Before 1886, calculators didn’t have keyboards.  They were based on the Pascaline, a calculating device invented by Blaise Pascal (1623-1662).  Pascal was only 18 years old when he conceived the Pascaline in 1642. A precocious French mathematician and philosopher, at the age of 12 Pascal discovered that the sum of the angles in a triangle is always 180 degrees. The Pascaline, built in 1643, was possibly the first mechanical adding device actually used for a practical purpose. It was built by Pascal to help his father, Etienne Pascal, a tax collector, with the tedious activity of adding and subtracting large sequences of numbers.

It was built on a brass rectangular box, where a set of notched dials moved internal wheels in a way that a full rotation of a wheel caused the wheel at the left to advance one 10^th. Although the first prototype contained only 5 wheels, later units were built with 6 and 8 wheels. A pin was used to rotate the dials. Subtraction was done by applying a cumbersome technique based on the addition of the nine’s complement.

Although the machine attracted a lot of attention in those days, it did not get wide acceptance because it was expensive, unreliable as well as difficult to use and manufacture. By 1652 about 50 units had been made but less than 15 had been sold. Initially, Pascal got a lot of interest in his invention and he even obtained a "privilege" protection (medieval equivalent of a patent) for his idea in 1649, but his interest in science and "material" pursuits ended when he retreated to a Jansensist convent in 1655 concentrating all his attention on philosophy. He died in 1662.

Even at the beginning of the 20th Century, several companies introduced models based directly on Pascal's design. One example is the Lightning Portable Adder introduced in 1908 by the Lightning Adding Machine Co. of Los Angeles. Another example is the Addometer introduced in 1920 by the Reliable Typewriter and Adding Machine Co. of Chicago. None of them achieved commercial success.

It was 1672 when the famous German polymath, mathematician and philosopher, Gottfried Wilhelm Von Leibniz (1646-1716), co-inventor of the differential calculus, decided to build a machine able to perform the four basic arithmetical operations. He was inspired by a steps-counting device (pedometer) he saw while on a diplomatic mission in Paris.

Like Pascal, Leibniz was a child prodigy. He learned Latin by the age of 8 and got his second doctorate when he was 19. As soon as he knew about Pascal’s design, he absorbed all its details and improved the design so as to allow for multiplication and division.  By 1674 his design was complete and he named this machine the Stepped Reckoner; then he commissioned the building of a prototype to a craftsman from Paris named Olivier.

The Stepped Reckoner used a special type of gear named Stepped Drum or Leibniz Wheel which was a cylinder with nine bar-shaped teeth of incrementing length parallel to the cylinder’s axis.  As in the Pascal device, there is one set of wheels for each digit. When the drum is rotated by using a crank, its movement is then translated by the device into multiplication or division depending on which direction the stepped drum is rotated.

Although the concept of the Leibniz Wheel became well know among the scientific community, there is no evidence that more than two prototypes of this machine were ever made.  

Even though Leibniz was one of the most outstanding polymaths of his time, he died in poverty and unrewarded. His machine remained in the attic of the University of Göttingen until a worker found it in 1879 while fixing a leak in the roof. Now it is in the State Museum of Hanover; another one is in the Deutsches Museum in München.

Pascal's and Leibniz’s designs were the basis for most of the mechanical calculators built during the 18^th Century. Giovanni Poleni made one in 1709, Lépine in 1725, Antonius Braun in 1725, Jacob Leupold in 1727, Hillerin de Boistissandau in 1730, C.L. Gersten in 1735, Jacob Isaac Pereire in 1750, Phillip Mathieus Hahn in Germany in 1773. Charles, the third Earl Stanhope, of England, in 1775; Johan Helfreich Müller in 1783, Jacob Auch in 1790, and Reichhold in 1792.

Special consideration deserves the Parson Phillip Mathieus Hahn (1730-1790) who developed in 1773 the first functional calculator based on Leibniz's Stepped Drum. Hahn's calculator had a set of 12 drums in a circular arrangement actuated by a crank located in the axis of the arrangement. Hahn made these machines until his death in 1790, however, his two sons and his brother-in-law, Johann Christopher Schuster, continued with the manufacture probably as late as 1820.

By the end of the 18th Century, calculating machines were still curiosities used for display purposes, rather than for actual use. The limitations imposed by the technology made it impossible to meet Pascal's dream of making them a practical calculation device.

It was not until the beginning of the 19^th Century that the calculator became a popular device. This happened in 1820 when Charles Xavier Thomas de Colmar (1785-1870), of France, made the Arithmometer   (French Patent N^o 1420, November 18, 1820), a machine based on Leibniz’s design which was capable of performing the four operations in a simple and reliable way. Because of its unidirectional drum, division and subtraction required setting a lever. The Arithmometer was also known as the Thomas Machine.

The Thomas Machine was very successful and one hundred years later, during the first half of the 20^th Century, the machine was still sold. Thomas received France's Chevalier of the Legion of Honor for his machine. About 1,500 machines were made by the Compagnie d'Assurance Le Soleil, founded by Thomas, and other contractors, between 1820 and 1930.

Like every successful product, the Thomas Machine had many clones, and the term Arithmometer became synonymous with four function calculating machine.

A significant change in Leibniz’s design was patented in 1875 by Frank Stephen Baldwin (1838-1925) of the United States. In the Baldwin Principle the nine stepped drums are replaced by a pinwheel which is a variable-toothed cylinder with sets of nine radial pegs that protrude or retract by the action of setting levers through slots located in the opposite side of the cylinder. A crank on a bi-directional handle activates the digit wheels geared in the variable-toothed cylinder. This invention was awarded the John Scott Medal for the most meritorious invention of the year.

A few years later, in 1878, Willgodt Theophil Odhner (1845-1905), a Swede working for Ludvig Nobel in Russia, patented a pinwheel device. Although Ohdner's patent is completely independent of Baldwin's invention, this device is so similar to Baldwin’s Cylinder that with the passing of time, the terms Baldwin type and Odhner type became synonymous for these kinds of calculating machines.

W.T. Odhner, Maschinenfabrik & Metallgiesserei started manufacturing calculators  under the name Original-Odhner, circa 1886, in St. Petersburg, Russia.  About 30,000 calculators were made in Russia and probably less than 20% were exported. 

In 1892 Odhner sold his patent rights to Grimme, Natalis & Co. A.G. of Braunschweig for sale in Germany and some neighboring countries. This company manufactured the Odhner-type calculators under the name Brunsviga, and later, the company itself was named Brunsviga. 20,000 units were sold between 1892 and 1912.

In 1917, the year of the Russian revolution, his son Alexander moved the production to Göteborg, Sweden, and sold the rights to manufacture this type of machine under the name Facit, which were produced until 1977.

The quest for a keyboard

Between 1850 and 1887, many attempts were made to develop a calculating machine that would use keys as means to enter data. In Europe, key-driven machines were made by V. Schilt (1851), F. Arzberger (1866), A. Stettner (1882), Bagge (1882), d’Azevedo (1884), Petetin (1885) and Maq Meyer (1886). In the United States: D.D. Parmelee (1850), Thomas Hill (1857), G.W. Chapin (1870), W. Robjohn (1872), D. Carrol (1876), Borland & Hoffman (1878), M. Bouchet (1883), C.G. Spalding (1884), A. Stark (1884), L.W. Swem (1885), P.T. Lindholm (1886) and B.F. Smith (1887) [7].

These attempts illustrate the difficulty in adapting the usage of keys to the wheeled mechanisms.

And then, on Thanks Giving Day of 1884, Dorr Eugene Felt (1862-1930) from Chicago, goes to the store in his neighborhood, buys a wooden macaroni box, staples, rubber bands, wire, string, and meat skewers, and builds a prototype for the first practical calculator to use a keyboard.  Soon, by 1886 he completed his first metal prototype, which he named the Comptometer,and in 1887 he was granted US Patent 371,496 and started the Felt & Tarrant Mfg. Co. in partnership with Robert Tarrant.

The Comptometer used a "multiple order keyboard" also called full keyboard which consisted of a matrix with 9 rows of keys, one for each digit (1 to 9). The number was entered by pressing one digit in each column. There were no Zero keys because zero was represented by the absence of a keystroke in the corresponding column. This arrangement of keys, initially introduced by Thomas Hill in 1857, became very popular during the first half of the 20^th Century.

Comptometer operation became a formal profession and required a lot of training.  A good Comptometer operator was able not only to perform additions and subtractions at fast speed, but also multiplications and divisions by applying repeated additions and complementary subtractions, respectively.

Few years before Dorr E. Felt started working on his Comptometer project, in 1880 William Seward Burroughs (1857-1898) started the development of an adding machine with a full keyboard and printing capabilities. He applied for a patent on January 10, 1885, almost two years before Felt applied for his patent, and was granted the patent on August 21, 1888, about a year after Felt was issued his patent.

In 1886, Burroughs founded the American Arithmometer Company in St. Louis, Missouri. By 1889 the company had sold about 50 machines, but they were difficult to operate. Burroughs immediately improved them by inventing the dash pot, a mechanism used to regulate the pull in the machine’s handle and combining his invention with several functional features of Felt’s inventions. This is reflected on the patent awarded to Burroughs in May 5, 1892. Burroughs achieved his objective but he only saw the beginning of his success for he died in 1898.

In 1904 the company moved to Detroit, and in 1905 it was renamed to Burroughs Adding Machine Co. Twenty years later they had sold about one million machines and had become the largest manufacturer of adding machines in the United States. It remained as one of the leading manufacturers of mechanical office equipment until the 1950's. After WW2 it expanded its operations to include computers, and in 1953 its name was changed to Burroughs Corporation. Finally, in 1986 it merged with Sperry Corporation to form Unisys Corporation.

The beginnings of the 10-key machine

A significant departure from the popular full keyboard came when William W. Hopkins of St. Louis, invented the Standard in 1901. This machine manufactured by the Standard Adding Machine Co., had one row of 10 digit keys.

[1] [2] [3] [4] [5] [6] [7] [8] [9] [0]

The following year, in 1902, James I. Dalton introduced the Dalton, an adding/printing machine designed by Hubert Hopkins, which had two rows of five digits arranged in a ay such that the lower digits were in the left, the higher digirts on the right and the five and zero on the middle:

[2] [4] [5] [7] [9]
[1] [3] [0] [6] [8]

 The Dalton was very successful and over 150 models were introduced until 1928.

[1] [3] [5] [7] [9]
[2] [4] [6] [8]
[0] [00] [000]

In 1914, when commercial mechanical calculators were just becoming practical, Oscar J. Sundstrand, founder of the Sundstrand Adding Machine Co., introduced the modern 10-key keyboard design with three rows and a zero key arranged as:

                           [7] [8] [9]
                           [4] [5] [6]
                           [1] [2] [3]
                               [0]

In 1927, Sundstrand sold the rights to Underwood-Elliot Fisher Co. and worked for this company until 1949. In 1950 he joined the Victor Adding Machine Co.

Sundstrand keyboard design became the de-facto standard for mechanical calculators, and then it was also adopted by handheld electronic calculators when they started to appear in the market at the beginning of the 70's. 

Since then, electronic calculators have evolved in many different ways, displays changed from LED’s to LCD’s, and every imaginable combination of shapes, material and colors have been used. However, they have all inherited the basic 10-key design from Sundstrand's mechanical calculator.

 The limitation of the mechanical technology has prevented from thinking about other ways to enter numbers that were not keying a sequence of digits, but our mind does not conceive a number as a sequence of digits.  We live in the year “two thousand seven” not in the year “two zero zero seven”.

Numbers are objects made up of small quantities supported by powers of ten acting as numerical structures.  These structures lead us to develop the concept of Verbal Numerals.  In a verbal numeral, digits are used to represent small quantities, and a symbol is assigned to each numerical structure. For example:

                                                H   for “Hundred”
                                                T   for “Thousand”
                                                M  for “Million.”

                                       Yes, we live in the year “2T7.”

The usage of Verbal Numerals requires a procedure to convert them into its corresponding sequence of digits.  This method covered by US Patent  5,623,433, extends a normal keyboard with keys for "Hundred," "Thousand" and "Million," so that the number can be entered as it is pronounced. This procedure can be used for all major languages, and have the following advantages:

•Consistent with the way the mind conceive a number.

•In many cases requires less symbols to be entered.

•Avoids mental conversion into a sequence of digits.

•Easily adapted as input means in electronic devices.

•Does not replace but expand the traditional input procedures