Early Engineering Computing

In the early 1970 I was assigned to work with the Advanced Systems organization. This was a part of the company that prepared proposals for new aircraft and other systems, and did the Research and Development work for those systems.

You could tell you were on the cutting edge of technology because there were several tables where you could sit and use one of the WANG calculators owned by the company. The calculator itself was in a large box under the table, and there were four integrated keyboard/display units on the table top, all connected to the box. The displays were all small screens with orange bulbs that could light up to show a single row of numbers, and, of course, the correct placement of the decimal point. Since the WANG could do true four-function calculation (add, subtract, multiply, and divide), it was considered a huge leap forward from standard office adding machines, which were unable to do division. Most engineers, however, still preferred to use their trusty slide rules whenever possible.

The company also had a computer system we could use, composed of three IBM 360’s linked together. It was called the triplex, and was housed in a large room and tended by a small army of technicians and computer experts. Input to the triplex was via punch cards, so there were keypunch machines situated around the floor that people could use to make their own cards. If you had a big enough punch card job, and the budget to pay for it, you could submit your coding sheets to a group of ladies who did nothing all day long but transfer data from coding sheets to punch cards. The ladies could produce punch cards much faster than most engineers, and were generally more accurate.

Once you had your computer program on punch cards, you submitted your stack to a Dutch window at the computer room, along with appropriately filled-out request forms, and went back to your desk to wait. Depending on what you wanted the computer to do, and the work load on the triplex at that given moment, you might wait a few hours or a few days for your job to run. Usually you didn’t get a phone call to tell you your data was ready, but most people learned to estimate about how long to wait and simply checked back to the Dutch window when they thought the time was about right.

Your output, when it was ready, usually came on large sheets of fan fold paper, about 9X17 inches with alternating white and green bars to make it easier to read across rows of numbers. If you were unfortunate enough to have made an error in your program, or with your input data, your output sheets would be blank and contain some cryptically worded error message. The messages could often only be interpreted by a computer expert, although common errors soon came to be thoroughly understood by users. After finding out what the error message meant and correcting the problem (often a keypunch mistake you had made yourself at one of the machines while preparing input cards), you resubmitted the job and went through the entire process all over again.

Eventually you got good results and put the output sheets into binders, and the punch cards into boxes. You could tell how complex a computer program was by the number of boxes it took to hold the cards for a single run.

! * !

Personal Calculators Come To Vought

In the early 1970’s there were few personal calculator options. You could use the Vought main frame computers for major efforts, or the WANG four-function calculators for smaller activities. The WANGs were housed in a box almost as big as a modern-day PC, and had four keyboard/display units that sat on a table and could be used simultaneously by four people.

But the only affordable pocket-size calculating devices available before that time were adding machines that could add, subtract, and multiply, but not perform division. To get that fourth function, division, raised the cost dramatically.

So it was with intense interest that we heard of an offer from Vought to allow purchase of a brand-new Texas Instruments (TI) scientific calculator for only $200. These were actually calculators that had failed TI quality control tests during manufacture, but had been refurbished and carried a full guarantee. Vought struck a deal with TI for their purchase, and you could sign up for payroll deduction to pay for them. The refurbished machines were collectively known as “E-models”, since they had the letter “E” stamped into the back of their plastic case. This apparently served to distinguish them from regular production calculators.

The SR-51 model immediately caught my eye. It was loaded with statistical functions, such as a random number generator, as well as trigonometric and other capabilities. But the most awesome task it could perform was linear regression. I had often reduced data from field maintenance activities through use of regression, but the task was long, labor-intensive, and contained many opportunities for error. The last regression calculation with the WANG had taken half a day and required several ruled sheets of paper filled with pencil notations. The worst part of the WANG process was the nagging doubt when it was over about whether you had made a mistake, and whether your answer was actually correct.

I eagerly read through the thick users manual when my SR-51 finally arrived, zeroing in on the regression section. Then I dragged out the pencil sheets from the last WANG session, the one that had taken a half-day. In less than ten minutes with the SR-51 I had my answer. And best of all, hallelujah!, the answer was the same as the one I had gotten from my WANG effort.

With regret, I tenderly replaced my bamboo slide rule into it’s leather case, never to be withdrawn again, finally fallen victim to the age of digital electronics.

That SR-51 was carried in my briefcase for many years, and I wore out and replaced several rechargeable battery packs. The number of regressions and other calculations that were done with that machine saved me an untold amount of time compared with what would have been necessary if I hadn’t had it. It was kept long after much more capable, and much cheaper, personal calculators were available. It was completely reliable, and was only retired after the keyboard finally failed. At that point, the SR-51 had been in my service for 20 years. And I kept the users manual for several years after the calculator itself was gone because it contained so many good equations and example problems.