Note: This article is over a year old and information contained in it may no longer be accurate. Please use the contact information in the lower-left corner to verify any information in this article.

Setting the pace for a new computing education

By Mara Kumagai Fink '11
November 16, 2009
St. Olaf Associate Professor of Computer Science Dick Brown is leading the way nationally in creating a new computer science curriculum. "The whole computing industry needs this kind of change," he says.

When the giants of the computer industry — including Google and Intel — put out a call last year for a new style of computing education, St. Olaf Associate Professor of Computer Science Dick Brown answered. His quick response earned him a $200,000 grant from the National Science Foundation to be one of the first to create materials for a new curriculum.

“The whole computing industry needs this kind of change, so it is a really big deal. We feel fortunate that we’re the ones who got the grant,” Brown says.

This new curriculum has the potential to transform the way professors teach computer science in colleges across the country. The reason for the change in teaching is that advancements in the way professors teach computer science have not kept pace with the advancements made in computers over the last few years.

Too fast
Brown says computers have actually doubled their speed every 18 months since 1965. Up until 2005 this wasn’t a problem because the software for computers also increased speed. However, in 2005 programmers were unable to make computers any faster without adding more power to them. As they added more power, the software engineers did not adapt their software to this new arrangement. Therefore, although computers were getting faster, the software wasn’t.

This is where the students come in. From here on out, it is critical that computer science students learn how to work with “multicore” computers. Previously, computing for multicore computers — known as “parallel computing” or “parallelism” — was only taught in a select few upper-level courses. However, because it is so important now, it needs to be taught to all computer science students.

Timothy Yates '12, seated at the keyboard of a cluster of multicore computers, is one of three students working on Brown's project. St. Olaf's new parallel computing room has the capacity to house one of the world's 500 most powerful computing systems.

“Multicore computers are changing the way computing is done, and we get the opportunity to help colleges learn how to teach computer science using parallelism,” Brown says.

Brown’s new curriculum will translate the advanced concepts taught in these upper-level classes to make them easily digestible for beginning computer science students. He and Macalester College Associate Professor of Computer Science Libby Shoop, along with the help of three St. Olaf students, are working to create a method of teaching this new computing. This fall St. Olaf students in lower-level computer classes started using parts of the curriculum, and Macalester is remotely using the computers at St. Olaf to implement the software for their classes.

However, it’s not just beginning students who are unfamiliar with the concepts; many college professors are not well versed in parallelism either. This means that the curriculum also must give professors who don’t have a background in this kind of computing the tools they need to teach it. The teaching tools Brown is developing will be self-contained, so there is no need for students to buy an entirely new textbook — instead, the curriculum can just be added.

A nationwide program
Before the grant ends in January 2012, Brown hopes to have the curriculum up and running at St. Olaf and Macalester, as well as be prepared to launch the nationwide program. He says they’ll focus on liberal arts schools first, although the same technology can be just as easily used at larger institutions. In the spring, Brown and his students will present a workshop on teaching parallelism at a conference of the Special Interest Group on Computer Science Education Technical Symposium.

“Our strategy is to focus on 4-year institutions first, because the nation's 3,000-plus liberal arts colleges include lots of students who won't otherwise have quick ways to start learning parallelism. At a major university, undergraduate courses may be influenced by graduate research in parallel computing on campus. The way we see it, smaller schools represent a bigger opportunity for achieving our main goal,” he says.

Also read about the Computer Science Department's MistRider Beowulf cluster ...

Contact David Gonnerman at 507-786-3315 or gonnermd@stolaf.edu.