01

From Vassar to the Mark I

Grace Brewster Murray was born in New York in 1906 and dismantled seven alarm clocks as a child before her mother set a limit of one. She studied mathematics and physics at Vassar, took a doctorate at Yale in 1934, and returned to Vassar to teach. When the United States entered the Second World War she tried to enlist, was rejected as too old and too light for her height, obtained a waiver, and joined the Navy Reserve in 1943.

She was assigned to the Bureau of Ordnance Computation Project at Harvard, where Howard Aiken had built the Mark I, a fifty-foot electromechanical calculator driven by paper tape. Aiken handed her the machine and told her to compute the coefficients of an interpolation function. She learned it by writing its manual, five hundred pages describing the architecture and its operating procedures, which became one of the first books about how to program a computer at all. She stayed with the project for the remainder of the war.

02

Compilers and a Moth

In 1947 the Mark II failed and the team traced the fault to a moth caught in a relay. They taped the insect into the logbook and noted the first actual case of a bug being found. Engineers had used the word for decades, but Hopper enjoyed the story and helped make debugging permanent. Her serious contribution began in 1949 at the Eckert-Mauchly company, where she worked on UNIVAC and concluded that programmers should not have to write machine code by hand.

Her A-0 system of 1952 assembled subroutines into a running program, the first practical compiler, and she later recalled that nobody believed a computer could be made to understand anything resembling English. FLOW-MATIC followed, using ordinary command words on business data. When the CODASYL committee produced COBOL in 1959, Hopper was a technical adviser rather than its author, and Jean Sammet and others did much of the design. Her decisive role was in arguing for the idea and then making the Navy adopt it.

03

Standards, Nanoseconds, and the Long Career

Hopper retired from the Navy in 1966 and was recalled within months, a cycle repeated until she finally left in 1986 as its oldest serving officer, having been promoted to rear admiral the year before. Her later work was unglamorous and consequential: writing the validation suites that let the Navy test whether a supplier's COBOL compiler actually conformed to the standard, and pushing for networks of small distributed computers rather than single central machines. She regarded that unglamorous compliance work as the reason the language survived at all.

She was a formidable teacher of non-specialists, handing out lengths of wire cut to a foot to show the distance light travels in a nanosecond, and telling audiences that the most dangerous phrase in an organisation is the claim that a thing has always been done a certain way. She received the National Medal of Technology in 1991 and died in January 1992. A destroyer, the USS Hopper, carries her name, as does a large annual gathering of women in computing.