Are Player Pianos Early Computers ?
By Philippe Rouillé
|Paris, Feb 27th 1996|
I read with very much interest in this digest various opinions about the ambition of Player pianos to be (or not to be) early computers. Perhaps you could be interested reading the english summary of an article I wrote 2 years ago for the quarterly Journal ("La Revue") of the Musee National des Techniques, in Paris. (The museum has a web site at:
where you may read all these summaries).
Memory gaps / Philippe Rouille (<email@example.com>).- Musee des arts et metiers, La Revue, feb. 1993, n 2, p.34-41. Translation from French into English: firstname.lastname@example.org, 3 oct. 1994
The "memory gaps" in question here are designed to restore pre-programmed actions or pre-recorded data. There is a great deal of similarity between the solutions adopted by very different branches of industry (textile, publishing, manufacture of musical instruments, etc), for recording often very complicated actions to be reproduced identically and repeatedly, without direct human intervention. A long technical and, above all, psychological journey was needed to arrive at these solutions which are in turn being challenged today by other methods which they have helped to create.
The cam, known and used since ancient times, and its derivatives, has long been the best means of reproducing a pre-programmed movement. Used for example in fullers earth mills or in the movements of "jacquemart" clockwork mechanisms, these systems "in relief" present several limitations:
- the duration of programmed operations depends directly on the
circumference of the cylinder which supports the cams;
- the cylinders can sometimes be changed, but the operation is difficult
and storage is cumbersome;
- all subsequent alteration of the programme is both difficult and costly.
The textile industry was to play a pioneer role in this respect. In 1725, Basile Bouchon, a textile worker from Lyon, used a sheet of perforated paper in a loom to ease the task of fellow workers whose job was to pull the cords on the loom. The way the perforations were arranged was established according to the pattern to be reproduced on the cloth. The fact that Bouchon penetrated the needles he wanted neutralized into the holes in the paper, while the active needles were pushed back by the raised parts, is significant of the psychological difficulty in imagining that the empty spaces could play an active role. However, the idea of commanding the repeated actions of a machine, not by cams firmly attached to a rotating shaft, but by holes made in a strip of paper, is a revolutionary one.
In 1728, Falcon, who worked with Basile Bouchon, replaced the paper strip by a set of punch cards attached to one another, enabling the programme to be changed rapidly. Vaucanson, in 1745, further automated the loom but used a perforated cylinder, less flexible than the strips of paper or card.
These innovations were taken up and improved by Jacquard in the 19th century with looms comprising up to two thousand hooks.
Other sectors were subsequently to make wide use of these "memory gaps". In 1843 in Lyon, Claude-Fe'lix Seytre patented a system using punch cards in mechanical musical instruments.
Following confidential tests on various factors, perforated card or paper became commonly used in pianos and automatic organs from around 1880 to 1890.
Several types of media coexisted with different types of reading methods. The punch card, folded accordion-style, was notably used in street, barrel and fairground organs; its thickness made it cumbersome, but its solidity enabled intensive usage and the mechanical use of the sensors.
The roll of perforated paper provided the memory for pneumatically activated instruments: pianolas, violins, banjos, harps and so on. Of smaller volume, it enabled a large repertoire of very long pieces, but its fragility restricted its use to pneumatic commands.
Disks - metal for music boxes composed of vibrating strips (symphonion, polyphon), or cardboard for the family of organettes, small portable harmoniums - were widespread from around 1890 to 1914.
Four reading systems were to appear successively and are still used, according to needs: mechanical reading, pneumatic reading, reading via electrical contact and, last of all, the optical reading of punch cards and tape in statistical or information processing machines. All these devices were to enable increasingly faster reading speeds.
Around the end of the 19th century, several industrial sectors began to grasp the potential of perforated media. Printing was one example, with the invention in the US in 1885 of the "Monotype" which operates via coded punched tape very similar to that used in pianolas.
Perforated media are also to be found in certain Braille writing machines for the blind, as well as in telegraph or telex tapes, in arithmetical, statistical and information processing machines and also, in industry, for commanding large numbers of machine tools.
However, the use of perforated media to record and reproduce data also has its limits. The system is deficient if it is required to instantly restore very complex or nebulous data. It has therefore been replaced by similar technologies: magnetic media or laser scanning techniques.
MUSEE DES ARTS ET METIERS - 292, rue Saint-Martin - 75003 PARIS - FRANCE
(Message sent Tue 27 Feb 1996, 23:01:04 GMT, from time zone GMT+0100.)