Séance de 8 mars

Lieu: MESHS, 2 Rue des Canonniers, 59000 Lille

Nous avons le plaisir de vous annoncer la prochaîne séance du séminaire « Histoire et Philosophie de l’Informatique”, mercredi 8 mars 2017 à 13h30, MESHS, Lille avec la participation de Ulf Hashagen, Marie-José Durand-Richard et Olivier Bournez.

Ulf Hashagen (Deutches Museum, Munich, Allemagne) Analog computing as a failed modernization program in Germany 1930-1960.

In the late German Kaiserreich the invention and production of
 mathematical instruments became a particular strength of the German  innovation system—this was part of a general development that made the  German instrument industry the world leaders during this period of time.  However, after WW I the financial crisis of the German state partly  hindered the further development of large scale instruments, and as a  consequence other national science systems partly overtook the world  leading German science system. The case of the epoch-making Differential  Analyzer built at MIT by the electrical engineer Vannevar Bush in 1931  serves not only to illustrate this argumentation, but it also shows that  the ideology of pure science did not allow building a large scale analog  computing machine in the environment of a German university. While some
 differential analysers were built in Britain and in Continental Europe
 in the 1930s, the German development started only after the outbreak of  WW II, when the rationality of the German science system had changed  totally to a rationality of applicability of science for military  purposes. In 1939/40 two different ambitious differential analyzer  projects were launched in connection with the Heereswaffenamt or rather  with the V2 rocket project lead by Werner von Braun. The talk shows how  the Third Reich used the scientific competence of mathematicians as  resources for the developments of weapons in the military system and  how, the other way round, mathematicians used the Third Reich to get  financial and personnel resources for the building of differential  analysers for their own purposes. Furthermore, it will be analysed, why  all the differential analyser projects in the Third Reich failed in the  end. The paper aims to focus on fundamental aspects of scientific and  technological developments in computing in the Third Reich and to  analyse the dynamics of its military-industrial-academic complex.

Marie-José Durand-Richard  (Sphère, Université de Paris), Douglas R. Hartree (1897-1958) et la résolution des équations différentielles par machine : méthodes analogiques et digitales

Dougals R. Hartree, (1897-1958) physicien et mathématicien anglais, est directement concerné par le calcul numérique,  
• d’abord calculs balistiques pendant la 1ère guerre mondiale sous la conduite de son père, pratiqué manuellement et à l’aide de machines de bureau,
• puis à partir des années 1930 sur l’analyseur différentiel, machine analogique, dont il a dirigé la construction à l’université de Manchester en 1934-35, à partir des plans du premier exemplaire de ces machines, réalisé par Vannevar Bush au MIT entre 1927 et 1931,
• enfin, après la Première Guerre Mondiale sur les premiers ordinateurs, de l’ENIAC aux USA, où il se rend en 1945; et sur l’EDSAC en construction à Cambridge.
C’est dire que toute sa carrière est concernée par la mécanisation du calcul. Cet exposé s’attachera à mettre en évidence et à comparer les méthodes de résolution numérique des équations différentielles sur ces différentes machines.

Olivier Bournez, Continuous Time Models of computation, Lix, Ecole Polytechnique

We will revisit several models of literature. This will include the General Purpose Analog Computer (GPAC). The GPAC was proposed as a mathematical model of Differential Analysers by Claude Shannon in 1941. We will compare their computational power to modern models, such as Turing machines, using arguments and tools from computability and complexity. This will lead to discussions related to Church-Thesis (and its variants) for continuous time models and continuous models in general.