Genie Logiciel et Logiciel pour l ' ingenierie

447 This discussion paper surveys the current state of software for scientific and engineering applications, and the foreseeable evolutions. It argues that a serious change in the attitude of scientists and engineers is necessary in order to master the growth of problem size. The discussion includes aspects of methodology, tools and languages ; we feel that much progress is needed in all three areas. The current actions of the ANSI Fortran committee are seen as particularly obnoxious with respect to the overall goal of providing scientific programmers with adequate tools.