First Programmable Computer

Michael D. Smith
Dean, Faculty of Arts and Sciences
John H. Finley Jr. Professor of Engineering and Applied Sciences, SEAS

The laptop tucked away in your backpack, the tablet computer on your coffee table, even the smartphone in your pocket all can trace at least part of their lineage to Harvard. The Mark I, built 8 feet high, 3 feet deep, and 50 feet long, was the first programmable computer in the United States. The brainchild of longtime Harvard Professor Howard Aiken, Ph.D. ’39, the Mark I launched the computer age, introducing automated computation as a tool to address problems in the natural, applied, and social sciences.

Consisting of 765,299 parts and 530 miles of cable, the Mark I in 1944 was easily the most complex electromechanical device ever constructed. To build it, Aiken relied on the ingenuity of engineers at IBM, demonstrating the importance of government funding and industry cooperation in large-scale, academic science and engineering projects.

The art of modern programming was also born with the Mark I. Robert Bloch, Robert Campbell, and, most famously, Grace Hopper developed some of the earliest instances of subroutines, branching techniques, code compression, and debugging procedures while at Harvard. Hopper not only wrote the manual of operation for the Mark I, but she documented the first physical bug — a moth found in the machine’s electromechanical relays — and helped usher the term “debugging” into common usage.

Automatic checking and debugging support were critical components of the design of the Mark I, because it produced accurate answers less than 95 percent of the time. Today it is unthinkable that our computing devices would incorrectly sum a column of numbers. But as we make the tiny transistors in modern devices ever smaller and faster, we also make them less reliable and reopen the concerns that Aiken and his team wrestled with 70 years ago.