A Learning By Doing Approach To Teaching Computational Physics

Scientific research is becoming unthinkable without computing. The ubiquity of computerized instrumentation and detailed simulations generates scientific data in volumes that no longer can be understood without computation. Computational physics is a rapidly growing subfield of physics and computational science in large part because computers can solve previously intractable problems or simulate natural processes that do not have analytic solutions. One can easily argue that all graduates of science or engineering programs should have the opportunity to develop good computing skills by the time they complete their studies. However, the depth and range of skills needed varies considerably – even in a single discipline such as physics. Moreover, the interests, backgrounds, and abilities of students taking physics courses vary widely, whereas the number of instructors with scientific computing skills has been rather limited. Providing appropriate courses and instruction in computational physics for such diverse student population is a challenge. On the other hand, computational physics provides exciting teaching opportunities that can complement traditional methods of teaching in the lecture or the laboratory. We use a laboratory project-based approach, where the students are learning by doing. The course is divided into two sections, lecture and laboratory session. During the laboratory session, the students work at mid-term and final projects, while the lecture the programming, numerical and computational techniques and methods are discussed. The usefulness of this approach is evaluated by surveys conducted every semester, and feedback from other educators is highly appreciated.