All which is beautiful and noble is the result of reason and calculations. Charles Baudelaire
As the 1870 census approached Charles W Seaton at the US Census Bureau was the first to identify that there would be a problem in tabulating the 38 million census forms. A machine was needed to tally the returns and print out the results.
Herman Hollerith, a German-American statistician working with one of his professors at the Census Bureau, met the problem first-hand in 1880. The then 50 million population took eight years to analyse. And by 1890 the population would be 63 million!
After Hollerith’s brief stint at the Census Bureau he taught for a year at MIT’s Mechanical Engineering Department where he began work on a tabulating machine.
A girlfriend’s father provided inspiration, ‘… there ought to be a machine for doing the purely mechanical work of tabulating population and similar statistics… … like a type distributing machine … using cards with the description of the individual shown by notches punched in the edge…’
Hollerith was inspired by the Jacquard Loom but he worked with punched cards to carry data rather than the Jacquard’s mechanical instructions. In 1884 while working at the Patent Office he applied for a patent for his Hollerith Electric Tabulating System.
This tabulating system was a device to make and break electrical circuits by the presence or lack of holes in cards fed through it. The telegraph, another key 19th century development, had a similar approach but transmitted messages rather than counted data.
The guinea pig for Hollerith’s device was the analysis of the City of Baltimore’s 1886 mortality statistics.
Each hand-punched card represented an individual, the holes showing location, ethnicity, occupation and cause of death. There was also a device to sort the cards into slots.
Hollerith’s tabulator won a competitive tender for the 1890 census and processed the 63 million census records in just three months, saving two years against the expected timescale, and reducing the bureau’s costs by $5m. The tabulator was soon managing censuses in the Austro-Hungarian Empire, Canada, Italy, Norway, and was used in Russia by 1897 and the UK by 1911.
Not only did Hollerith start the business of computing, he also set another trend that many would follow – litigation. In 1905 his patent expired and he soon became embroiled in a damaging series of legal actions against the Census Bureau that had unilaterally decided to modify his equipment. The action failed.
In 1907 the Census Bureau hired and funded James Powers, a Russian-born mechanical engineer, to produce an alternative to Hollerith’s equipment. Powers patented a mechanical approach and gained the majority of the contract for the 1910 census.
Power’s cards were the same size but the holes were circular and in 21, 45 or 64 columns. Hollerith’s system had 80 columns and square holes. In 1911 Powers founded the Powers Accounting Machine Corporation and the two operations battled it out around the globe.
In 1911 Charles R Flint bought Hollerith’s business for $2.3m and merged it with others to form the Computing-Tabulating-Recording Company. In 1914 Thomas J Watson Sr joined the organisation, when the chairman died in 1924 it was renamed IBM.
Watson Sr and his son, Thomas J Watson Jr, led IBM for sixty years.