Return to A Short Bio
Bio - relative to digital computing
I'm from a small rural town, we think simple.
I'm told there are two kinds of people (or is it politicians?).- liars - damn liars
So, hoping to avoid the "damn liars" category, here goesI early discovered I had 10 easily countable fingers, including 2 thumbs,
and so fell in love with the decimal system.
All else is funny counting,000, 001, 010, 011, 100, ...
is not really natural to me ;-))My mother wanted to get me into college as soon as possible, so got kindergarten to bend the rules about minimum age for admittance. I was the youngest, smallest, least physically able, most socially inept of my class. I had thick glasses for my eyes, ("Hey - 4 Eyes, get over here!!") and soon had braces for my teeth :-((
- I was a prototype for Dilbert of the "funny" papers.
- Even girls could and did physically threaten me. I didn't have enough social status to be worth beating up on.
So I found of solace in being alone, and being curious, fell into reading about science. In the seventh grade I discovered the chemistry room for ninth graders and the chemistry teacher, Miss Marquart - who found I was fascinated in the chemistry she taught and loved to talk about.
By the ninth grade I felt I had absorbed all the science in the local library. I didn't include arithmetic, and math as "science". I thought they were foreign to sparks, magnets, chemistry, fire works, jet engines, ... the fun, spectacular stuff.
So in the 10th grade, I started taking the bus (Saturdays) to the nearby big city, St. Paul, Minnesota, and visiting the big public library there, and an associated technical library next door. I would borrow thick aeronautical books that discussed "slugs" "bars" "g's" and other unknown nouns. I paged through them, and even slept on them, hoping that knowledge would somehow pass from the pages into my brain - Apparently that is not the the path to knowledge for me :-((
During the 11 grade, my mother said there was a Saturday class in St. Paul that might be interesting. - The class was how to wire a large complicated plug board of a machine I had never seen - likely an IBM 402 or 405 accounting machine. The whole exercise seemed puzzling, I didn't "get it", so quit after a couple of session.
Not computer related, but fun for me.
Around 12th grade, (1949) Saturdays I start lurking about the physics building at the University of Minnesota. Since the building was locked, I lurked about, trying to go in with folks, or dash into closing doors left by departing folks. I found out and saw the inside of the big silo back of the Physics building, a 4,000,000 volt Van de Graff generator/linear accelerator. Kind students let me into the room at the base, where they showed me saw complex wiring, and strange instruments. I was in heaven. They said that the space between the Van de Graff generator and the outside steel silo was filled with pressurized DRY air to prevent electrical leakage and sparks. They said that when there was a "discharge" (brief arc to the steel "ground") it sounded like someone hitting the steel silo with a hammer. ... lots of time goes by ...
(1957) - out of the army, where I was in the Nike program and the Nike Analog Computer.
I found an article in the St. Paul Dispatch that there was a class for school teachers about a new computer that Remington Rand Univac was making - that might be interesting - so I showed up - with out registering ;-)). A designer, I think it was ?Phelps?, was instructing the class. It was about their new machine, the UNIVAC 1103A, a digital computer with magnetic core memory. sales brochure of 1103 (magnetic drum memory) 2.5 megabytes.
I was permitted to "audit" the class, and receive training material. :-)) By the time the introductory classes were over I had a notebook of blue prints, carefully folded into 4 parts to fit the notebook. I kept that priceless notebook for many years, through many moves - discarding it about 1980 - who would be interest in that old stuff? The next class was to be programming the machine, and formal registration would be required :-(( I decided that I could not fake being a school teacher, and did not try to register. So ended my first brush with digital computers. 1960, I was a senior at the Milwaukee School of Engineering and took a class on the newly arrived Royal Precision LGP-30. I was fascinated, hooked, lost forever, ... I decided to make a simulation of my Senior Thesis "An Analog Multiplier" which I had constructed out of some 9 vacuum tubes using a Phantastron circuit for one of the two input variables. I got the key to the computer room, and probably spent 30 hours a week there. Between the above two projects, my senior academic grades declined from "very respectable" to "buddy, you better watch it" :-(( But I was having a BALL !!! with my true love, digital computers. :-))
much time passes
While at Control Data, in the late 1960s, between projects, I did some benchmarking, and learned about "flexibility" ;-)) and hand coding challenging sections of benchmarks :-| - Ed PS Oh golly, earlier, while at General Electric, I found how the GE salesmen estimated how many records per second a 225 could process. Find the maximum speed of the output device - printer - magnetic tape and assume the computer was fast enough to keep the output device going at full speed. Likely the GE 225 was designed to compete with the IBM 1401. The 225 could haul a 20 bit word out of memory (21 us) in about the same time a 1401 could get 3 characters - The 225 was well adapted for parallel I/O and compute :-)) If a customer complained (loud enough) that the 900 line/minute printer http://www.ed-thelen.org/comp-hist/GE-235-SystemManual.pdf page VI-1 was going at 100 lines/minute instead of full speed during a COBOL written run of importance, an Applications Engineer was sent to hand code in assembly some of the I/O code - to replace calls to the standard I/O routines which were big, generic, and slow. Say you needed to print a date on each bill, three binary numbers to say three decimal 2 digit numbers on a print line. Instead of popping say 13 parameters and addresses onto the stack and calling the generalized I/O routine three times - You could take a binary value into the extended AQ register, divide by 10 convert both registers to the appropriate printer codes - stuff 'em away into the print line, three times and be done in less than 1 % of the time. Do that enough places in the code, and your printer output might go from 100 lines/minute up to 700 lines/minute - enough to pacify the customer :-)) One of those guys showed up at my site, showed me how, and the customer was less mad :-)) He had lots of other things to be irritated at !! http://www.ed-thelen.org/EarlyGE-Computers.html#crappyness
Originated October 13, 2009