colonels on our staff, who were supposed to help us create something that looked real. One of their more important functions was to remind their active-duty military friends that we needed to have our contract continued.