Page 28

CIM36

IN DEPTH SPECIAL REPORT Gérard Berry, PROFILE PORTFOLIO 28 CNRS INTERNATIONAL MAGAZINE Computer Conductor BY LAURE CAILLOCE CNRS 2014 Gold Medal. Gérard Berry is awarded France’s highest scientific distinction, for his pioneering work on real-time programming languages. B eing slightly late for our appointment, Gérard Berry apologizes. He has just received a call from a French daily, eager to know his opinion on computer science education in the country. As a CNRS Gold Medal winner, the computer scientist is much sought after—and only too happy to oblige. Lively and outgoing, he belies the stereotype of the “computer nerd” living in his own world. He welcomes us to his spacious office at the Collège de France, where he created the first permanent chair in computer science in 2012—an impressive record for an unconventional researcher who successfully forged a career at the interface of pure theory and industrial innovation. The Esterel language, which made him famous, is now used to operate real-time systems as complex and sensitive as airliners, trains, and nuclear power plants. A youthful passion He could have very well become a chemist. He learned to read from the chemistry books of his mother, a professor, and spent many hours of his teenage years manipulating beakers and reagents in the basement of his family home in the outskirts of Paris. He discovered computers at the age of 19. “I saw my first computer—an old Seti PB250—in 1967, when I started out at the École Polytechnique,” Berry recalls. “They had just begun offering a computer class and I took to this emerging discipline immediately: it combined the experimental aspect that I loved with pure logic, which I found highly stimulating.” The young student was instantly fascinated with the man-machine dichotomy: an odd couple in which man, “clever but slow,” tries to give instructions to a computer, “quick but brainless, since it only does what it is told to do.” After the École Polytechnique, Berry completed his training working as an engineer in the Corps des Mines, a French public service inter-ministerial institution where he began conducting research in computer science in 1970. It soon occurred to him that the programming language was the stumbling block in the man-machine relationship, and solving it became the work of a lifetime. “I understood immediately that it was very difficult to get it right,” he admits. His first research focus, also the subject of his thesis, was at the interface of mathematics and computer science: lambda calculus, a mathematical language that serves as the basis for many programming languages. This choice of a purely theoretical subject might have also been borne out of necessity. “Back then, no research laboratory in France had a computer worthy of the name,” Berry recalls. “At least until the launch of IRCAM1 in 1977, when the composer Pierre Boulez rang the alarm bell loud and clear. He subsequently obtained the first machine suitable for proper research. The other laboratories, including mine, were not equipped until 1982.” Working in computer science without a computer had its advantages, though: the experience made Berry an exceptional theoretician and laid the foundation for a French school of computer science, which is now recognized around the world for its power of abstraction. Esterel: a brilliant intuition In 1977, Berry moved from Paris to Sophia-Antipolis in southeastern France, where the École des Mines had just opened the automation control theory and computer science research unit, which it soon shared with Inria.2 The turning point of his career came in 1982, when Microsystems, one of the world’s very first trade magazines, launched a competition for robot cars. “My colleagues in control engineering had produced an ultra-sophisticated vehicle but The Esterel language is now used to operate real-time systems as complex and sensitive as airliners, trains, and nuclear power plants. weren’t sure how to program it,” the researcher says. “Until then, I had always dealt with programs that simply involved some input data, a calculation, and an output result. Yet this particular machine needed to react constantly to its surrounding environment. This is how we devised a language that would be totally different from what everyone else was doing. That language would eventually become Esterel.” 1. Institut de recherche et coordination acoustique / musique. 2. Institut national de recherche en informatique et en automatique.


CIM36
To see the actual publication please follow the link above