‎COUFFIGNAL, Louis‎
‎Les machines à calculer. Leurs principes, leur évolution‎

‎Paris, Gauthier-Villars, 1952, in-8, VIII-[2]-86-[6] p, 27 fig. dans et hors-texte, broché, couverture imprimée en noir et rouge de l'éditeur, Première édition, rare, introduite par une préface de Maurice d'Ocagne dont Couffignal était l'élève et dont il hérita de la même passion pour des machines à calculer. A la mort de son maitre en 1938, Couffignal fut considéré alors considéré, en France, comme le meilleur expert en calcul numérique. A partir de 1946, il eut la charge de mettre au point et de construire le premier ordinateur français digital. Couffignal eut l'idée de concevoir sa propre machine à calculer en 1933, l'année même où il publia le présent ouvrage : une approche historique du développement des machines à calculer. "Couffingal's findings, as reported here, persuaded him that calculating machines, like living organisms, must necessarily evolve from the simple to the complex, in a way that would as far as possible eliminitate the need for human intervention. He also became convinced that the guiding principle in mechanical calculation should be organization of the calculation procedures themselves. Couffignal sought to bring his calculating machine plan to fruition while serving in the administration of the CNRS. During the war years little was done to advance Couffignal's project, and it was not until 1947, with the foundation of the Institut Blaise Pascal, that Couffignal at last obtained the grants that would enable him to build his machine. Couffignal's design for a calculating machine differed substiantially from those then being constructed in the United States and England in that it was founded not on the principle of the stored-progam, but on the classification and organization of calculation procedures. At a small conference on new developments in electronic computers held in Paris in 1947, Couffignal "explicitly disassociated himself from the American and English views on stored-program computer design. Later at the first electronic computer conference at Cambridge in 1949 in which ran against the convincing evidence being presented by all other speakers at the conference. In his view the stored-progam design, which reduced the calculating part of the machine to a simple counter and gave priority to the memory, was going agains the law of progress... In Couffignal's view, it was the scientific organization of the calculation that should take priority" (Ramunni 1989, page 252). Needless to say, Couffignal was wrong in his assessment : computer development proceeded along the paths charted by the Americans and British, and Couffignal's machine was never built, except in a small prototype version demonstrated at the Pascal Institute's 1951 symposium on Les machines à calculer et la pensée humaine. The time and enerby invested in this technological dead end delayed France's entrance into the electronic computer age. The CNRS did not obtain a stored-progam computer (British) until 1955. This machine was on of the first of its kind installed in France" (Origins of cyberspace, 281). Bel exemplaire de cette monographie rare écrite par un des pionniers de l'informatique française. Papier uniformément jauni, petit manque en pied du dos, notes marginales manuscrites. Couverture rigide‎

Reference : 62855

‎Bon VIII-[2]-86-[6] p., 27 fig.‎