TITLE: Science Students Should Learn How to Program, and Do Research AUTHOR: Eugene Wallingford DATE: August 23, 2015 10:12 AM DESC: ----- BODY: Physicist, science blogger, and pop science author Chad Orzel offered some advice for prospective science students in a post on his Forbes blog last week. Among other things, he suggests that science students learn to program. Orzel is among many physics profs who integrate computer simulations into their introductory courses, using the Matter and Interactions curriculum (which you may recall reading about here in a post from 2007). I like the way Orzel explains the approach to his students:
When we start doing programming, I tell students that this matters because there are only about a dozen problems in physics that you can readily solve exactly with pencil and paper, and many of them are not that interesting. And that goes double, maybe triple for engineering, where you can't get away with the simplifying spherical-cow approximations we're so fond of in physics. Any really interesting problem in any technical field is going to require some numerical simulation, and the sooner you learn to do that, the better.
This advice complements Astrachan's Law and its variants, which assert that we should not ask students to write a program if they can do the task by hand. Conversely, if they can't solve their problems by hand, then they should get comfortable writing programs that can. (Actually, that's the contrapositive of Astrachan, but "contrapositively" doesn't sound as good.) Programming is a medium for scientists, just as math is, and it becomes more important as they try to solve more challenging problems. Orzel and Astrachan both know that the best way to learn to program is to have a problem you need a computer to solve. Curricula such as Matter and Interactions draw on this motivation and integrate computing directly into science courses. This is good news for us in computer science. Some of the students who learn how to program in their science courses find that they like it and want to learn more. We have just the courses they need to go deeper. I concur with all five of Orzel's suggestions for prospective science students. They apply as well to computer science students as to those interested in the physical sciences. When I meet with prospective CS students and their families, I emphasize especially that students should get involved in research. Here is Orzel's take:
While you might think you love science based on your experience in classes, classwork is a pale imitation of actual science. One of my colleagues at Williams used a phrase that I love, and quote all the time, saying that "the hardest thing to teach new research students is that this is not a three-hour lab."
CS students can get involved in empirical research, but they also have the ability to write their own programs to explore their own ideas and interests. The world of open source software enables them to engage the discipline in ways that preceding generations could only have dreamed of. By doing empirical CS research with a professor or working on substantial programs that have users other than the creators, students can find out what computer science is really about -- and find out what they want to devote their lives to. As Orzel points out, this is one of the ways in which small colleges are great for science students: undergrads can more readily become involved in research with their professors. This advantage extends to smaller public universities, too. In the past year, we have had undergrads do some challenging work on bioinformatics algorithms, motion virtual manipulatives, and system security. These students are having a qualitatively different learning experience than students who are only taking courses, and it is an experience that is open to all undergrad students in CS and the other sciences here. -----