Apocrypha: Traditions, Myths, & Legends

DIMWIT and Doctor John

PINBALL WIZARD. Prof. John Hancock headed up the team that built Reed’s first computer, DIMWIT.

By David W. Digby ’57

Prof. John Hancock [chemistry 1955-89] arrived at Reed in 1955, the same year I transferred in as a junior in mathematics. He was an Englishman and a bachelor, and lived in the New Men’s Dorm (now Foster-Scholz), where he fit right into our fun and games, such as the “music wars” between his madrigals and our rock ‘n’ roll.

John was fascinated with pipe organs, and fancied building one out of beer cans. He opened the door to his room one morning to find the entire doorframe blocked by empty cans, with a “Good morning John” note attached. Nearly every Sunday he would fill his car with students and visit a local church to check out the pipe organ. He never failed to sign the guest register with a flamboyant imitation of the famous signature on the Declaration of Independence.

John’s favorite research project was the quest to synthesize dodecahedrane, a hydrocarbon which featured 20 carbon atoms in the shape of a dodecahedron (one of the five Platonic solids), with a hydrogen atom bonded to each carbon atom.

We kept in touch after I graduated and went on to Oregon State University. One day he told me that he had been wondering what would happen if some of the hydrogen atoms on this hypothetical molecule were replaced with chlorine atoms, which led to the question: “How many different ways are there to do this?”

John passed that question onto me, since I had access to a computer at OSU. (It would be several years until Reed purchased its first commercial computer, an IBM 1620 that was installed in the basement of Eliot Hall.) 

I wrote a program to enumerate the answer, which took about a hundred hours to run on OSU’s gigantic vacuum-tube ALWAC computer. The result was a printout of 17,000 unique solutions. Each of these was a “canonical” value representing 120 different rotations and reflections of the same shape, due to the symmetric nature of the dodecahedron.

This led to another question: “What if I make one of these compounds, and the value I find for it is not on the list?” That called for another program to rotate the value he found into the corresponding one in the list. This only took a few seconds to run, but there was a problem—Reed had no computer.

John came up with a fiendish idea. Pinball was then illegal, and the county sheriff had a warehouse full of confiscated machines. He couldn’t sell them, but he could donate them in the name of science. Pinball machines are full of relays, and John suspected we could make a primitive computer out of them. After looking over the available parts, we decided to give it a shot.

I designed and programmed the machine, Reed students put it together, and John named it DIMWIT for Dodecahedrane Isomer Machine With Internal Translation. It worked—sort of. It took about 5 minutes to get an answer, which was correct about 10% of the time. This was not quite satisfactory, but John had another ace up his sleeve.

It turned out that the telephone company was upgrading some of its machinery and was willing to give us a few tons of older parts. Telephone relays are much more reliable than pinball relays—after all, customers get upset when their calls don’t get through.

Using telephone relays, DIMWIT-II got the correct answer 90% of the time—not too bad, considering that the occasional wrong answer was easy to pick out. It did take a long time to produce a result, so John had one final request: “Can we have it turn on the coffee maker when it has finished the calculation?” Unfortunately, this was not programmed into DIMWIT-II.

The quest for dodecahedrane was finally achieved in 1982 by a team at Ohio State. By then, of course, the most basic programmable pocket calculator could complete our calculations in an instant. Nonetheless, I have always been proud of DIMWIT, and so, I think, was John. He died of a heart attack in 1989, and I still miss him. 

Go Further

Reed is seeking alumni to help us develop a computer science program. Find out more in the April 2015 Sallyportal post Digital Pioneers Help Reed Design CS Program or email Gaynor Hills ’95.

—Author info: David Digby worked in computer engineering, design, and genetic analysis. He lives in Atlanta, Georgia, and is currently a professor at the (imaginary) University of the Dancing Mind.