IBM Cards
When I first got out of the service in
1967, I got a job at a large oil company in downtown Baltimore as a stock
clerk for the computer operations.
Back then operations were often mechanical. One top
skill for a programmer was the ability to pick a two foot tall stack of
punch cards from a sorter and turn them 180 degrees and place them into
a slot of another tabulator machine.
One day, taking a break from reading a book by Willy
Ley, I took a brown lunchbag filled with keypunch holes to the 26th floor
of the downtown skyscraper. The stairwell was open in the center down
to two floors below the ground floor. The bag touched a railing about
10 floors down, skimmed another, then broke apart, raining pale confetti
over the stairs.
Dropping the bag and walking to the Little Italy warehouse
on Wednesday to mark the boxes of cards and forms the computer room would
need in the next week were the highlights of my months as a stock clerk.
GE Timeshare for Junior Project
As a junior
in the chemistry department, I had an independent project - calculate
something in the Debye-Hückel theory. The department had an account
with the GE Timesharing system that I would access through a terminal,
with paper tape, in the library.
I spent a lot of time on the project converting mathematical
formula into algorithms until I found the library of differential equations
the system supported. I thought I was home free. I set up the equations
and in a few moments the results typed out.
When I went to doublecheck the first result, it didn't
match my hand calculations. Over the next couple of weeks, I isolated
a bug in the mathematical libraries. The common logarithm was substituted
for the natural logarithm.
I had doubted the Debye-Hückel theory which
had angered the head of the department. I had doubted my methematical
understanding which had not yet included differential equations. Everything
was open to doubt.
Then finally the answer was essentially a typo - in
an computer routine!
