Harvard Mark I

Harvard Mark I

The Harvard Mark I — officially IBM's Automatic Sequence Controlled Calculator — was the machine that proved a computer could be big: 765,000 electromechanical parts, a 51-foot case eight feet high, and a rotating shaft running its length, distributing mechanical motion to 72 adding counters like a factory floor. Conceived by physicist Howard Aiken in 1935, pitched to IBM in 1937, and personally approved by Thomas Watson Sr. in 1939, it was built at IBM's Endicott plant, shipped to Harvard in February 1944, and began working for the Navy Bureau of Ships in May.^[1]

The machine's timing made it a war asset before it was a university computer: one of its first programs was run on 29 March 1944 by John von Neumann, calculating whether implosion could detonate an atomic bomb — the Manhattan Project's first computing job. When not doing secret work it printed mathematical tables, which is what Charles Babbage had wanted a machine for in 1837 — and the connection was literal: Aiken had found his way to Babbage's Analytical Engine via a demonstration model Babbage's son had donated to Harvard decades earlier, and the ASCC "brought Babbage's principles almost to full realization."^[2]

Programmed on paper tape, staffed by legends

The Mark I read its instructions from punched paper tape and computed in decimal — 23-digit numbers, additions in about a third of a second — with no conditional branching in the original design (loops meant physically taping the tape into a loop). Its operators, who called it "Bessie, the Bessel engine," included a young lieutenant (junior grade) named Grace Hopper, who wrote the machine's manual of operations and would go on to invent the compiler.^[3]

A note for the record: the famous moth taped into a logbook — "first actual case of bug being found" — was found in the Mark II in 1947, not this machine. The Mark I's own folklore is better anyway: it is the computer on which computer programming was first documented as a profession.^[4]

Where it sits

The Mark I was obsolete almost at birth — ENIAC's electronics would outrun its relays within two years — but that misses what it was: the first automatic general-purpose calculator, reliable enough to run for fifteen years until its 1959 disassembly. It was the last and greatest monument of the electromechanical era, the machine that proved Babbage's century-old idea in hardware, and the classroom where Hopper's generation learned that programming was a job. Parts of it still stand in Harvard's Science and Engineering Complex.

Check your understanding

• What was the Mark I's official IBM name — and who designed and funded it?
• What famous physicist ran one of its first programs, and for what project?
• Which Grace Hopper story belongs to the Mark II, not the Mark I?
• In what sense was the ASCC "Babbage's principles almost fully realized"?