The first really practical machine of this type, however, was invented by a Frenchman named Charles Xavier Thomas, in 1820, and has since become known as the “Thomas Arithmometre.”
The Thomas machine is made and sold by a number of different foreign manufacturers, and is used to a considerable extent in Europe and to a limited extent in the United States.
The key-set principle more practical for recorders
But two-motion calculators, from Leibnitz down to date, have always been constructed so that the primary or first action involved merely the setting of the controlling devices and performed no function in the supplying of power to operate the mechanism which does the adding. With such machines the load was thrown on to the secondary action.
This, of course, made the primary action of setting, a very light action, especially when keys came into use, and as there are several key depressions to each secondary or crank action, it may be understood that while the action of Felt’s printing or paper shift-lever was light, the action of the keys which were called upon to perform most of the work was much harder than it would have been if his adding mechanism had been designed on the key-set crank-operated plan of the regular two-motion machine such as illustrated in the Pottin or Burroughs patents described.
Thus, when Burroughs applied the Felt recording principle to his key-set crank-operated adding mechanism, he produced a type of recording machine which proved to be more acceptable from an operative standpoint than the recorder made by Felt; and yet the writer has read testimonials given by those who had both the Felt key-driven recorder and the Burroughs key-set crank-operated recorders, who claimed they could see no advantage.
From Drawings of Burroughs’ Patents Nos. 504,963
and 505,078
Probably the best proof lies in the fact that Felt finally abandoned the key-driven feature in his recorders, as may be noted from the later-day “Comptograph.”