How the US census led to the first data processing company 125 years ago



The first automated data processing system was used in the 1890 U.S. census.

The population count is required by the US Constitution.

The census has always been charged with political significance. There were controversies in the run-up to the 2020 census.

In my book, Republic of Numbers: Unexpected Stories of Mathematical Americans through History, I tell a story about how the census helped develop the US computer industry. The Tabulating Machine Company was founded 125 years ago on December 3, 1896.

Population growth.

The only use of the census is to allocate seats in the House of Representatives. More populous states have more seats.

The overall population of each state would be reported in a minimalist interpretation of the census mission. The census has never been limited to this.

There was a complicating factor at the beginning, with the Constitution's distinction between free persons and three-fifths of all other persons. The Founding Fathers made a mealy-mouthed compromise between the states with a large number of enslaved persons and the states with relatively few.

The first census made distinctions by sex and age. In subsequent decades, many other personal attributes were probed as well: occupational status, marital status, educational status, place of birth, and so on.

Each census required more effort than the last to collect the data and make it usable. The processing of the census took place in the late 19th century.

It had become boring and error-prone, and it was rarely seen.

The population was growing at a rapid pace, so those with enough imagination knew that the 1890 census would be gruesome.

Advertisement

There is a new invention.

The Census Office had assigned a physician named John Shaw Billings to help with health statistics. He expressed his concerns to a young mechanical engineer who was assisting with the census.

The US Patent Office recorded a submission from the young Hollerith titled "Art of Compiling Statistics" in 1884.

The electric tabulating machine was used in the 19th century.

By improving the ideas of this initial submission, Hollerith would win an 1889 competition to improve the processing of the 1890 census.

The technological solutions were created by Hollerith. The first important innovation was to translate handwritten census tally sheets to patterns of holes in cards. In the 1889 revision of his patent application, Hollerith said, "A hole is punched corresponding to person, then a hole according to person is a male or female, another recording whether native or foreign born."

Special machinery was developed to ensure that the holes were punched with accuracy and efficiency.

The machine was designed to read the card by probing the card with pins, so that only the hole in the card would allow the pin to travel through the card to make an electrical connection.

If a white male farmer passed through the machine, a counter for each of these categories would be increased. The card was made sturdy enough to allow passage through the machine multiple times, for counting different categories or checking results.

The state-by-state numbers needed for congressional apportionment were certified before the end of November 1890.

Advertisement

The 1950 census used a mechanical punch card sorter.

The card was punched.

After his census success, Hollerith went into business. After he retired, the Tabulating Machine Company would become IBM. IBM was the leader in the development of card technology for recording and tabulating large sets of data.

Many businesses used cards for record-keeping procedures in the 1930s. Astronomers were also finding the cards. IBM had developed machines that would change little for decades.

After World War II, card processing became one of the most important parts of the computer industry, and IBM was the third-largest corporation in the world. Card processing was a way of building up the computer industry, with little evidence remaining of the old regime.

Those who have grown up with computers that are portable and easy to use may be unfamiliar with the room-size computers of the 1950s and ’60s, where the primary means of loading data and instructions was by creating a deck. This was the default procedure for many computers.

Grace Murray Hopper said that back in those days, everyone was using punched cards and they thought they would use punched cards forever.

Hopper was an important member of the team that created the first commercially viable general-purpose computer, the Universal Automatic Computer. The US Census Bureau was still hungry to improve its data processing capabilities when the first UNIVAC was delivered in 1951.

The Apollo Moon-landing program and the height of the Cold War used punched cards, but computer users wouldn't use them forever. The direct descendants of the 1890s census machinery would likely have been recognized by Hollerith.

The article was published on October 15, 2019.