The MNINE Weight Loss Program - Part 1

The summer between high school and college, I worked as a programmer at Giffels Associates, an architectural and engineering firm in Detroit.

My nominal boss was "Nathan" (short for his middle name, Nathanchandra), but in practice I was supervised by an engineer named Stan, in whose office I had a table. Nearby was the office of Stan's colleague Chris - not Christopher - so named because he had been adopted on Christmas Day.

My responsibilities included writing a project time accounting system, and maintaining various engineering programs. Of course, I had no experience in either accounting or engineering. But somehow that didn't turn out to be a problem.

MNINE listing - obverse

Giffels had a refreshingly simple way of naming their programs. Each program's name was a single letter - the first letter of the department's name - followed by a number, which was incremented for each new program. Since I worked for the Mechanical Engineering department, our oldest program was named M1. However, for reasons unknown to me, the program names were typically rendered by spelling out the number in English. Thus, the engineering program I worked on most frequently - M9 - was rendered as MNINE. It's unclear what would have happened once the 11th program was written; the name MELEVEN would likely have exceeded the system's identifier length limit.

In years since, having had to cope with the busywork associated with periodic product renaming, I've had reason to long for the simplicity of the Giffels software naming convention.

Giffels had a couple of computers: an old IBM 1130 and a seemingly even more ancient Univac 9300. The IBM was a modest, mini-computer-like machine, slow but easy to use if you could get by with the very limited disk capacity. The Univac was rarely used, to the frustration of some higher-ups who felt it was going to waste.

I wrote the time accounting system on the IBM 1130. However, neither of these computers was worth much if you wanted to do engineering calculations. Thus, Giffels rented time on an elderly Univac 1108 located at Illinois Institute of Technology in Chicago and rented through Utility Network of America (UNA). This computer was old enough to have drum, rather than disk, storage. Unlike the 1130, its operating system, EXEC 8, was rather complex, and I never really learned more about it than what I needed to edit, build, and run the engineering programs.

I don't know why we had to resort to using a computer a couple of states away. But it worked for us. We had a remote batch system that was linked to IIT via a synchronous dial-up line. We submitted jobs from Detroit, they ran on the mainframe in Chicago, and the results were transmitted back to the office, where they were printed.

MNINE listing - reverse

For some crazy reason, the output was always printed on carbon copy forms, so there were two copies of each printout, plus the piece of carbon paper between the two copies. When I went off to college, I took with me numerous spare printouts of MNINE. This wasn't because I thought I would need to calculate the flow of fluids through building pipes. Rather, I used the backs of the printouts as scrap paper, or to make carbon copies of my philosophy papers. (Evidently I felt that my essays were so brilliant that the professors might not want to return them, leaving me with no way to publish them.) The carbon paper that came with the forms was pretty good quality, and could be reused several times before it wore out. I was living high on the hog!

Carbon-copy forms

One assignment I remember involved enhancing a program to calculate the width of pipe needed to handle a given flow of steam. (To this day, I have to resist the urge to pronounce that word as "stem", the way that Nathan, who spoke pretty good English, unaccountably pronounced it.) Surely MNINE must have had the ability to compute this sort of thing before I arrived, so perhaps this particular calculation was for a new type of pipe. At any rate, Stan showed me a book in which there was a graph of desired steam flow vs required pipe size. The curve looked well-behaved. So, using a ruler, I simply read a number of points from the graph, then fed them into a curve-fitting program. I tried all possible curve types, even though only a few were legitimate candidates. Then I programmed the winning formula (which fit the points quite closely) into the application.

On another occasion, engineers on a different floor discovered a bug in an application that had been written in ALGOL-60 by some wise guy who had moved to another department and was no longer available. Now, ALGOL-60 was a significant step forward in programming languages in its day, and a lot of ALGOL-60 compilers were written. But especially in the US, the language never really caught on, and very few engineers knew it. My colleagues held the puzzling attitude that computers were mere tools to accomplish some task, and had little interest in learning a new language just to fix someone else's bug. Thus, no one would touch that program. Stan lent me out to the team to see if I could help. Fortunately, the bug proved easy to spot, and did not require intimate knowledge of ALGOL oddities such as call-by-name. So I came off looking like a genius.

To get to that other floor, I walked up a couple of flights of stairs. (I'm not sure why I bothered with stairs in those pre-fitness days.) In the dimly-lit stairwell, I often encountered a couple of shady-looking guys, clutching dollar bills and muttering numbers to each other. Someone - maybe Stan - told me they were playing "Dollar Bill Poker". I never figured out just what dollar bill poker was, or how it could be a fair game. Running this gauntlet was rather intimidating to an 18-year-old. But I kept taking the stairs, and tried to ignore the shenanigans.

At one point during the summer, Stan and Chris were struggling with a difficult engineering problem. Eventually, in their desperation, they cooked up a brilliant scheme: they would concoct a bogus solution to the problem, write it up as an article, and submit it to some second-rate engineering journal. The solution would have to be plausible enough to make it past the hopefully low editorial standards of the journal. Then all they'd have to do is sit back and wait for the angry letters to the editor to roll in. Surely one of those letters would have a legitimate solution. Yes, there'd be a cost in terms of pride, but pride is overrated. Sadly, Stan and Chris never followed through.

Toward the end of my time at Giffels, the powers-that-be prevailed upon Stan to get me to convert some existing software to run on the disused Univac 9300. Unfortunately, the only compiler we had for the 9300 was for RPG, a disgusting language. I'm sorry to say that I did not tackle the job with much gusto, and no software conversion was completed before the summer ended and I left for Michigan State University.

It wasn't until decades later that my exposure to the Univac 1108 was to reap health benefits for me. But that is a story for another day.