, Kalamazoo College
Pages: pp. 11-15
You know one if you see one, but can you define a computational physics course in general? Even more fundamentally, can you specify what role numerical computations should have in any standard physics course?
The quest to address such questions was the motivation for a project that has culminated in the publication of this special issue. I believe and hope that our nonphysicists readers will regard this special issue as an opportunity to gather information and learn lessons that extend beyond physics to other disciplines.
To appreciate CiSE's concern, I refer you back to the remarks of the editor in chief at the magazine's inception. George Cybenko had this to say about the merger between Computers in Physics and IEEE Computational Science & Engineering (vol. 1, no. 1, Jan./Feb. 1999, p. 1):
" CiSE is setting up camp at the confluence of two great intellectual rivers—the physical sciences and the computational sciences. This camp will grow into a town and then a city but only if we learn each other's languages and trade in good faith.
[…] By publishing novel ideas from such a broad array of topics and specialties, we must be open-minded and helpful to each other. More likely than not, what might appear to be a wrong-headed argument or approach in one community is what another community takes as gospel. This creates opportunities both ways. We stand to learn a lot or to reach an enormous new audience by setting the record straight, whichever the case may be."
The genesis of this theme issue began at a discussion on computation in physics courses hosted by the American Association of Physics Teachers (AAPT) Educational Technologies Committee in 2005. It was a lively exchange that covered a range of questions and issues: What is computational physics? Is computational physics a new, third branch of physics? What factors encourage or impede the development of computational physics? What is the state of undergraduate computational physics programs? What can we do to promote computational physics? What are the most appropriate mechanisms for sharing information? Is there a need to train faculty?
As in any academic dialogue, the participants voiced a spectrum of opinions, but one consensus emerged—namely, the need to conduct a national survey on the state of computational physics at the undergraduate level. Furthermore, the person conducting the survey should possess two qualifications: to be a known and respected leader in physics education and to have an open mind about the nature and role of computations in undergraduate physics courses. Following this recommendation, Norman Chonacky ( CiSE's current EIC) commissioned Robert Fuller, a professor of physics emeritus at the University of Nebraska, to conduct a national survey on the uses of computation in undergraduate physics in the US.
More than 250 physics departments responded to Fuller's 28-question survey, which revealed a strong overall interest in computational physics but a range of opinions on the functional details. Ultimately, the survey results contained some interesting and thought-provoking responses.
Building on this study, the AAPT Educational Technologies Committee hosted a three-hour invited undergraduate computational physics programs session followed by a two-hour poster session at the AAPT 2006 summer meeting, which ultimately featured six invited papers and 17 posters. The organizers selected the presentations and posters to give voice to a range of opinions on computational physics from experienced faculty. Most of the effort and progress in this field has come from dedicated and often isolated faculty. We hope this special issue of CiSE—and the special sessions on which its articles are based—helps grow a supportive community that will facilitate the development and integration of numerical computation into all aspects of undergraduate physics curricula.
Norman Chonacky opens the issue with a discussion on the rationale for Fuller's survey; you'll also find a detailed report of Fuller's findings.
In "Computational Physics: A Better Model for Physics Education?" Rubin Landau offers a peek at Oregon State University's new BS in computational physics, which includes five multidisciplinary courses. The physics department has developed textbooks for each course and makes extensive use of Web-based curricula materials.
In "Implementing Curricular Change," Marty Johnston describes why it's not an easy task to integrate computational elements across the entire curriculum, particularly because of the widely varying levels of computational skills among the faculty.
In "Using Computational Methods to Reinvigorate an Undergraduate Physics Curriculum," Jaime R. Taylor and B. Alex King III explain how to incorporate computational methods into a physics curriculum by either creating a separate computational methods course or teaching it across the curriculum. Regardless of the approach, specific elements are important to prepare students for careers in physics.
Finally, in "An Incremental Approach to Computational Physics Education," Kelly Roos discusses a program that integrates a successive approach to incorporating computational methods into a traditional physics curriculum. Students are introduced to, and produce, numerical solutions involving chaos and nonlinearity.
The Fuller survey's results are revelatory and promise to continue to be useful as the computational physics community moves forward. Repeating my opening remarks, I hope that readers from other disciplines might consider doing similar studies. The study has given us a much clearer view of the state of undergraduate computational physics and provided the opportunity to initiate the formation of a community of interested people. The sidebar lists abstracts for all the poster materials. If you'd like to get involved with the discussion, join the recently initiated discussion board on computational physics hosted by the National Science Digital Library's ComPADRE facility ( www.compadre.org/ psrc/bulletinboard/ForumDetails.cfm?FID=28).
The American Association of Physics Teachers' Educational Technologies Committee sponsored a "Computation in Undergraduate Physics Courses" poster session at its 2006 summer meeting held at the University of Syracuse. The poster session followed six invited papers, delivered during a two-hour session, on the same subject.
The following posters describe instructional uses of computational physics at the undergraduate level. You can find more detailed information on each project by contacting the authors directly, or at http://opac.ieeecomputersociety.org/ opac?year=2006&volume=8& issue=5&acronym=cise.
Charles Leming (Henderson State University; firstname.lastname@example.org), "Minimal Matlab for Mechanics Courses"
Matlab is recognized as one of the best tools for computing applications in physics research, but how easy is it for students to use? This poster gives examples of elementary Matlab applications in an intermediate-level undergraduate mechanics class. Students can acquire the small subset of Matlab capabilities used in the course in only one or two class periods. As a result, numerical methods impose minimal alterations to a traditional mechanics course and require no prior computing knowledge.
Tim Sullivan (Kenyon College; email@example.com), "Computation in Undergraduate Physics at Kenyon College"
Kenyon's physics department has a long history of innovation in its use of computers, and this poster details a novel, interdisciplinary program in computational science. The goal is to produce a computational science program appropriate to a liberal arts college while simultaneously enhancing opportunities for physics majors to learn more about computational techniques. The poster also describes three student-faculty research programs at Kenyon that depend heavily on computation: medical imaging, complex systems, and the dynamics of phase transitions.
Michael Dennin (University of California, Irvine; firstname.lastname@example.org), "Reviving Continuum Mechanics: How Computation Helps"
Independent of their value as a research tool, numerical methods open up an entirely new class of problems that was previously inaccessible to students. Instructors face multiple challenges when integrating numerical methods into a curriculum. Should they be taught in a stand-alone course? Should they be integrated across the curriculum? How does a large research university achieve such integration when faced with diverse faculty interests? This poster looks at numerical methods' impact on reviving continuum mechanics at the undergraduate level and how UC Irvine has worked to integrate them across the curriculum.
Richard Gass (University of Cincinnati; email@example.com), "The Role of Computation in Undergraduate Physics at the University of Cincinnati"
Defined broadly, computation encompasses visualization and symbolic calculation as well as more traditional numerical work. The University of Cincinnati uses all three of these aspects in both its laboratory and lecture courses. This poster describes the university's current efforts and examines future directions; it also discusses how these efforts fit into the broader, statewide initiative in computational science.
Kelly Roos (Bradley University; firstname.lastname@example.org), "Incremental Integration of Computational Physics into Traditional Undergraduate Courses"
To properly introduce students to such topics as chaos and nonlinearity—as well as to the stochastic nature of many topics in statistical mechanics—computational techniques don't just augment the treatment, they're utterly necessary. This poster describes such an approach's effectiveness at Bradley, arguing that computational material fits well into its "natural habitat"—that is, within the context of the physics that generates the need for computational treatment. The poster also shows some nonequilibrium computational examples.
Norman Chonacky ( CiSE magazine; email@example.com), "How Can Computing Change Physics Courses? Need for a National Discussion"
This poster provides the elements of a tentative strategy that Computing in Science & Engineering ( CiSE) magazine, in collaboration with the AAPT/AAS ComPADRE project, plans to use to facilitate a national discussion about how computing could change undergraduate physics curricula.
Robert Fuller (University of Nebraska Lincoln; firstname.lastname@example.org), "The Details of a Study of Computations in Undergraduate Physics Courses in the USA"
CiSE magazine has commissioned a study of the use of numerical computations in physics courses in the US. So far, more than 250 physics faculty members have responded, representing more than 220 different physics departments. This poster features the details of the study, the range of responses, some sample quotes, and a preliminary analysis. A more complete analysis appears in this special issue of CiSE.
Rubin Landau (Oregon State University; email@example.com), "The Beginnings of a National Repository of Computational Science/Physics Courses"
This poster describes progress in the US toward establishing a national repository of university-offered undergraduate courses and modules in computational science. The repository is to be a collection from various pioneering programs throughout the country, which together would cover the entire field of computational science at various levels. Once in place, other schools could adopt these courses as a way of including modern computation and multidisciplinary studies in their curricula. The conversion of computational physics courses in the University of Wisconsin's system has already begun.
Jaime Taylor and B. Alex King III (Austin Peay State University; TaylorJR@apsu.edu and firstname.lastname@example.org), "Matlab: A Fast Development Environment for Computational Methods"
The focus of a computational methods course in physics is computational methods, rather then developing programming skills. Matlab is an ideal solution. The Matlab environment allows for quick implementation of computational methods discussed in the classroom. In this environment, students are less likely to lose sight of the concepts of the computational techniques because they spend less time focused on syntax or the many other programming issues that arise from working with a lower-level language.
Martin Johnston and Adam Green (University of Saint Thomas; email@example.com), "Introducing Students to LabVIEW without Tying Up the Semester"
Like many other worthwhile endeavors, programming in LabVIEW has a significant learning curve, but by integrating the language into sophomore- and junior-level courses, the authors have successfully introduced students to using it. Students' first introduction to LabVIEW is in the modern physics laboratory, where they use prewritten applications to collect data. Formal development of LabVIEW programming skills occurs in the experimental methods course and is further refined in an optics course. Student response to the spread-out approach to learning the language has been excellent. Future plans include debuting LabVIEW in the introductory classical physics sequence.
Wolfgang Christian (Davidson University; firstname.lastname@example.org), Harvey Gould (Clark University; email@example.com), and Jan Tobochnik (Kalamazoo College; firstname.lastname@example.org), "The Open Source Physics Library and Curricular Material"
The Open Source Physics (OSP) project has developed a Java library and an easy-to-use XML vocabulary for the development of physics software and the distribution of Web-based curricular material. This technology lets instructors associate a simulation, a set of initial conditions, and a narrative into a topical unit as well as combine topical units into a curriculum module that users can further adapt and customize. The authors show how they've used this framework to write and organize course material from computational physics, classical mechanics, statistical physics, and quantum mechanics.
Richard Martin (Illinois State University; email@example.com), "The Computer Physics Bachelor's Degree Program at Illinois State University"
After integrating computational assignments into physics-major courses in the 1990s, Illinois State University's Department of Physics launched a degree sequence in computational physics in 1999. The goals were multifold: to satisfy student requests, offer an alternative to students not interested in a traditional physics career, and enhance physics-major recruitment by exploiting a perceived student demand for computer-oriented careers. This poster describes the program's development and structure, updates some preliminary assessment results, and offers practical suggestions for other departments interested in going the computational physics route.
David Cook (Lawrence University; firstname.lastname@example.org), "Computation in the Lawrence Physics Curriculum"
The physics curriculum at Lawrence University introduces online data acquisition, statistical data analysis, and curve fitting in its introductory laboratories; includes the required course "Computational Mechanics," in which sophomores learn to use computer-based symbolic, numerical, and visualization tools to address intermediate mechanics problems involving ordinary differential equations, integrals, eigenvalues, and eigenvectors; and offers the junior-senior elective "Computational Physics," which focuses on numerical solution of wave, diffusion, and Laplace equations via finite difference and finite element methods.
Bruce Sherwood and Ruth Chabay (North Carolina State University; Bruce_Sherwood@ncsu.edu and Ruth_Chabay@ncsu.edu), "Computing in the Introductory Physics Course"
VPython is a programming language that is unusually easy to learn, produces navigable 3D animations as a side effect of physics computations, and supports full vector calculations. The high speed of current computers makes sophisticated numerical analysis techniques unnecessary. Students can use simple first-order Euler integration, decreasing the step size until system behavior no longer changes. In mechanics, iterative application of the momentum principle gives students a sense of the time-evolution character of Newton's second law, which is usually missing from standard courses. This poster studies the impact of introducing computational physics topics in an introductory course.
Louis F. DeChiaro (Richard Stockton College of New Jersey; Louis.DeChiaro@stockton.edu), "Helping Undergraduates Climb the Computational Physics Learning Curve"
Today's science graduates need considerably more computational skills than their predecessors if they wish to strategically differentiate themselves from their competitors in nearly any technical profession. This poster summarizes the author's strategies to identify the optimal combination of software environments and computational physics topics. The goals are, first, to ensure that students acquire a minimum level of programming competence and, second, to build on this and give a firm foundation that will prepare them either for graduate study or for immediate entry into the job market.
Gary Pajer and Alexander Grushow (Rider University; email@example.com), "Computation in the Physics Curriculum at Rider University"
Rider University is a four-year comprehensive with a small active science faculty that exposes its physics students to the "new pyramid" of theory, experiment, and computation, with all three having equal status. For most courses, the faculty use a cross-platform, open-source environment consisting of the Python language and its extensions VPython (3D graphics), numpy (numerical extensions), scipy (special functions, ode's), and matplotlib (graphing). The Python computing environment provides all the functionality of popular commercial products at none of the cost.
Seamus Lagan (Whittier College; firstname.lastname@example.org), "Computational Oscillations and Waves at the Sophomore Level"
Whittier College created a sophomore-year computational course required for all its physics majors; it introduces the physics of oscillatory systems and wave motion yet serves as a vehicle for learning how to use Maple and the Interactive Data Language. Students learn how to perform tasks such as evaluating integrals, solving differential equations, solving simultaneous equations, plotting functions or data, creating series expansions of functions, finding roots of equations, and working with complex numbers.