The first computers did not execute words such as READ, ADD, or
PAYROLL. They understood numerical codes tied closely to their
hardware. Grace Hopper devoted her career to narrowing the distance
between human intent and machine operations.
Her contribution cannot be reduced to one invention or to the solitary
creation of COBOL. She programmed one of the first large automatic
calculators, led pioneering work on program translation, developed a
language for business data with her team, and championed standards. One
idea connects it all: programs should express the problem being solved
rather than the peculiarities of a particular computer.
Learning to program before the profession existed
Grace Brewster Murray was born in New York in 1906. She studied
mathematics and physics at Vassar College and earned a doctorate in
mathematics from Yale in 1934. She taught at Vassar until World War II
changed her path.
She joined the US Naval Reserve in 1943. The following year she was
assigned to Howard Aiken’s team at Harvard, which operated the IBM
Automatic Sequence Controlled Calculator, better known as the
Harvard Mark I.
The Mark I was an electromechanical machine several meters long. It read
instructions from punched tape and automated long calculations. Hopper
learned to program it, produced tables for the Navy, and helped write
its manual. The work forced the team to formalize a new activity:
breaking down problems, arranging instructions, checking results, and
reusing sequences. Programming was becoming a profession distinct from
building hardware.
What the moth story really tells us
In 1947, the Harvard Mark II team found a moth trapped in a relay. It
was taped into the laboratory logbook beside a note about the first
actual case of a “bug.”
The anecdote is often distorted into a claim that Hopper invented the
term. Engineers had already used bug for technical faults in the
nineteenth century, and the logbook does not establish that Hopper
removed the moth herself. The joke mainly shows that bug and
debugging were already part of computing vocabulary.
That nuance better reflects her importance. Hopper’s place in history
rests not on an insect, but on her work to make programming
understandable, teachable, and independent of hardware.
Why A-0 automated part of programming
In 1949, Hopper joined the Eckert-Mauchly Computer Corporation, soon
absorbed by Remington Rand, to work on UNIVAC I. Businesses wanted
to process customer, inventory, and payroll records, but coding every
operation in machine language remained slow and costly.
In the early 1950s, Hopper’s team developed A-0. A programmer
identified available subroutines; A-0 found their code, loaded it, and
organized their use. By modern definitions, it resembles a loader and
linker as much as a complete compiler.
Calling it one of the earliest compiling systems is therefore more
accurate than declaring it the first compiler without qualification. Its
significance does not depend on the label: A-0 demonstrated that one
program could automatically prepare another for execution. A-1, A-2, and
MATH-MATIC continued this approach and supported Hopper’s then
controversial argument that computers could reliably translate
human-oriented notation.
FLOW-MATIC brought code closer to business work
Scientific languages primarily represented formulas. Business
applications instead described operations such as reading a file,
comparing records, calculating a total, or printing a report. Hopper
believed their notation should reflect that vocabulary.
Her team developed FLOW-MATIC in the mid-1950s. Its English-like
instructions described flows of data. Readability did not remove the
need to learn programming, but it brought code closer to procedures
understood by analysts and organizations.
FLOW-MATIC also taught a lesson about portability. If an instruction
expressed a business operation rather than a hardware detail, a compiler
could translate it for different computers. Programs became less
dependent on a manufacturer. This approach directly prepared the
discussions that produced COBOL.
COBOL was a collective achievement
In 1959, manufacturers, government agencies, and users met through
CODASYL, the Committee on Data Systems Languages. They wanted a
common language for business applications that could run on machines
from different vendors.
Hopper attended the first meeting, and FLOW-MATIC strongly influenced
the project. She had long advocated English-like instructions and
machine independence. Nevertheless, committees produced the COBOL
specification, with a drafting group responsible for its initial text.
Saying Hopper invented COBOL alone erases that cooperation. She was a
pioneer, technical influence, and determined advocate, and later helped
adoption and standardization, notably through the Navy. COBOL let
organizations retain programs as hardware changed. Its longevity
reflects decades of business rules whose replacement would be costly and
risky.
Why portability requires more than a common language
A specification alone does not make programs portable. Vendors may
interpret a rule differently, accept incompatible extensions, or
disagree in edge cases. Hopper therefore emphasized compiler
validation: test suites checked whether an implementation recognized
required constructs and produced expected behavior.
Validation did not prove a compiler free of every defect. It established
measurable, comparable conformity. The progression is continuous: A-0
automated subroutine assembly, FLOW-MATIC used business vocabulary,
COBOL supplied a shared language, and validation sought to preserve that
language’s meaning across machines.
Standardization can look less inventive than syntax design, yet it
determines whether an idea becomes durable infrastructure or remains
tied to one vendor.
From language design to standardization
Hopper continued at Remington Rand and Sperry Rand, publishing, teaching
programmers, and presenting languages as a way to broaden access to
computers.
After leaving the Naval Reserve in 1966, she returned to active duty in
1967. The Navy needed COBOL programs to remain compatible and verifiable
across systems, and Hopper worked on validation and standardization.
She served until 1986 and retired as a rear admiral. In lectures, she
made computing scales tangible, famously using pieces of wire
representing the distance light travels in a nanosecond to explain the
physical cost of time.
Hopper died in 1992. Her legacy lies in compilers and in a conviction
now central to software: languages should bring programs closer to human
problems, while standards should keep that progress from being trapped
on one machine.
Timeline
- 1906: Grace Brewster Murray is born in New York.
- 1928: She graduates from Vassar in mathematics and physics.
- 1934: She earns a doctorate in mathematics from Yale.
- 1943: She joins the US Naval Reserve.
- 1944: She joins the Harvard Mark I team.
- 1947: The moth is discovered in a Mark II relay.
- 1949: She joins Eckert-Mauchly to work on UNIVAC.
- 1952: Work on A-0 is presented.
- 1955–1959: Her team develops FLOW-MATIC.
- 1959: She participates in the first CODASYL discussions.
- 1960: The first COBOL specification is published.
- 1967: The Navy recalls her to work on COBOL standardization.
- 1986: She retires as a rear admiral.
- 1992: She dies in Arlington, Virginia.
- Today: Compilers, portable languages, and standards continue the
work to which she devoted her career.
Frequently asked questions
Did Grace Hopper invent the first compiler?
She led the development of A-0, often described as the first compiler.
Its function also resembles what is now called a loader or linker, so
“one of the earliest compiling systems” is more precise.
Did she create COBOL alone?
No. CODASYL committees designed COBOL collectively. Hopper’s ideas and
FLOW-MATIC directly influenced it, and she later played an important
role in adoption, validation, and standardization.
Did she invent the word “bug”?
No. The word already described technical faults. The Mark II moth
inspired a memorable joke about an “actual” bug, but did not originate
the term.
Why is FLOW-MATIC important?
It showed that a business language could use words close to
organizational procedures while remaining translatable by a computer. It
directly influenced COBOL.
Why is COBOL still present?
Organizations have accumulated decades of proven business rules in
COBOL. Replacing them is expensive and risky, while the original
emphasis on portability helped the code survive successive hardware
generations.
What connects Grace Hopper and John Backus?
Both sought to move programming away from machine code in the 1950s.
Hopper focused on business processing through FLOW-MATIC and COBOL;
Backus led FORTRAN for scientific
calculation.
