How strange it was to find an old-fashioned computer science course repackaged as a new masters program in scientific computing, and presented in a journal called Future Generation Computer Systems! And when I say old-fashioned, I mean it as a compliment: this paper describes an excellent masters curriculum, which will make better programmers and scientists out of physicists, chemists, biologists, and engineers. As the author points out, “even those who appear to be very familiar with computers are often not used to formulating algorithms or writing programs.” Having wrestled with Pascal word size on an ICL 2980 in the late 1970s, while solving a nonlinear diffusion equation that modeled charge transfer in a CCD, I welcome programs of this type, which are available at several American institutions.
Computational science and engineering (CSE) is now regarded as a crucial element, along with theory and experiment, of scientific investigation and engineering design. The Technical University of Munich is solving problems of research training for European postgraduates in scientific computer science by supporting the seven core modules of this course, across seven faculties. The curriculum covers core computer science subjects, such as scientific programming, algorithms, software engineering, and parallel programming, and balances hardware topics, such as networks and computer architecture, with options in core subject problem solving and applied mathematics. It has been designed to meet the Society for Industrial and Applied Mathematics (SIAM) requirements for CSE education. A slight drawback to this paper is its editorial sloppiness. This is no fault of the author, who produces sound arguments in favor of this form of pedagogy and well-thought-out examples.