Mechanization of bookkeeping in Western Europe was similar to expansions of offices in the United States between the late nineteenth century and the Second World War. This mechanization enabled the Tabulating Machine Company and the Powers company to extend their positions in the United States to Great Britain, Germany, and France. However, their control was not tight enough to prevent discernible differences from surfacing among punched-card-based bookkeeping systems in the various countries in Europe and in the United States in the interwar years.
British punched-card applications were shaped by the emergence of the earliest letter printing tabulator, which was marketed by the British Powers company in 1921. Having this capability failed to become a major asset for the company, which only developed an alphanumeric system after the Second World War. Instead, the British Powers company focused on developing and marketing cheap punched-card machines using small nonstandard cards, a strategy that proved rather successful.
In Germany, the Powers agency marketed the alphabet printing tabulator, which had been developed by the British Powers company, but its success was limited. The German demand at that time called for numeric calculation capability—while the need for alphanumeric punched-card systems only emerged during the Second World War. In contrast, French punched-card applications were already developing alphanumeric printing systems in the 1930s; these subsequently became crucial for establishing a national register between 1940 and 1944. These bookkeeping systems were shaped through a combination of varying demand in the different countries, different national business conditions, and different relations to IBM and to the Powers company in the United States.
The first development of punched cards for bookkeeping in Great Britain originated in the sphere of social legislation. The welfare reforms of the Liberal governments from 1906 to 1914 extended the range of government intervention, although the significance of much of the legislation was in its pioneering nature rather than in the level of expenditure involved. Three of the most important innovations were old age pensions, health insurance, and unemployment insurance. These innovations were derived from two basic principles. The old age pension program created in 1908 was based on local organization of public expenditure, which was simple to administer.1 In contrast, the compulsory national health insurance and unemployment insurance from 1911 were financed through contributions related to the individual employee. Virtually all employees were covered and thereby entitled to sickness benefits and assistance with medical treatment. The national health insurance program relied on weekly contributions from all employees and their employers, supplemented by state contributions. This program required much more administration than one based exclusively on public funding. The collection of contributions and payment of benefits were administered by “government-approved societies,” mostly consisting of trade unions and commercial insurance companies, which individual employees could opt to join.2
The government-approved commercial insurance companies were all industrial life insurance companies. This market was dominated by the Prudential Assurance Company in London, which held several million policies. In 1911, the only aids the company had to administer their policies were simple key office machines and address plates that were used to address notices and receipts.3 The new large-scale project of health insurance was expected to be substantial, and in 1913 Prudential established a completely new system to administer it.
Prudential’s chief executive officer, Joseph Burns, had attended James Powers’ demonstration of his machines in Berlin in 1913. He shared the Stationery Office’s enthusiasm for Powers’ printing tabulator, and Prudential acquired a Powers installation in 1914 to compile statistics. Initially, Prudential used punched cards to compile the obligatory statistics required by national health insurance and, in 1919, punched cards were also used to produce statistics of the company’s life insurance policies. Two years later, punched cards were successfully applied in the annual valuation of the company’s insurance policies.4
Meanwhile, Prudential used typewriters to produce the lists of policies for the frequent collection of industrial insurance premiums. In 1914, Frank P. Symmons of Prudential suggested to the British Powers agency that they could ease production these lists by using punched cards. This would require letter representation and a tabulator that could print the names of policyholders in addition to policy numbers and amounts to be collected. Charles Foster, an engineer working at the Powers agency, accepted the challenge and designed an alphabet printing unit for the standard numerical American produced tabulators. The alphabet printing unit was completed in 1916 and could print a reduced alphabet of twenty-three letters but not digits. The remaining three letters of the English alphabet were provided through double use of three letters.5 This alphabet printing unit could print the names of policyholders for internal use in the company, but it was not able to print full addresses, as these required a combination of letters and digits. This limited the alphabet printing unit’s usefulness in policy administration, but the unit was used to print specifications in bookkeeping projects elsewhere.
The British Powers company was a then simple agency, owned by the American company and selling and maintaining exclusively American-produced machines. Therefore, the alphabet printing unit would either have had to have been produced in the United States or the British company would have had to establish production of machines. As the American Powers company had no interest in an alphabet printing unit, the British decided to establish their own production.6
At the same time, the First World War was raging, making it precarious for the Prudential Assurance Company to rely on the continued import of office machine equipment, particularly from the Powers company, which was experiencing financial problems. Moreover, transatlantic transport was unreliable due to the First World War, which complicated acquisitions of new machines and spare parts from the factory in the United States. Therefore, Prudential decided that they needed to control the supply of punched-card equipment before they could extend their own use of punched cards.7
For these reasons, in 1918, the Prudential Assurance Company of London acquired the British manufacturing and sales rights for the machines from the American Powers company. They paid $90,000 (£20,000) for the rights to manufacture and sell Powers equipment throughout the British Empire, and they would neither have to pay royalties nor be bound by any pricing restrictions. Therefore, despite the hefty price, this arrangement was highly preferable to the 1908 agreement between the British Tabulating Machine Company and the Tabulating Machine Company in the United States.
The new British Powers company became closely tied to Prudential. It acquired a board of high-ranking Prudential officers, which lasted until 1945 when the Vickers company gained a substantial shareholding. In 1919, Frank P. Symmons became general manager of the British Powers company, a post he retained until 1945. He came from a position at Prudential and he continued his career in that company, becoming its chief executive officer in 1925.8
From the outset, the new British Powers company chose to use domestic machine production. Initially, they attempted to subcontract manufacturing, but probably due to the lack of experience in the industrial manufacture of office machinery in England, they were not able to make a satisfactory arrangement. Instead they built a factory of their own in Croydon, in 1920. This facility gradually took up full production of all punched-card machines, starting with the sterling currency attachment and the alphabetical printing unit in 1921 for the imported tabulators. The tabulator was the most complex punched-card machine and was the last to go into production, in about 1924.9 This made the British Powers company entirely independent of punched-card machines produced by the Powers company in the United States. The two companies remained in contact and exchanged patents and designs, but even so two distinct lines of punched-card machines evolved.
The American Powers company accepted the independence of the British company. However it twice tried to improve its position in the 1920s. First, immediately after the American Powers company had been reconstructed in 1922, the new management tried to renegotiate the territories specified in the agreement from 1918, but the British Powers company saw no reason to make any changes. As the British company by this time was approaching self-sufficiency in machine production, the American Powers company instructed its suppliers of printing ribbon and card stock to stop supplying the British company. The ribbons were easily replaced, but the British company was only able to find inferior card stocks. Despite this, they managed to run the machines with these stocks until relations with the American company improved.
This story was repeated when Remington Rand acquired the American Powers company in 1927. This time, the British company managed to get a British paper producer to produce cards of suitable quality. At the same time, Remington Rand tried to improve its position by selling its equipment within the British Empire, and the company began marketing in India, in breach of the 1918 agreement. The British Powers company threatened legal action, and Remington Rand withdrew.10 The two foiled attempts by the American Powers company to strengthen its position in relation to its former affiliate demonstrate the strong position attained by the British Powers company through the acquisition of the patent rights as opposed to a license.
However, though the alphabet printing unit had been a major reason for Prudential to invest in the British Powers company, it proved only a limited success in the company’s policy administration. In the 1920s, Prudential mechanized the administration of their “ordinary” life insurance policies by using a combination of Powers punched cards and address plates. “Ordinary” policies were for middle- and upper-class people for considerable insurance sums and in which premiums were collected between one and four times a year. The policy administration used a punched card containing all numerical information for every policy, which was used for internal operations in Prudential. In addition, an address plate for every policy existed, containing the policy number, the holder’s name and address, the insurance company agent, the total amount insured and the premium to be paid.
To ease identification in the administration, the policy punched card showed an imprint of the corresponding address plate. The address plate was utilized to generate renewal notices, receipts, and letters to the policyholder, while the punched card was used to value the company’s policy holding, compute bonuses, and print the lists needed to control the collection of premiums.11 It is noteworthy that the British Powers company’s reduced alphabet unit was neither applied nor developed for this application and that the dual system of punched cards and address plates remained in place until after the Second World War.12
The alphabet printing unit was used in business other than insurance companies in the 1930s to specify entries on invoices or ledger pages, while addressing was accomplished by the use of a separate address plate system. For example, a customer number ensured the customer received the correct invoice, as it was printed twice on the invoice, both from a punched card and from an address plate.13
In 1925 the British Powers company realized that their sales strategy, which was based on contacts in a few insurance companies, neither attracted many insurance companies nor new customers outside the field of insurance. They decided to improve the current sales and extend their scope to include both commercial and local public organizations. To this end, they outsourced their sales operations to Morland and Impey, who sold American Kalamazoo loose-leaf binders in Britain and who had successfully conducted negotiations with the American Powers company in 1918 on behalf of British Powers. Morland and Impey’s sales people had contacts to many bookkeeping departments and they drew on these when they started to sell Powers’ punched-card machines.14
Improved numerical capability of the British Powers company’s machines was essential if they were to attract customers from commercial and local public organizations and so, in 1926, a separate development department was established.15 Their initial ambition was to introduce the use of punched cards for various bookkeeping and accounting purposes, including costing and sales analysis. Punched cards were used in factories to compile payrolls and carry out wage analysis, to perform stock-taking, to produce purchase journals, and to monitor waste material.16 In other words, all applications requiring calculations or lists of separate letters and numbers printed on rolls of paper that the current Powers machines could produce.
To enhance their competitive edge over the machines from the British Tabulating Machine Company, the British Powers company improved the numerical capability of their machines in two significant ways: They introduced Y-wiring in their connection boxes, and they integrated automatic group control in their tabulator. These improvements ran parallel to the simultaneous machine development in the American Powers company; in fact, the British company’s introduction of automatic group control seems to have been based on the American company’s implementation of this feature.
Y-wiring in connection boxes was introduced in 1926 and facilitated more complex calculations. Connection boxes were interchangeable units composed of a rigid frame into which rows of wires were assembled that transmitted the movement from the hole sensed in the card through to the unit performing the calculation and to the printing units on the tabulator. Y-wiring enabled information to be transmitted from a perforation on a card to two different units on the tabulator, which was of great importance in preparing invoices and statements for which debit and credit values needed to be recorded in separate calculating units, because the tabulators at that time were unable to perform subtractions. In comparison, competing machines from the Tabulating Machine Company in the United States had, from the outset, the ability to transmit information from a perforation on a card to two or more calculation units because the information was transmitted electrically.
Automatic group control was introduced by the British Powers company in 1927, the same year as the American Powers company. Automatic group control ended the need for the time-consuming insertion and removal of total cards, which informed the tabulator where it was to print the total of certain amounts. The automatic group control mechanism allowed the tabulator to pick up the change of designation on the cards without action by the operator, to record the total at the end of each card group (for example a customer number), and to proceed to the next group of cards without interruption. The first automatic group control mechanism was subsequently improved so that it could control computing of both subtotals and grand totals.17
Tabulators with automatic group control had first been marketed in the United States in 1921 by the Tabulating Machine Company and were introduced on the British market in 1927 by the British Tabulating Machine Company. The basic Hollerith and Powers patents for this feature had been granted in 1916 and 1917 in Great Britain, where, in contrast to the situation in Germany, this did not give rise to litigation.
In 1929 the first paper-feeding system for the British Powers company’s machines was created, which allowed better control over printing. However, the tabulator still used rolls of blank paper and separate items, like invoices, were torn off the roll. The heading of an invoice was rubber stamped on the paper. Paper control was improved during the 1930s to enable the use of preprinted forms, although the addressing operation still relied on separate address plate systems.18
The strategy from 1925 of outsourcing sales was not a success. Much later, L. E. Brougham of the company’s engineering staff described Morland and Impey’s sales efforts as a “flop.”19 Selling loose-leaf binders was different from selling punched-card machines, and so special punched-card salesmen were introduced. It should be remembered, however, that the British Powers company’s machines only started to become well suited to bookkeeping applications through the improvements introduced between 1927 and 1929.
In 1929 a dedicated sales company, the Powers-Samas Accounting Machines Limited, was established owned jointly by the British Powers manufacturing company and Morland and Impey. SAMAS, the abbreviation of Société anonyme des machines à statistique (Statistics Machines Company Limited), was the agency for the British Powers machines in France that Morland and Impey had established earlier. The British Powers company bought a controlling interest in SAMAS in 1929. Powers-Samas became the brand-name of Powers’ machines in Britain, even though the Powers-Samas company was much smaller than the manufacturing company. Integration of the manufacturing and sales companies was achieved through product planning and development committees, an organizational structure that worked well. However, the two-company structure was confusing for outsiders, and after the Second World War the sales company was absorbed as a sales division of the manufacturing company.20
To improve sales, an American, Harold R. Russell, was hired in 1929 as the first general manager of Powers-Samas. Russell had worked in office machinery sales since 1910. He joined the American Powers company in 1919 and became its general sales manager after the company was reorganized in 1923. He returned to the United States in 1931 to become general manager of the Powers Division of Remington Rand. Russell’s appointment indicated the desire of British Powers-Samas to professionalize the company’s sales efforts. This he did, reorganizing sales entirely. The new organization was centered in London and had district offices in Birmingham, Manchester, Leeds, Glasgow, and Dublin.
Russell introduced selling on commission, he organized the sales force in such a way to ensure frequent calls to customers, and he advertised Powers-Samas machines in accountancy journals. In fact, prior to his appointment, there does not appear to have been any advertisements for punched-card machines in British journals. Finally, a company magazine, Powers-Samas Punch, was introduced. (It was renamed Powers-Samas Magazine in 1935.)
So far, competition between the various companies manufacturing and selling punched-card applications had been based on the practical capabilities of the machines to perform certain administrative tasks. However, IBM fundamentally changed the entire basis for competition in this field in 1928 when they introduced an 80-column card that was intended as a new proprietary standard to replace the existing industry standard of the 45-column card. There were essentially two ways to counter the new card: either to introduce another extended card or to diversify by offering cheaper punched-card systems with smaller cards.21
Remington Rand, the American Powers company, chose the option of an extended card which was created through a redefinition of the old standard 45-column card. They divided this card into two decks, which yielded ninety columns. The British Powers company could thus attain an extended card by adopting Remington Rand’s card and machines, or by following IBM’s lead and squeezing the forty-five columns on the existing card, or, finally, they could introduce a longer card. British Powers considered squeezing sixty columns onto the existing 45-column card, which should have been possible using the existing technology.22 Instead, however, they decided to diversify by offering a cheaper punched-card system with smaller cards to attract customers who could not afford to use a 45-column punched-card system.
The cheaper punched-card system was launched in 1932 as Powers-Four. It used smaller punched cards of 2 by 411/16 inches (5.1 × 11.9 cm) with 26 columns each of 11 punching positions, one less than on the 45-column and 80-column cards. The new card was only 39 percent of the size of a standard 45-column card but had 53 percent of its punching positions.23 Further, it had the same column spacing as the 45-column card, which minimized the amount of machine development required. Both cards and machines cost only about half the price of the 45-column cards and machines.
The lower price of the machines was achieved by a reduced number of units and a simpler design. Powers-Four offered four units, instead of up to seven units on the 45-column card machines. The small card made the machines physically smaller, they had fewer components and operated at a lower speed. Initially a range of only three machines was built: a hand punch, a sorter, and a numerical tabulator.24
The success of Powers-Four led the company to broaden its scope through additional machines and an alphanumeric capability. Between 1932 and 1935 the company added an automatic key punch, an implementation of automatic group control, and an alphanumeric tabulator with a reduced alphabet, and subsequently a full range of ancillary machines followed. The performance of the alphanumeric tabulator was constrained by the card only containing eleven punching positions. It had only twenty-nine characters, ten digits and nineteen letters, and each remaining letter was substituted by a digit or another letter.
This tabulator had fewer letters than the alphabetic unit for the 45-column tabulator, but it was alphanumeric, which meant it could print either a letter or a digit in all its printing positions. The alphabet capability was improved in 1935 to encompass twenty-one letters in addition to the ten digits.25 Powers-Four’s alphanumeric capability was marketed specifically for entries on invoices or ledger pages, just like the alphabet printing unit for the 45-column cards; a further similarity was that Powers-Four recommended a separate address plate system for the addressing operation.26
The success of Powers-Four for the company was not unqualified, however. Powers-Four was launched in 1932 when British Powers was still suffering from the effects of the 1929 economic crisis. The company only managed to raise factory output in 1933, and these are the only figures handed down as a record of the turnover. In 1933, doubts arose in the company surrounding the choice of marketing small cards. It was feared that Powers-Four would have a negative impact on the sales of 45-column equipment, which was already under pressure from the competitor’s 80-column card equipment. Some at Powers-Samas were even afraid that Powers-Four would become the main product of the company. The higher sales volume of the small machines could only partly offset their lower profitability. It was therefore decided to condense the columns on the 45-column card to restore the competitive position of the company’s big machines.27 The columns on a punched card could be condensed without infringing IBM’s 80-column card patent, as this only covered rectangular or oblong perforations.28 Smaller, round perforations worked well for the Powers company’s pin-sensing method.
The process of condensing the columns of the British Powers company’s cards was achieved in two phases. First, in 1936, they put sixty-five columns on the original 45-column card, which was accomplished by squeezing the columns together and using smaller perforations. Second, the card was made to contain eighty columns in 1954, twenty-six years later than IBM’s card. The number of columns on the Powers-Four card was similarly increased in two phases, from twenty-six to thirty-six columns in 1936, and then to forty columns in 1950.29
There was an important distinction between IBM’s conversion from forty-five to eighty columns and British Powers’ change from forty-five to sixty-five to eighty columns. While the IBM machines for the new card could be plugged to handle the old cards, Powers’ mechanical technology was less flexible. British Powers’ customers had to replace all their old machines each time the card changed. To ease the transitions, both Powers and IBM supplied machines that transferred information from the old to the new punched-card formats.
In the mid-1930s, the British Powers company was not only working toward increased representation on their cards, but they also designed a punched-card system, launched in 1936, that was the smallest and cheapest punched-card system on the market. It used a 21-column card, 2 by 2¾ inches (5.1 × 7.0 cm) with eleven punching positions. The tabulator was and remained a standard numerical machine with one calculating unit that operated at a modest speed. Powers-One was originally developed for the Cooperative Wholesale Societies to compute dividends for their members, with the expectation that the system would spread to other organizations that were unable to justify the greater cost of punched-card equipment using larger cards.30
The new machines improved British Powers’ competitive position in relation to the British Tabulating Machine Company, and the two smaller sets of punched-card equipment facilitated cheaper, though simpler, punched-card processing. The enhanced competitiveness was reflected in the company’s rising factory output until 1937, but output was lower in 1938 and 1939. Output figures are not available from the British Tabulating Machine Company, but the company had steadily growing assets from 1932, including 1938 and 1939, which indicate improved performance. In 1938 and 1939 British Powers produced approximately the same number of Powers-Four and standard punched-card size machines, while the number of Powers-One machines was approximately half as large, which shows the limited success of Powers-One. The falling machine production indicated a company in crisis, while the distribution of the company’s production between the three different machine lines showed that the forebodings of some employees’ in 1933 had been realized. Measured in number of machines, the company had primarily become a producer of small-size machines; any advantages gained from this did not, however, offset the disadvantages facing the Powers-Samas’ 65-column card in the competition with the 80-column punched cards from British Tabulating Machine Company.
Punched-card production at the Powers-Samas Company in London, 1937. Production of the cards themselves contributed up to a third of the punched-card producers’ revenues. (Powers-Samas Magazine October 1937, 7)
Problems of raising capital had been a major reason for the sluggish start of the British Tabulating Machine Company (BTM) from 1902 and until 1908, and the company had financial problems with the Tabulating Machine Company in the United States until the early 1920s. In 1912, the company fell behind in its royalty payments, and the American parent company pressured for payments. This was repeated in 1916 when the British company raised the question of whether they were to pay royalty on the part of the revenues that they used to pay the company tax, which had been imposed on British industry due to the First World War; according to the 1908 agreement, the 25 percent royalty was to be based on the gross revenue.
At that time, the British company tried to get the American company to lower the royalty percentage and take the royalty debt as shares. However, the Americans were unwilling to accept shares instead of royalty payments, as they were only interested in a controlling interest in the British agency. Only first in 1919 could the British Tabulating Machine Company sell additional stocks, which, finally, enabled the agency to pay off the debt to the American company.31
The British Tabulating Machine Company had established a workshop in London back in 1912 to assemble machines out of parts supplied from the United States. The growth in business during the First World War turned this facility into a bottleneck, and the company decided to build a factory for their assembly—similar to the one at the British Powers company. The British Tabulating Machine Company built a factory in Letchworth, near London, which opened in 1921. During the 1920s, their main job was to assemble pieces imported from the mother company in the United States. By and large they sold machines identical to IBM’s but for adaptations to sterling currency.32 However, the introduction of products in Britain lagged behind a few years, for example the number-printing tabulator was introduced in Britain in 1924, compared with 1921 in the United States.33
Assembly in Britain made the British Tabulating Machine Company’s operations less transparent for the Tabulating Machine Company in New York. In 1922, the American company questioned the British company’s calculation of royalties, as goods supplied by the American company did not agree with the royalties remitted. The American company demanded detailed royalty statements and threatened to take over the British company by buying up shares, a threat which was credible in light of the takeover of its German affiliate the previous year. Dishonest accounting would entitle the American company to void the contract, which would force the British Tabulating Machine Company out of business, as it only assembled machines, except for the attachment designed to handle sterling currency.
A report was prepared by auditors from the American company who found a deficit of $183,000 (£40,000), 28 percent of the British company’s assets in 1922. During the subsequent negotiations, the British Tabulating Machine Company raised the problem of the fluctuating prices based on currency exchange rates, which was a contrast to the fixed rate that had been the basis for the agreement between the two companies since 1908. In the agreement reached in 1923, the British Tabulating Machine Company only had to pay 10 percent of their deficits, and they were granted the freedom to decide on the cost of rentals. Further, they no longer had to pay royalties on the devices and machines which they designed and produced themselves.34 At that time, this arrangement was only relevant for the tabulator modification to handle sterling currency, but it made machine development advantageous for the British company, just as it was for the German subsidiary of the American company, Deutsche Hollerith Maschinen Gesellschaft mit beschränker Haftung, or Dehomag.
To exploit this enhanced freedom, the British Tabulating Machine Company formed a development department in 1923. Its first accomplishment, in 1926, was a “pence translator.” The new numeric printing tabulator also required reasonable printing of amounts in sterling currency, which required an improvement of the existing modification to handle sterling.35 Apart from converting machines for sterling, very little technical development took place at the British Tabulating Machine Company during the 1920s, and the important improvements, such as numeric printing and subtracting tabulators, came from the United States. The subtracting tabulator greatly widened the company’s scope for application in bookkeeping.36 However, as late as 1924, the company only envisioned the application of punched cards for statistics processing.37
Originally, the British Tabulating Machine Company’s sales operations were exclusively based on the managers’ contacts and word of mouth, like Hollerith’s original sales promotion in the United States. In 1922 the British Tabulating Machine Company improved their sales promotion. They held the first training course for service engineers that year, and they established a branch office in Birmingham, followed by offices in Manchester (1923), Glasgow (1924), and Leeds (1930).38 In 1936 the company started to publish a regular sales bulletin, The Tabulator.
The company’s sales organization evolved more gradually and systematically than that of the British Powers company. A historian explained this by observing that several of the company’s directors were familiar with selling capital goods when they joined the company.39 Additional inspiration came from their frequent contact with IBM’s efficient sales organization.
During the economic crisis that started in 1929, protective barriers sprang up in Britain as they did in France, Germany, and most other countries. In 1932 the government introduced general tariffs, which for punched-card machines amounted to 25 percent of their value and 10 percent for punched cards. As British Powers produced almost all the machines themselves, they were less affected than the British Tabulating Machine Company, whose business was based on machines produced in the United States. Although the British Tabulating Machine Company assembled about 70 percent of the machines, which reduced their duties, duties caused an 8 percent rise in costs, giving British Powers a competitive advantage.40
This advantage was furthered by the growing national hostility to non-British manufacturing expressed in the “Buy British” campaign of populist newspapers like The Daily Mail beginning in 1930. British Powers chose to capitalize on this trend and advertised that their machines were completely manufactured with British material, with skilled British labor and were backed entirely by British capital. This put pressure on the British Tabulating Machine Company as an importer of machines produced in the United States. Their brilliant counteraction was to appoint to their board of directors Conservative politician Leo S. Amery, who was a leader of the “Buy British” campaign.41 Thus, in spite of the new tariffs and the “Buy British” campaign, the British Tabulating Machine Company experienced steadily growing assets from 1932.
Perhaps ironically, these impediments probably contributed to the British Tabulating Machine Company embarking on its first major building of a new subtracting tabulator, as the contemporary restrictions in Germany encouraged the German subsidiary to build its own tabulator. In Britain, the company’s stated objective was to reduce their royalty payments. Since the late 1920s, the development department had established the basis for this endeavor through inventions by Harold Hall Keen. Their new tabulator from 1933 was based on 80-column punched cards and it could subtract and print numbers. After some initial problems, this machine became the British Tabulating Machine Company’s standard tabulator, and they no longer marketed the IBM-designed tabulators.42
Although this new tabulator was exclusively numeric and intended for statistical tasks, both IBM in the United States and the competing British Powers company saw bookkeeping as their main field of application at this time. To accommodate bookkeeping, an improved paper-control device was designed and produced for the new tabulator to facilitate printing on forms, and an alphanumeric version of the tabulator was produced in 1935. This British-designed machine was supplemented in 1939 by a multiplying punch that could handle sterling currency. The British Tabulating Machine Company also produced its own hand punch, hand verifier, and sorter. However, the other new machines, such as smart punches, reproducing punches, and collators were all IBM products.43
As stated earlier, the British Tabulating Machine Company’s main objective in developing their own tabulators had been to reduce royalty payments. Their claims were based on three of Harold Hall Keens’ patents, but IBM found that they all incorporated inventions that they already had sole ownership of. The outcome was a “proportionate” reduction in the royalties for the Keen claims IBM found to be novel.44
On the British market, the promotional newsletter from the British Tabulating Machine Company indicated that they envisioned their tabulator’s alphanumeric capability nearly exclusively to be used to write specifications, for example, on invoices.45 However, punched cards were used for addressing in a few of the reviewed applications, while the alternatives were to use typewriters or address plates.46 This observation that specifications were the main use of the British Tabulating Machine Company’s alphanumeric capability can also explain why the company in 1936 introduced smaller cards and machines to compete with the small card machines from the British Powers company. Both companies seem to have thought that they would attain more business in Britain using these cheaper formats than using alphanumeric applications of their bigger machines, a fact that is confirmed by the record of punched-card applications in British insurance companies mentioned earlier.
The British Tabulating Machine Company introduced a half-size 38-column card measuring 311/16 inches by 3¼ inches (8.3 × 9.4 cm). The shorter card enabled them to halve many of the internal components of the machine, while at the same time still be able to apply many parts from the 80-column card machines. Subsequently, the British Tabulating Machine Company introduced similar 24-column and 60-column cards, but none of these short-card formats ever caught on.47
In contrast, the British Tabulating Machine Company was unparalleled when it came to large-card machines, which Powers could not rival. In fact, the markets of the two companies were starting to diverge. The British Tabulating Machine Company went for the high-end market for their 80-column equipment, while Powers focused on the cheaper small-card machines.
Germany: Numeric Punched-Card Bookkeeping
The Tabulating Machine Company’s agency, Dehomag, successfully introduced punched cards in Germany in 1910. The use of punched cards grew rapidly for public and business statistics in the few years up until the First World War, and Dehomag’s business soared during the war.
The German postwar crisis resulted in a runaway inflation that was halted by both the introduction of the new mark in November 1923 and also the first mutually agreed arrangement for Germany’s payment of reparations the following year. Following this, Germany experienced substantial economic growth until the world economic crisis in 1929. This growth caused the relative cost of labor to rise and a major wave of rationalization swept German industry.48
In the postwar economic crisis lasting to 1923, many punched-card users gave up their rented equipment. But once the German mark stabilized, punched-card applications once again grew rapidly both in private companies and public organizations. First, punched cards regained their function of compiling public statistics and operational statistics in big companies. Then punched card-based operational statistics spread to the public sector and the cards started to be used for bookkeeping tasks. In the public sector, traditional statistics punched-card applications returned and new ones emerged between 1923 and 1927. The German National Statistics Department (Statistisches Reichsamt) in Berlin resumed punched-card processing of their foreign trade statistics in 1923, and early the next year the National Bank (Reichsbank) used punched cards to improve their surveillance of the foreign debt and assets, crucial to stabilize the new mark after 1923. The use of punched cards facilitated more detailed and quickly available statistics than did traditional methods.49
In addition, public institutions started applying punched cards to processing operational statistics, as had been the case in German industry since 1911. The German national railroads and the city of Frankfurt am Main exemplify the dynamics of this application. The German Reich was a federation of many states and in many respects had a decentralized structure. The organization of the railroads illustrates this. They were not a nationwide company, but they were organized through eight regional railroad systems operated by different states.50 By the end of the First World War, the Versailles treaty caused large railroad sections to be surrendered, and the regional railroad systems merged, forming the German National Railroad (Deutsche Reichsbahn) in 1920.51
According to the 1924 agreement between Germany and the Allies on reparations, the national railroads had to provide a third of the reparations through the amortization of nearly half the national railroad’s estimated total assets.52 Therefore, reparations became a decisive motivation for extensive rationalization during the following several years to achieve the advantages of scale in the amalgamated company. To this end, an important tool was extensive punched-card-based operational statistics. Further, within a few years, the operational statistics applications were extended to monitor line construction and train building to minimize stocks and costs. Finally, this was supplemented in 1928 with the introduction of punched-card-based bookkeeping for line construction expenses.53
Similarly, the city of Frankfurt am Main acquired punched cards in 1927 to process their municipal statistics. Then in 1929, they started to apply punched cards for expense control, which facilitated daily surveys of payments for various purposes by the use of numeric lists of expenses.54 Punched-card-aided bookkeeping was starting to emerge.
Energy supply, like income tax and real estate tax, brought a municipality in direct contact with a large part of the population. Municipal energy suppliers introduced punched cards to make various lists on consumption and to monitor the expenses of the utilities, for example, in Dresden (1925), Berlin (1929), and Frankfurt am Main (1930). While these applications only required addition, punched-card-based calculation of the amounts to be paid using the meter readings was a more complex task, the solution to which only emerged in 1935.55
German industry provided a parallel story. They too experienced a rapid expansion of operational statistics in the late 1920s, after which punched card-aided bookkeeping emerged. Between 1910 and 1912 the chemical and electrical industries were the first to introduce punched cards, which they used for various kinds of operational statistics. Three big ironworks introduced punched cards for operational statistics in 1920, three additional steelworks followed in 1926, and the Friedrich Krupp concern the following year.56
The extension of the use of punched cards in big companies to bookkeeping started in the late 1920s and can be exemplified by the development at the Siemens concern. After two years of preparations, Siemens established two punched-card installations, one in Berlin with Powers equipment, and the other in Nuremberg using Dehomag equipment.57 The Dehomag installation in the Siemens factory in Nuremberg was established in 1928 for internal accounting of materials in two production units. In the early 1930s, the application was extended to wage administration and wage statistics and, by 1937, an exclusively numeric punched-card register of workers, including their marital status and number of children, which influenced the calculation of wages, was operationalized. Now all Dehomag punched-card handling was performed in a centralized office, which, in 1942, had seven members of management and 158 additional employees.58
The Siemens factory in Berlin introduced Powers punched-card machines in 1927 for wage administration, including wage calculations, the handling of deductions like health insurance, and the printing of records of payment of wages and health insurance contributions. In 1928 tasks at this Siemens factory were extended to encompass operational statistics, similar to the Nuremberg factory.59
In a textbook on punched cards from 1929, Robert Feindler argued for the use of punched cards for wage calculation and wage control. However, the lack of multiplication capability in the punched-card machines impeded wage calculations being done with punched cards using the time rates and the number of hours. Further, the absence of subtraction prevented punched card-based calculation of the payable wage by deductions from the gross wage. Therefore, Feindler proposed that each worker’s deductions be added separately and to subtract the total amount from the gross wage either manually or by the use of a key calculation machine. Feindler suggested lists of wage components and deductions as output.60
The German punched-card companies, Dehomag and the German Powers company, pursued different roads to bookkeeping. Following the stabilization of the mark, economic growth in Germany was reflected in Dehomag’s rising turnover, which already in 1925 in dollars was 4.7 times that in the peak year of 1918–1919. This growth accelerated during the rationalization wave in the second half of the 1920s and became the basis for renewed machine development and production consented to by IBM as majority shareholder. The company also established branches in major German cities: Frankfurt am Main (1925), Hamburg (1925), Stuttgart (1926), Düsseldorf (1926), Dresden (1927), Leipzig (1929), and Munich (1930).61 Dehomag, which already in 1924 had bought a machine works factory in Sindelfingen in the industrialized Stuttgart area, moved machine repair and spare part production from Villingen.62
In the ten years following 1925, Dehomag’s main development endeavor was to improve the calculating capacity of punched-card machines, which reveals their perception of the role of punched cards in bookkeeping. First, they improved balancing and later features for the calculation of interests. The two main figures in this endeavor were Hermann Adalbert Weinlich, the company’s chief engineer since its inception in 1910, and Ulrich Kölm, an engineer who joined the company in 1925.63 They built numerical punched-card machines with calculation capabilities that were unsurpassed elsewhere. The prime users of this shaping of punched-card equipment seem to have been the banks.
Dehomag had seen the potential of the bank market as early as 1914, particularly at the large and internationally recognized Grossbanken, which they approached. Dehomag envisioned punched cards being applied to deposit administration, transaction of payments, credit administration, and possibly current accounts (Kontokorrent).64 However, these applications were initially held back by their tabulator’s lack of a printing capability but eventually emerged when improved tabulators included this feature.
In 1923, Dehomag started to import number printing tabulators from the United States and subsequently secured several banks as customers, who primarily used their installations for current account administration. Short-term loans through overdrafts on current accounts were a prime way to supply working capital for German industry, and many German banks had a large number of current accounts. Further, many current accounts had frequent and large entries, which made it crucial to establish effective monitoring of transactions to ensure that each costumer and bank would be able to meet their outstanding debts.
Using a separate punched-card for each transaction, the numeric printing and adding tabulators of the mid-1920s facilitated printing of sorted and summed up lists of the activities on every current account, which was cumbersome and time-consuming to compile by other means. However, the absence of subtracting tabulators hampered the appeal of punched cards for current account administration, as balancing was crucial due to frequent mixed debit and credit movements. Further, calculating interest on current accounts was complex due to the frequent activities on these accounts, but punched-card processing of this task would require a punched-card multiplier. Consequently, punched-card applications for current account administration called for developing features for subtraction and multiplication.65 Current account administrations that included the use of punched cards were established at Vaterländische Bank in Berlin in 1924, at Provinzialbank Pommern in 1925, and at Dresdner Bank in the late 1920s.66
The then-available account applications showed the potential of three possible punched-card machine improvements, namely subtraction, interest calculation, and addressing statements to the customers. Due to the absence of subtraction, the calculation of balances was awkward. Credits and debits were recorded in separate fields on the punched cards and balances were established by use of complementary figures, making it difficult to decide whether a balance was credit or debit. The punched-card companies subsequently addressed these shortcomings, but Dehomag distinguished itself by being the company that most stubbornly concentrated on improving the calculation capabilities, while disregarding alphabetic features.
In 1926, the engineers at Dehomag solved the balancing problem by modifying an IBM tabulator produced in the United States to provide this capability, and in 1928 IBM started to produce a tabulator that included subtraction capabilities. However, this modification came with the major disadvantage that negative numbers needed to be punched as complement figures, which was demanding for the operator and bred errors.67 IBM marketed a tabulator with a similar subtraction capability in 1928 and a full subtracting tabulator in 1933.68
The success in 1926 of Dehomag’s development team in modifying an IBM tabulator to perform subtractions became the basis for their subsequently building tabulators with high computing capability and forms control system, which was crucial for advanced printing on forms.69 In 1933, they completed an improved tabulator that did not require negative figures to be punched as complement figures. This meant that all numbers could be punched regularly; negative signs were indicated by using perforations in row 11. IBM marketed similar improvements the same year in the United States, fundamentally facilitating the use of punched cards for bookkeeping applications involving addition and subtraction. Dehomag produced 250 copies of these tabulators from 1933 to 1936, until the D11 tabulator replaced them.70
Following this, the tabulator was improved to also multiply. In 1934, Dehomag completed a version that could multiply by up to three digits, which, for example, could calculate interest on current accounts in a bank.71 By the following year, this line of development was completed by Dehomag’s D11 numeric tabulator, which could multiply, divide, and perform complex calculations. This tabulator was produced in Germany from 1935 to 1960, with a total of 1,120 machines supplied by 1943.72 The absence of letter printing during this same period is striking. Until the outbreak of the Second World War, little interest in letter printing using punched cards appeared in Germany.
The German Powers agency had the disadvantage in developing its road to bookkeeping that its initial start-up had been foiled by Dehomag’s patent infringement suit in 1914, the First World War, and the subsequent runaway inflation. Thus, the Powers agency reestablished itself from scratch in 1923 in Berlin. Jan Büchter headed the agency, and it retained the former name, Deutsche Gesellschaft für Addier- und Sortiermaschinen mit beschränkter Haftung (German Company for Adding and Sorting Machines with limited liability). Their business, based on machines imported from the United States and a card printing shop, marked the first permanent presence of the American Powers company on the European continent. They made the German agency the center of their continental business.73
Dehomag had established a robust first-mover position in Germany because of their business since 1910 and because the Powers agency had been prevented from operating in the country between 1914 and 1923. Jan Büchter now countered his inferior position through an offensive strategy in marketing and machine development that illuminated his visions for new punched-card application fields. Büchter’s marketing strategy focused on the Grossbanken and insurance companies. The main technical asset the Powers agency had compared with the IBM machines was the British Powers company’s tabulator with a reduced alphabet of twenty-three letters.
The German Powers agency, which implemented the alphabet reduction differently from the British company, had no feature for four national German characters (ä, ö, ü, and ß).74 However, in spite of this shortcoming, the German Powers agency argued in 1926 that punched cards be applied to wage administration systems that encompassed printing lists with the worker names for internal use in the company. Printing worker names for the workers’ use required a more complete alphabet.75 Moreover, the German Powers company tried to improve its position through machine development.76 These actions indicated an independent position for the German Powers agency during the American parent company’s weakness in the 1920s.
Remington Rand’s acquisition of the Powers company in the United States in 1927 caused changes in Germany. Remington Rand maintained a strategy of conducting marketing abroad through its subsidiary companies, which distinguished it from its punched-card predecessor. Therefore, they took over the German agency and established a child company, Powers Gesellschaft mit beschränkter Haftung (Powers company limited) with new management.77 During the next several years, the German Powers company marketed machines produced in the United States and printed punched cards in Berlin, while terminating their own machine development efforts. As in the United States, the German Powers company introduced the double-deck 90-column punched card in 1929, and began marketing the new Remington Rand range of machines (Model 2) intended for the new punched card.78
It is difficult to assess the German Powers company in the 1920s, as no turnover figures survived from this period. Company revenues in 1940 only amounted to 13 percent of Dehomag’s revenues that year, implying a modest-sized turnover.79 However, Dehomag launched a second patent infringement suit against the German Powers company in 1924, indicating that they saw the German Powers company as a significant opponent. This patent infringement suit was in reference to Herman Hollerith’s and James Powers’ automatic group control patents. Both patents were filed and granted in Germany, Powers’ patent in 1921 and Hollerith’s patent in 1924, but Dehomag contested the validity of Powers’ patent in German courts. This dispute was only solved in 1929 when the Supreme Court ruled that Powers’ patent was valid.80
During the world economic crisis starting in 1929, the German right-wing governments of the early 1930s chose to concentrate on the balance of payments by introducing government import controls. This contributed to a fall in imported industrial goods, in particular, with imports between 1929 and 1932 being reduced in value by 45 percent.81 When the Nazis assumed power in January 1933, they promised full employment through the creation of new jobs and advocated for “self-sufficiency” and “independence from the World Economy,” to be implemented by strengthening the previous governments’ import restrictions, thus causing further reductions in the import of manufactured goods.82
These Nazi objectives made Dehomag vulnerable. Though the introduction of punched cards was frequently justified because it saved on manpower, nearly all of the company’s machines were imported, and the company was 90 percent foreign owned. However, Dehomag’s substantial machine development, which culminated in 1935 with the D11 numeric tabulator, improved the company’s position both in Germany and simultaneously in relation to IBM in the United States.
In the summer of 1933, Dehomag knew its work building machines would come to fruition, requiring it to have manufacturing capacity. The company bought a site in Berlin and built a new factory, which it opened in 1934 with the explicit consent of IBM. From 1935, this factory housed the highly successful D11 tabulator production, which became the major reason for Dehomag’s fast growing turnover during the second half of the 1930s. Also, the number of employees increased from 462 in 1933, to 1,100 three years later.83 This accomplishment enabled Dehomag to avoid conflict with the National Ministry for Trade and Industry (Reichswirtschaftsministerium), which controlled foreign trade, whereas Dehomag’s competitor, a Powers’ subsidiary, had to commence production in Germany using American-designed machines to avoid importing complete machines.
Cabling of D11 tabulators in the factory of IBM’s German subsidiary, Dehomag, in Berlin, 1935. This tabulator shows the size and technical complexity of large punched-card machines in the 1930s. (Festschrift zur 25-Jahresfeier der Deutschen Hollerith-Maschinen Gesellschaft. Berlin: Dehomag, 1935, 45)
Further, Dehomag’s management took the opportunity at the opening of their new Berlin factory in 1934, where there were representatives from the government and the Nazi party present, to voice their allegiance and enhance their connection to the new regime as well as stress the company’s “Germanness.” Director Willy Heidinger told the audience that, “We have firm trust in our physician and will follow his orders blindly, because we know that he will lead our nation towards a great future. Hail to our German people and their Führer.”84 In 1935, the company further emphasized its Germanness in a publication commemorating Dehomag’s twenty-fifth anniversary.85 The German origin of punched cards was asserted because of Herman Hollerith’s being the son of German immigrants, his first name (mis-)spelled with a double n, according to German tradition. He was referred to as a German-American, though he was born in the United States and had never lived in Europe. Finally, the company drew attention to its own development and production work. However, development and production did not carry enough status within the company for an engineer to gain a position among the executive managers before 1945. The delegation of authority to national leaders within IBM eased the requisite adaptation during the chauvinistic mood of the 1930s.
Dehomag’s second director, Hermann Rottke, also used the factory opening in 1934, attended by the Nazi party and Deutsche Arbeitsfront representatives, to claim that Dehomag punched-card systems created jobs, as people with Dehomag punched-card training easily found work in spite of the high unemployment nationwide. But, he avoided any discussion about whether punched-card machines saved on the number of jobs or not.86
The various sectors of German industry experienced different ramifications of the self-sufficiency and rearmament policy. While, for example, the production of shoes for the civilian population declined and caused shortages, industries of military importance, like the Siemens concern, experienced rapid growth.87 Dehomag experienced even bigger growth than Siemens. Between 1936 and 1942, in addition to a growth in turnover 19 percent larger than Siemens’, Dehomag experienced equivalent relative increases in its workforce. The company also increased its number of branches, spreading its distributor net across Germany. They opened branches in Bielefeld (1935), Breslau (1935), Nuremberg (1935), Saarbrücken (1935), Königsberg (1937), Bremen (1938), Dortmunt (1938), Essen (1938), Hannover (1938), Karlsruhe (1938), Cologne (1938), and Magdeburg (1938).88
Between 1933 and 1940, Dehomag’s high level of profit caused IBM yearly problems, because the repatriation of funds was not possible, and because Willy Heidinger, who was the sole German shareholder, wanted his dividends. Declaring the profit was delayed, and then, with the exception of payment to Heidinger, they were either reinvested or invested in property.89 Punched cards became increasingly important for the arms buildup and later for warfare.
Hitler’s assumption of power in 1933 signaled the start of the establishment of a totalitarian state. Punched-card-based control became a major sales argument for Dehomag, who supplied equipment for statistical monitoring of the members of the Nazi party, SA, SS, Hitler Jugend, and Bund Deutscher Mädeln.90 These applications were remarkable for their limited ambitions, compared with the simultaneous introduction in France and the United States of punched cards for managing records.
The German Powers company was much more vulnerable than Dehomag to the import restrictions of the German autarchy, as it was exclusively owned by Remington Rand in New York and was a much smaller company. Like Dehomag, the German Powers company voiced its support for the new age by offering their expertise for the construction of a scientific national planning system, but even that could not eliminate the company’s problems.91
Machine development in Germany had been terminated in 1927 when Remington Rand took over the agency, leaving the German Powers company two options: to import machines produced in the United States or to establish production in Germany of machines designed in the United States. The first option had already been selected before the economic crisis started, but tight German currency restrictions from the early 1930s curtailed this possibility, and production started in Germany of punches designed in the United States but not of sorters or tabulator. This eased the German subsidiary’s relations with the authorities for a while, and the company earned money in the rearmament. Profits, however, were invested in bonds or equities in Siemens and Halske, IG Farben, and other German companies because the authorities did not allow them to be exported.92
The choice of Siemens and Halske was hardly a coincidence, as they had used Powers’ punched cards for several years and were substantially developing punched-card equipment.93 Siemens and Halske included the low-voltage part of the Siemens group and produced telegraph, telex, and telephone equipment. It was a proliferating, innovative company with technology similar to Powers’ punched-card technology, especially Siemens’ alphanumeric telegraph transmitters. Siemens and Halske, who had acquired a Powers’ installation for wage administration in 1927, soon began developing their own punched-card machine.94 By 1934, they had designed their own punched-card code, a sorter, and a card reader. The Siemens card had ninety columns organized in two decks of four rows and was inspired by the Powers double-deck card even though that card had five rows in each deck.95
Further, Siemens utilized their expertise in the transmission of digital information. In 1934 they had completed a link between a key subtracting and multiplying machine and a punch that facilitated the punching onto cards without extra action, with information that had been keyed in and processed using an ordinary bookkeeping system. This feature enhanced the modest calculation capability of a standard tabulator and resembled a similar Dehomag construction from 1929. In addition, Siemens had built a machine that transferred information from punched strip tapes for telegraph transmitters to punched cards.
If Siemens wanted to market this technology, they had the choice of either cooperating with one of the existing punched-card producers or producing and marketing their own machines, as they had done for many other products when they diversified in the 1920s. Siemens possessed several punched-card patents and vast technological abilities, but they lacked key patents, like the automatic group control patent Dehomag and Powers had in Germany. In addition, the economic crisis caused falling revenues for Siemens from 1930 to 1933.96
In 1934 Siemens chose to establish a partnership with the German Powers company. Powers acquired the right to the use of the Siemens’ innovations, and Siemens would produce the machines they designed. However, the agreement exclusively concerned the machines designed by Siemens, while Powers would continue to import sorters and tabulators from the United States.97 For Powers, the alliance would prevent a new punched-card competitor from emerging, and Siemens was an attractive, prestigious partner. The alliance, which would ease the concerns of the National Ministry for Trade and Industry regarding imports, could not solve the problem that Powers continued to rely on complete tabulators imported from the United States.
Two Siemens punched-card machines appeared on the market in 1935, which consisted of a sorter that could handle regular 45-column cards, as well as double-deck Siemens and Powers cards, and a link between a key calculating machine and a punch.98 The innovation of a converter to change from punch strips to punched cards ran into trouble when no eager first customer appeared, grounding development in 1938. It was only revived due to the subsequent emergence of an eager customer, the army’s punched-card service (later called Maschinelle Berichtswesen).99
In 1934 the Siemens-Powers partnership envisioned two contradictory developments of the punched card. They planned to develop the Siemens punched-card either into a triple-deck 135-column card or to introduce alphabet representation. Siemens’ four-row number representation provided space for a third deck on the card, but four rows only allowed fifteen characters, whereas exclusive letter representation required at least twenty-six characters (five rows) and alphanumeric representation needed at least thirty-six characters (six rows).100 These plans were probably based on Siemens’ extensive expertise in alphanumeric telegraphic transmitter systems with paper strips.
Neither of these options was pursued by the German Powers company. They did not even adopt the Remington Rand 90-column alphanumeric system that was introduced in the United States in 1938. The German Powers company only marketed the old reduced alphabet printer for 45-column cards, designed by the British Powers company, which had been on the market since the 1920s.101 The German Powers company’s focus on bookkeeping remained the calculation capability of their punched-card systems, while letters were only used as an auxiliary for specifications.
During the second half of the 1930s, the German Powers company contracted with Siemens to produce punches and verifiers based on designs from the United States.102 Remington Rand, which only allowed Siemens to produce minor machines like punches and verifiers, wanted their sorters and tabulators to be imported from the United States. Siemens produced a sorter of their own design, but the production of tabulators became a problem. The Ministry for Trade and Industry pressured the German Powers company to produce all their machines in Germany. In 1935, Powers was permitted to import tabulators for another three years, while preparing for production in Germany. However, import was blocked in 1937 as the Ministry for Trade and Industry realized that Powers did not plan to move their tabulator production to Germany, forcing Powers to start their own tabulator production in Berlin. The American company would not allow Siemens to produce their machines, as they feared Siemens would use their expertise to enter the punched-card market when the Siemens-Powers partnership expired in 1949. Siemens, for their part, was eager to adopt tabulator production.103
France: Alphanumeric Accent in Punched-Card Bookkeeping
The first bookkeeping application of punched cards in France was security administration at the Bank of Alsace and Lorraine (Banque d’Alsace et de Lorraine) in Strasbourg in 1930.104 It was derived from equipment from the American Powers company, and the tasks involved monitoring holdings and frequently producing account statements for account holders with letter specifications of the various securities. Letter specifications in these reports were produced by the reduced alphabet printing unit from the American Powers company. Customer account statements were addressed by use of address plates.
By 1933, the Bank of Alsace and Lorraine had established an additional Powers punched-card installation for current account administration. As soon as the bank was informed of a transaction, it was recorded on a key bookkeeping machine that produced advice for the customer and carbon copies for use in the bank. On this basis, information of the transaction was punched onto a card, which was used daily to produce various sorted and consolidated lists for internal use in the bank. Some of these lists were sent as statements to the account holders. A special subtraction unit on the tabulator consolidated the entries for each department or account. Similar bookkeeping applications using punched cards were established at several French banks in the 1930s, making French banks a noticeable market segment.105
Another distinction of the punched-card trade in France was the emergence of the state as a user that contributed to shaping punched-card systems for bookkeeping in France. The public sector diffusion of punched cards started later than it did in the private business sector. The first installation in the public sector was only established in 1926, which grew to six installations in 1932.106 At that point, most punched-card machines were still operating in businesses, but in 1936, 60 percent of punched-card machines operated in the national administration.107 The introduction of punched cards in the national government was part of the process to reform government from the ground up, while attempts for a general reform failed.
The Third French Republic (1870–1940) suffered from basic constitutional weaknesses. It had a two-chamber parliament that faced severe problems due to the numerous, frequently changing parties that rendered it virtually inoperable. The French president, who was elected by parliament, held little power, the government’s position was seriously weakened by the many unstable parties, and by the fact that it was not empowered to respond to a vote of censure by dissolving parliament. These constitutional inefficiencies reinforced the volatile political development in France during the 1930s and contributed to the failure of a general reform. The outcome was the disintegration of the Third French Republic which, in 1940, transformed into the autocratic Vichy state.108
The state’s majority share of the punched-card business in the 1930s distinguished France from Germany and Great Britain, with the French state gaining a significant role in shaping punched-card systems. Simultaneously, the emergence of the Compagnie des machines Bull in France, independent of the IBM and Powers companies, made the competition in France unique in the 1930s.
The installations at the Bank of Alsace and Lorraine were modeled on American Powers equipment, which was marketed separately from British Powers equipment in France starting in the late 1920s. Furthermore, these installations were associated with plans to establish a production of punched-card equipment from the American Powers company in Hagenau, Alsace, an idea supported by two polytechniciens, Georges Vieillard and Elie Doury, engineers trained at the distinguished École polytechnique (Technical University) in Paris.109 However, this plan was not implemented, and Vieillard and Doury moved to assist in establishing the Bull company in France, while the American Powers company chose to merge their business into SAMAS (Société anonymes des machines à statistique, Statistics Machines Company limited), the French subsidiary of the British Powers company.
SAMAS, which imported foreign-produced machines, had neither production nor machine development in France, making the company vulnerable when foreign trade became complicated after the breakdown of the international gold standard in 1931.
The state came to control imports and, simultaneously, the French state became the most important punched-card customer. National preferences grew in public acquisitions and became evident during the Pierre Laval government from 1935 to 1936.110 Furthermore, the French government insisted on Frenchifying foreign-owned businesses in France, which meant having French ownership of a majority of the stocks, and a majority of French nationals on the board. To accommodate this request, Frenchmen took over a majority of the shares of SAMAS. To accomplish this, all the British directors on the board resigned and French nationals were appointed in their places. However, the British Powers company retained the right to dismiss the directors, which implied they had not relinquished control of the French company.111
SAMAS established their own production facilities to assemble foreign-produced parts and started machine development. However, they ran into financial problems due to the costs of production and development combined with the falling value of the French franc, which declined by 63 percent compared with sterling between 1935 and 1937, and by 61 percent compared with the dollar. By the end of 1937, this had caused SAMAS severe deficits, and the British Powers company and Remington Rand began negotiating a reconstruction. The outcome was that the British Powers company paid for the reconstruction of SAMAS in 1939, while Remington Rand terminated their business in France, selling their French interests to the British Powers company in 1939.112
IBM companies in France prospered in the 1920s under the rationalization of businesses. Their total profits rose nearly every year from 1922 to 1931. Net profits are the only figures available for the IBM companies in France, but the impression of progress is upheld by a rising number of customers. IBM in France had thirty customers in 1927, which grew to seventy-six customers in 1935 and to 102 customers in 1938. In addition, the company established branch offices in Lille (1923), Lyon (1924), Marseille (1928), Bordeaux (1932), Nantes (1935), Alger (1936), Strasbourg (1938), Casablanca (1938), and Toulouse (1940).113
In 1935, the French IBM companies were also pressured by the French government to Frenchify, just as SAMAS had. To accommodate this request, IBM’s total business was united in a single company in 1935 with the lengthy name Société française Hollerith, machines comptable et enregistreuse (French company for bookkeeping and registering Hollerith machines), which the following year was simplified to Compagnie électro-comptable de France (CEC, electric bookkeeping machine company of France). Although the economic control of the company in France remained with IBM in the United States, Frenchmen took over the board in France. Prior to that, there had only been one Frenchman on the boards of the two IBM companies in France. Now two-thirds of the members were French.114
National development and production became crucial components of IBM’s strategy in France after the IBM company got a board with a French majority. However, this strategy had been initiated much earlier. As early as 1924, IBM had acquired a factory proper at their Vincennes machine shop to establish production.115 In addition, they had started machine development in France in the late 1920s that seems to have focused on improving machines designed by IBM in the United States.116 The Vincennes workshop finished a verifier in 1931, which they subsequently produced. Further, they improved the calculating capability of an IBM tabulator in 1934 to meet the special needs of French railroads, and they finished an improved paper feed for tabulators in 1935.117
In 1935 the French IBM company concluded an agreement with the parent company in the United States to assemble a printing numeric tabulator and a sorter, in addition to the verifier they had produced since 1931. This decision was the result of severe problems with importing completed machines from the United States. Customs duty was substantial and customs officers were zealous. By 1937 most machines were produced in France, including IBM’s alphanumeric tabulators.118
The falling and low total net profits of IBM’s French subsidiaries between 1933 and 1940 indicated poor business for IBM in France, but these numbers were influenced by the prices in the trade between IBM and its subsidiaries as well as the fluctuating value of the franc compared with the dollar. Moreover, the dynamic actions of French subsidiaries indicated a different and more positive development. The competing Bull company did receive preferential treatment as a genuine French company. In 1935, the French army acquired Bull equipment to monitor army transport expenses, in spite of the recommendation from the office processing those expenses to purchase IBM machines. They considered the IBM machines technically superior.119 However, Bull remained a small company in the 1930s compared with the French IBM company, and the French IBM company had a dynamic development of punched-card applications in the French army, showing the importance of the army in the transition from statistics to bookkeeping applications in France and of IBM’s role in that transition.
René Carmille was the key person in the French army’s use of punched cards in the 1930s. He was trained in engineering at the elite École polytechnique in Paris and was employed in the French army after graduation. In 1924, he was assigned to the Corps du contrôle de l’administration de l’Armée de terre, where he rose to become the head in 1936.120 This corps was a tool for controlling the French army’s expenses and for promoting efficient administration.121
Back in 1931 or 1932, René Carmille became interested in using punched cards while studying ways to control costs in the army’s artillery workshops. As a polytechnicien —a graduate from École polytechnique—he was influenced by the stories of rationalization propagated by his fellow engineers stemming from their key roles in the rationalization of the industrial sector dating back to the 1920s. Still, the Bull company had not yet emerged on the French market, and in 1932, IBM machines were chosen for use in cost control at the national weapons factory in Puteaux, near Paris. Management control was to be improved by the use of frequent and regular budgetary statements.
Key numeric information on invoices received or wages to be paid for work to produce ammunition and weapons was punched on cards as these invoices or wage accounts were approved for payment. This produced statistics on expenditures before they were met, facilitating a more detailed breakdown of costs according to the various departments or projects than could be provided by the records on payments. During the 1930s, the Puteaux factory maintained its IBM installation while it was expanded to include additional applications and machines.122
Between 1934 and 1937, the army established similar punched-card installations for army operational statistics at the army’s explosives establishment in Sevran-Livry near Paris and in four other locations. In contrast to the Puteaux installation, the army’s punched-card installations were all modeled on equipment from the French Bull company. However, national preferences in public acquisitions were not mandatory. In 1935, René Carmille started to extend the scope of punched-card applications in Sevran-Livry. Carmille intended to move beyond operational statistics to include bookkeeping of the weekly payroll. Since the start of the first army installation at Puteaux in 1932, punched-card producers had improved the capability of their machines. Most important, the machines had become alphanumeric—they now included a full alphabet and printed combinations of numbers and letters.
The new application at Sevran-Livry was based on the IBM punched-card standard and required some IBM machines to complement the original Bull installation.123 The Sevran-Livry application involved a bookkeeping system that allowed machines to add together various wage components, to subtract any deductions, and, of particular interest to the army, to print lists with names and various figures, as well as a receipt with the wage earner’s name and the amount to be paid.124 The IBM machines were selected, in spite of being American, because it would have taken an additional year before Bull could supply a subtracting tabulator.125
In 1936, Carmille described in a book his conception of punched-card applications as demonstrated at the Sevran-Livry works. This conception encompassed general statistics and bookkeeping, and used letters to improve the readability of various items, for example, for goods listed on inventories or the names of employees and wages owed to them. By and large, until this time punched cards had remained a processing tool and the cards were discarded when processing ended. However, in the Sevran-Livry wage administration application, each worker had a basic punched-card that was kept and used each time wages were paid. Each worker’s basic card held his name (up to thirteen letters), wage number, numeric wage, and the tax and insurance deduction information used to calculate the wage to be paid. The basic cards comprised a register, though this only constituted an auxiliary function to the system, which each week consumed four additional cards per worker that were subsequently discarded.126
The full alphabet found on the IBM tabulators was crucial to print receipts with the wage earners’ names, which distinguished this from similar applications in Germany. In 1935, the French army also began developing the idea of a register with alphanumeric information for use in conscription and mobilization. A substantial register for this purpose would be an attractive and prestigious application that all three punched-card companies on the French market tried to attain. However, the order for this application was only decided after the fall of France to the German invasion in 1940 and was awarded to the Bull company.
The original Bull punched-card machines were designed in Norway by engineer Fredrik Rosing Bull and built by a local precision maker. Bull died in 1924, but his machine building was continued by engineer Knut Andreas Knutsen in Norway. In 1927 and 1930, Swiss-Belgian interests acquired the Norwegian patents and expertise, which only encompassed a punch, a sorter, a nonprinting tabulator, and the design for a number-printing tabulator.127 Building of these machines relocated to Paris in 1931, where a new company, Egli-Bull, was established for this purpose. The Egli name came from the majority share holder, the H. W. Egli adding machine producer in Zürich. Production was established in an existing workshop in Paris, which changed from manufacturing American office machines to producing Bull machines. The new company had a share capital of $140,000 (3.6 million francs). However only one-third was paid, which provided very low funds for starting production of the machines built in Norway and for designing new machines, as the remainder was bonus shares awarded for the patent rights and the workshop facilities.128
The new company did not have an easy start. The transfer of technology from Norway to France proved complicated, demand was curtailed by the economic crises, and the new company had to design and produce additional punched-card machines to become a full competitor to IBM and SAMAS. In spite of these problems, the Bull company during its first year succeeded in building and supplying their first major punched-card machine to the French Ministry of Labor. However, this accomplishment exhausted the company’s economic resources and they ended up with a significant deficit. By the end of 1931, the company needed about $94,000 (2.4 million francs) in additional share capital to alleviate their current economic problems and to complete building a competitive set of punched-card machines; in dollars this amount was 2.4 times the originally paid share capital.
To aggravate matters, the Swiss majority shareholder, H. W. Egli adding machine producer in Zürich, now needed to sell some of its shares due to falling demand caused by the world economic crisis.129 The need for additional capital to the small Egli-Bull company and financing for further development of punched-card machines could be solved either through a takeover by a foreign producer or through the acquisition of French capital combined with further machine development within the Bull company. Both possibilities emerged successively in 1931.
Back in April 1931, Remington Rand had approached the main shareholder in the burgeoning Egli-Bull company, the H. W. Egli Company in Zürich, to buy the rights to Bull machines outside France. Remington Rand had acquired the only competitor to IBM in the United States in 1927, and they now needed to improve the performance of their machines. In the summer of 1931, directors from Remington Rand visited the workshops in Paris.130 During these negotiations, which lasted until December 1931, Remington Rand became aware that Egli-Bull was seeking capital, and they offered to invest $230,000 (5.9 million francs), implying that Remington Rand would take over the company. In addition, Egli-Bull would be allowed access to Remington Rand’s technology, though the advantages of this would be curtailed due to their different basic technologies.131
As the negotiations between Egli-Bull and Remington Rand were approaching an agreement, Georges Vieillard and Elie Doury emerged with an alternative. For several years, they had been working on a failed project to establish the production of Remington Rand punched-card machines in France. Now, they offered to promptly raise the needed capital in France for the Bull company.132
While Remington Rand’s offer was backed by a big company, Vieillard and Doury rapidly needed to raise the required capital from a group of punched-card users they were trying to organize. The most urgent concern was to raise $24,000 (francs 600,000) to acquire shares from the Swiss H. W. Egli company in order to relieve that company’s financial situation. In early December 1931, Georges Vieillard established a consortium of punched-card users in France who were willing to provide the amount required to buy the H. W. Egli shares in Switzerland and finance the much-needed expansion of Egli-Bull’s capital. The consortium consisted of various industrial producers, railroads, and insurance companies. Prominent among them was the Société des Papeteries Aussedat, a manufacturer of paper and cardboard, which later became a large punched-card producer.
The consortium of users offered the Egli-Bull company a second choice, and it chose this French option. In April 1932, the capital increase was achieved, allowing the French to attain a majority of the company’s share capital, and the Egli company in Zürich received the funds they needed.133 However, the French rescue meant no technological alliance with Remington Rand as required by that firm if the French were to become the majority shareholder. Therefore, the Paris company had to accomplish the much-needed development of their machines alone.
Again during 1932, development costs exceeded revenues, which once more were relieved by a substantial investment by the consortium of punched-card users in France in early 1933. The same year, the company was renamed Compagnie des Machines Bull (CMB).134
The increases in share capital in 1932 and 1933 reflected the young company’s problems in producing reliable machines and in developing a full line of punched-card machines to compete with IBM. Only in 1935 did Bull accomplish producing a basic line of punched-card machines, but the company’s financial situation was aggravated as a large part of its trade was leasing machines, which required more capital than selling. The outcome resulted in deficits throughout 1935, making the need to raise additional capital urgent. Simultaneously, the company was bolstered by a growing number of customers and increasing orders.135
Meanwhile, the Bull company had for several years been working to alleviate their problems and build up exports. From 1932 to 1934, they had tried in vain to find a partner in Great Britain, Italy, Czechoslovakia, the Soviet Union, and the United States. In all of these negotiations, it was a condition that two-thirds of the share capital should remain with French and Belgian shareholders.136 In 1935, Bull’s financial problems grew serious, and the company initially tried to obtain aid from the French state to support research and development, but the state hesitated. Then, Bull turned to find a partner in the United States, where they approached both Remington Rand and IBM. The chairman of IBM, Thomas J. Watson, came to Paris where he met with Jacques Cailles and Georges Vieillard of Bull’s board of directors and offered them an agreement of cooperation on the condition that IBM took over the Bull company. The discussions were tough, and many years later, Georges Vieillard remembered Watson to have said, “Mr. Frenchmen, you got up too late to conduct this trade.” The French answer should have been, “Mr. Watson, the sun shines on all of us, we will not do anything against IBM, but France has the right to have a punched-card industry. That is exactly the objective we pursue and which fills us with enthusiasm. We will stay independent and do everything to remain so.”137 IBM did not manage to buy the Bull company, which would continue to navigate alone.
As a consequence, Bull had to rely on French capital supplied by the Cailles family, who owned the Aussedat paper mills. The family had contributed to raising additional capital for the Bull company in 1932 and 1933, and Jacques Cailles was their representative on Bull’s board of directors. The Cailles family now became the majority shareholder through additional investment. There can be two reasons for this investment, to safeguard the Bull company in French hands or to consolidate the Aussedat company through a forward integration.
The reason for safeguarding the company in French hands is evident in the previously mentioned story regarding Watson’s negotiations in Paris, and this reason explains why Frenchmen invested in the Bull company, though it does not explain how the Cailles family was selected to accomplish this task. This can perhaps be explained through a strategy by the Aussedat paper works for forward integration. By 1935, they were providing Bull with punched cards, a business relation dating back to the early days of the Bull company in France. Then Bull approached Aussedat to acquire punched cards, but Aussedat had difficulties manufacturing cardboard of the same quality as the paper mills in the United States. To improve the quality of their punched cards, Aussedat obtained a license from the Racquette River Paper Company in the United States, who was a manufacturer with a reputation for punched-card production. Since that time, Bull had been providing ever-increasing prospects for the Aussedat paper mills. Now the threat of a takeover by IBM worried Aussedat, as IBM required their costumers to buy punched cards from them.138
After Bull had received a third increase in share capital, the company faired well during the remaining part of the 1930s. From 1935, the company managed to produce well-functioning machines, and by 1942, it accomplished becoming full competitor to IBM when it built the last of several required punched-card machines. From 1936, the French Bull company attained a surplus. While the increases in share capital in 1932, 1933 and 1935 had been caused by operating deficits, from 1936 to 1939, the surpluses provided the basis for the company’s expansion.
This is a case that appears to confirm the success of a company in France during the 1930s, that is, once national capital was attained. However, this story does not explain how the asset of national capital provided business, a feat that was accomplished by people in the Bull company via three interrelated social networks that also provided capital for the company’s infusions of capital in 1932, 1933 and 1935: the Association of French Industrial Producers (Confédération générale de la production française), a network of (former) commissioned officers in the French army, and a network of polytechniciens.
First, the Bull company was admitted to membership of the Association of French Industrial Producers, distinguishing them from their foreign-based competitors, the IBM subsidiary and SAMAS.139 Notably, this took place as early as December 1931 at time when the majority of the company’s capital was foreign.140 This admittance was most likely facilitated through the good offices of Lieutenant-Colonel Emile Rimailho, who was a member of the board of directors between 1931 and 1937 and its chairman from 1931 to 1934. Rimailho was a distinguished French artillery officer who had contributed to the design of a gun and, subsequently, had moved into industry, where he became a specialist in scientific management—hence his interest in office mechanization.141
Rimailho illustrates the company’s early close relations to a network of commissioned officers in the army and to French industry. Close relations to French industry were also mediated through two polytechniciens, Georges Vieillard and Elie Doury, who suddenly turned up in late 1931. Using their extensive network of polytechniciens, they accomplished within a short period the establishment of a consortium of punched-card users in France, which acted as the network that raised capital for the company in 1932, 1933, and 1935. Vieillard was appointed managing director of Bull in 1933 to secure his continued services for the company.142
Jacques Cailles, who was the second commissioned officer who became important for Bull, graduated from the Military Engineering University (École spéciale militaire) of Saint Cyr. He succeeded Rimailho in 1934 as chairman of Bull’s board of directors and represented the Cailles family, who had contributed to the raising of additional capital for the company in 1932 and 1933 and became majority shareholders in 1935.
Although René Carmille, the third key military officer in the story of Bull’s success in France, was employed outside the company, he became a core supporter, which was managed by his fellow polytechnicien Georges Vieillard. In 1935, Carmille advised the government to make a substantial direct investment in the company, which was suffering from financial problems arising from their difficulties in producing reliable machines.143 Carmille’s investment proposal was not accepted, but he subsequently provided advice that was favorable for the Bull company. Carmille resolved in 1935 to acquire Bull equipment to control army transport expenditures—in spite of a recommendation from the responsible office to purchase IBM machines.144 Further, Carmille selected Bull equipment for the tests for a conscription and mobilization system in 1935, which failed due to malfunctioning machines, and as the basis for the national register that started in 1940.
These examples illustrate the initial foundation of Bull’s success in the 1930s. The effective establishment of business networks provided access to a dynamic set of customers, and the establishment of a basis of French capital for the company was but one component in this process. Similarly, in Germany, Dehomag’s well-established and extensive business network was a crucial reason that it avoided the brunt of autarchy until 1940, in spite of the fact that the company was 90 percent owned by IBM in the United States. In contrast, SAMAS and IBM in France focused on becoming Frenchified and IBM lost steam, which can be explained due to a weaker business network. The second foundation of Bull’s success in the 1930s was the company’s ability to develop and produce reliable machines with the features required in France, which again resembled Dehomag’s experience.
The young Bull company had serious problems until 1935 in producing reliable punched-card machines. In addition, they needed to extensively develop the machines to catch up with the competitors. This was the basic reason for the company’s need to raise additional capital in 1932, 1933, and 1935, a situation that was aggravated by the company’s original low-paid capital and the leasing of 37 percent of its tabulators.145
It proved difficult to move the production of machines from Oslo and Zürich and to establish an industrial production in Paris, and many of Bull’s machines frequently made errors. The company only managed to produce reliable machines in 1935.146
The Bull company’s development of punched-card machinery in the 1930s was closely linked to its attempt to produce equipment that equaled the technical capabilities of that of IBM and SAMAS, which extended the scope of Bull’s strategy from statistics processing to encompass bookkeeping applications. In this process, Bull distinguished itself by shaping its own path to bookkeeping with punched cards. In addition, though the Bull machines were modeled on the same basic electromechanical technology as IBM’s, the Bull people chose separate designs.
Bull started as a producer of statistics-processing equipment in France.147 Their first production of equipment derived from earlier machine building in Norway and Switzerland of a nonprinting tabulator, a horizontal sorter, and a simple manual punch. In comparison, its competitors had produced numeric printing tabulators for several years. Demand for numeric printing capability had been mentioned by the people from the Egli company during negotiations with Norwegian engineer Knut Andreas Knutsen in Oslo in 1927.
In 1924, Knutsen had succeeded Bull as the leader of the development of the machines.148 This demand caused Knutsen to design a numeric wheel printer, an accomplishment that proved to be a major reason why the French Bull company was established and why Knutsen was hired to manage Bull’s development activities. In late 1930, Knutsen hired Norwegian engineer Anders Eirikson Vethe to innovate Knutsen’s patented design for a printing tabulator for production. In September 1931, the first numeric tabulator-printer was supplied to compile social statistics at the Ministry of Labor. It used printing wheels and printed at a speed outdistancing its competitors’ tabulators. The reliability of this tabulator was subsequently improved.149
The Bull company had the choice of focusing on the development of several features for their equipment for the extension of the scope of punched-card applications to encompass bookkeeping, alphanumeric capability, an extension of the 45-column card, and improved calculation capacity. First, influenced by considerations by insurance companies to introduce punched cards for insurance policy administration, they focused on alphanumeric punch card printing.150 However, by 1934 they had only gained two French insurance companies in addition to Rentenanstalt Zürich (today, Swiss Life), which had used Bull equipment since 1926.151
The problems involved in implementing alphanumeric systems were so great that all producers initially introduced systems that were not yet completely alphanumeric, for example, the British Powers company’s reduced alphabet unit from 1921 could not print numbers, and IBM’s first alphanumeric system with a reduced alphabet, introduced in 1931. Knut Andreas Knutsen initially built a tabulator with alphanumeric representation in a third of its printing positions and exclusively numeric representation in the remaining two-thirds in 1932. However, it had different representations of numbers in the two sections, reducing flexibility. At the same time, this tabulator was based on printing wheels instead of type bars, which proved to be an excellent basis for an extension from numeric to alphanumeric representation.152
The alphanumeric printing tabulator was particularly suited to print addresses that required a combination of letters and numbers, possibly indicating that insurance policy administration was the objective of this development. IBM’s research department in New York also developed a wheel printing tabulator, but their work was discontinued as a result of Knutsen’s patents.153 Eventually, in 1949, at which point Knutsen’s patents had expired, IBM marketed a wheel printing tabulator with a printing mechanism similar to that on the Bull tabulator.154
The Bull company’s alphanumeric code, 1935. The uppermost row (12) was used for qualitative indications, for example, a negative sign. (K.A. Knutsen, [U.S.] Patent 2,175,530, filed in France in 1935 and issued in the U.S. in 1939)
After completing the tabulator with reduced alphanumeric capability, Knutsen finished an alphanumeric design in 1934 that used the old digit representation standard for combinations with letters that allowed calculation on all numbers punched on to cards. This representation was distinct from IBM’s and from the subsequent alphanumeric systems designed by the American and British Powers companies, making alphanumeric punched-card systems proprietary, distinguishing them from the exclusively numeric systems that was an industry standard. A tabulator based on Knutsen’s improved design was produced in 1935 and continued to be manufactured without major modifications until 1968.155
Extending the old standard 45-column card was the second basic feature of the Bull punched-card system to be improved. After 1928, IBM used their new 8-column card and began supplying alphanumeric machines in 1933. From 1931 the Bull company prepared their machines to be built for 80-column cards.156 However, because of Bull’s electric reading of cards, this implied adopting rectangular or oblong perforations that would constitute an infringement on IBM’s French patent on this card design.157 Bull first applied 80-column cards for the machines used by the French army to test a new conscription administration system in 1935, which failed because of unreliable machines. For this application, the extension from forty-five to eighty columns facilitated additional information about the conscript to be stored on the card, avoiding the use of a second card for every conscript that would have made processing more complex.
The French army required 80-column cards for this application, as the extra capacity was needed and as IBM had alphanumeric machines for 80-column cards. Subsequently, the French army made 80-column cards a prerequisite for orders, and the Bull company marketed the 80-column machines in 1938.158 This caused IBM to instigate legal proceedings. Bull lost this lawsuit in 1941 in the lower courts, but they later won the appeal in 1947.159 Only during the appeal was it disclosed that this patent had been dissolved in the United States in 1933, as the design had been disclosed in the United States prior to the filing of the patent application.160 Thus, IBM’s suit had been based on a dissolved patent.
It is remarkable that a subtracting tabulator first reached the market only in 1936, subsequent to the alphanumeric facility and the initial introduction of the 80-column card. One of the partners of the burgeoning Bull company was the Egli company in Zürich, which produced key subtracting machines. The 80-column card was introduced before the subtracting tabulator for reasons of demand, but the late introduction of a subtracting tabulator was the result of problems the people developing the Bull machines had in completing a reliable design. Bull’s development people planned a subtracting tabulator as early as 1931, and they started to build a prototype in early 1934.161 However, it took another two years before the subtracting tabulator, which was based on the designs of the Toledo key adding machine, to reach the market.162
The Bull company had a development department, which from the outset was managed by Knut Andreas Knutsen. In 1936 he had six engineers and three assistants, and he proved apt at attracting able engineers to accomplish the substantial development of the Bull machines in the 1930s. He hired Anders Eirikson Vethe to implement his basic design for a printing tabulator in 1930. Georges Ziguelde was hired from the French IBM company and André Ziguelde from the key office machine producer, Olivetti, in 1935, to develop a subtracting tabulator; they subsequently designed the punched-card multiplier, launched in 1938. Roger Clouet was hired and designed a tabulator with improved programming facilities that was marketed in 1939.163
Dynamics of Distinctions in Europe
Punched-card based bookkeeping systems were shaped in the years between the two world wars in Great Britain, Germany, and France with discernible differences. The earliest letter printing tabulator was developed and marketed by the British Powers company in 1921. However, this feature did not become a major advantage for the company, which only developed an alphanumeric system after the Second World War. Instead, the British Powers company focused on developing and marketing of cheap punched-card systems based on small nonstandard cards, which had substantial success.
In Germany, the Powers agency marketed the letter printing tabulator developed by the British Powers company, but they also experienced limited success. The German demand focused on numeric calculation capability, demand for alphanumeric punched-card systems only emerging during the Second World War. In contrast, the French punched-card applications developed an emphasis on alphanumeric printing systems in the 1930s, which Bull responded to with machine building, while the French IBM subsidiary marketed machines designed in the United States.
These varying systems for bookkeeping were shaped through a combination of different interpretations of national demand and in response to IBM and the Powers company (which joined with Remington Rand in 1927) in the United States. Like in the United States, most of the development of punched-card systems in European companies originated in actual and imagined applications. An example of actual applications was in Britain, when the Powers company based their building of the first alphabetic punched-card system on specifications from the Prudential Assurance Company in the years around 1920. An example of imagined applications was the Powers company’s envisioning the potential of their small size punched-card equipment for medium-size companies.
In Germany, the market account administration in banking in the 1920s provided the focus for Dehomag’s development of a numeric interpretation of punched-card bookkeeping. In France, the emphasis on alphanumeric features at Bull originated with the Swiss producer of office machines, Egli, which was instrumental in the transfer of the original Norwegian technology to France. However, alphanumeric features only became prominent at Bull when the French army started to demand alphanumeric punched-card equipment in 1935. Subsequently, the French army became the prime user of the Bull company, which shaped the technology of their alphanumeric features and the conversion of the company’s basic punched-card from forty-five to eighty columns. In contrast, IBM’s French subsidiary did not contribute significantly to the shaping of the technology in France. Instead, they concentrated on producing equipment designed by IBM in the United States, which satisfied the French government by saving foreign currency.
Through the development of these different punched-card systems, European punched-card producers contributed significantly to shaping punched-card technology. The most noticeable examples were the British Powers company making the first letter printing tabulator, Dehomag producing punched-card machines with advanced calculation facilities, and the Bull company making alphanumeric tabulators that used printing wheels. After the Second World War, alphanumeric became a standard capability in punched-card systems, Dehomag’s D11 tabulator was marketed by IBM in the United States, and IBM started to produce a tabulator with printing wheels once the relevant Bull company patents had expired.
The variations served to enhance the independence of the punched-card companies in Europe in relation to the parent companies in the United States. The relations between subsidiaries and parent companies had two interrelated dimensions, legal connections and a degree of operational freedom. The Tabulating Machines and Powers agencies in Great Britain and Germany were established by 1914. Originally, they were domestically owned, which provided great operational freedom. In contrast, the subsidiaries in France were only established in the early 1920s and were from the outset completely owned by the parent companies, which provided less operational freedom. Reduced operational freedom was also caused by the Tabulating Machine Company’s exploitation of the German mark crisis of the early 1920s by taking over their German agency, Dehomag. However, Dehomag’s successful founder, Willy Heidinger, remained and managed the company with substantial operational freedom. Assigning operational autonomy to nationals leading the subsidiaries in Europe was in contrast to the limited delegation within the company in the United States.
Heidinger’s subsequent success and autonomy in relation to IBM was closely linked to Dehomag’s development and production of punched-card machines in Germany, independently of IBM. Similarly, the degree of operational freedom in the two British punched-card companies was not based upon their legal independence of the parent companies in the United States, but on their success in business as well as the development and production of punched-card equipment. Powers-Samas had more independent development and autonomy than did the British Tabulating Machine Company. National development of punched-card technology became a means to gain autonomy.
While the various agencies, child companies, and the Bull company in France had different relations to IBM and the Powers company in the United States, they interacted on their national markets. The high degree of competition in Germany and France seems to have focused the national emphasis in the two countries, with computation in Germany and with alphanumeric systems in France. In contrast, the low competition British market segmented into two, punched-card systems based on standard punched cards from the British Tabulating Machines Company like those in the United States and cheaper systems based on small nonstandard punched cards.
Finally, shaping diverse punched-card equipment and applications in the various countries in Europe was enhanced by the reduction of international trade after the breakdown of the international currency standard in 1931. British development was least affected, as both the British Tabulating Machine Company and the British Powers company exclusively used British capital. Only the British Tabulating Machine Company relied to a large extent on foreign design, but they do not appear to have been much effected by the “Buy British” campaign in the 1930s, as they had steadily growing assets in sterling after 1925.
In contrast, the punched-card business was very much affected by the Frenchifying that occurred in France in the mid 1930s and the German autarchy from 1932. These measures did not eliminate the foreign control of the Powers and IBM companies in either country, but in Germany they contributed to the development of a separate version of punched-card technology with heavy emphasis on computation capability.