Mims's Bits

Robots Make Computer Science Fun Again

Students who program robots are more likely to stick with their computer science curriculum.

Christopher Mims 07/19/2010

  • 3 Comments

The long-term trend for interest in computer science at the university level is relatively bleak. As the graph at above makes apparent, interest has declined precipitously since the 2001 bursting of the dot-com bubble, leading to something of an existential crisis in the field of computer science instruction.

The latest survey on the subject, which charts 2007-08 data, showed a widely-reported up-tick in enrollment of 8 percent, which is great for a year-on-year change, but neglects the long-term trend.

One has to wonder whether it's the very ubiquity of computers that has made them uninteresting to students--note the spike of interest in the early 80's, when the advent of personal computers slaked a pent-up demand for access to the instruments that everyone believed would define the future.

Robots, in contrast, are still rare in our everyday lives--plus, they're the furthest thing from remote and abstract. So goes the reasoning behind a new effort to get them into classrooms, described earlier this month in a paper by Tom Lauwers and Illah Nourbakhsh, in which they unveiled the Finch.

The Finch is cheap, simple and avoids the major complexifying factor most previous efforts to add robots to computer science have encountered: namely, robots break, and debugging physical objects is a headache students don't need.

The results were profound: retention rates for the 2009 computer science classes in which the Finch was used (shown below, in red) increased by 25 percent.

And why not? The Finch sounds like exactly the kind of Maker project everyone's inner geek cries out for:

The Finch can express motion through a differential drive system, light through a color- programmable LED, and sound through a beeper and using computer speakers. Similarly, it can sense light levels through two photoresistors, temperature through a thermistor, distance traveled through two wheel encoders, obstacles placed in front of it, and its orientation in three dimensional space through an accelerometer [...] In addition to these hardware-based capabilities, the accompanying software allows students to easily have the Finch speak or play songs over computer speakers, read real-time data from internet RSS feeds, and react to video from computer webcams.

Follow Christopher Mims on Twitter, or contact him via email.

Print

Close Comments

To comment, please sign in or register

Forgot my password

Mapou

356 Comments

  • 569 Days Ago
  • 07/19/2010

Humans Love to Play with Toys

Interesting article. Back in the 80s, there was a challenge. People wanted to play with the new machines until they could master them. New operating systems, new programming languages and countless applications needed to be created. It was an untamed frontier, perfect for making new discoveries. Likewise, robots are an untamed frontier. They must become truly intelligent in order to take over most tasks from humans. Problem is, nobody knows how to do that. The opportunity for a major discovery in AI is tempting. Some of us like to dream about the possibilities.

Another reason for student interest in robotics is that programming a robot is more like tinkering with a toy. Tinkering is part of human nature. I believe that the ability to tinker with software objects will be the most powerful part of future computer programming environments. This is why I hate all computer programming languages; they are not fun to play with and they don't provide for easy 'what-if' type tinkering. In my opinion, unless your development environment feels like a toy, it's no good.

Reply

jesup

17 Comments

  • 567 Days Ago
  • 07/21/2010

Re: Humans Love to Play with Toys

You should check out Scratch (http://scratch.mit.edu/)  Not perfect, but an interesting new take on programming.  There's now an App builder for Android based on Scratch.

Reply

LJB

1 Comment

  • 568 Days Ago
  • 07/20/2010

Not a scientific study

If you read the article you point to, you'll see that retention was only improved for one of the two pilot semesters. When comparing the entire year to four previous years, retention is not significantly different. We have no idea how many students were involved either, so there's no way of making judgments about the statistical test. In addition, the title "robots make CS fun again" has very limited evidence in the study. The authors write "Though we don’t have comparable data of students in regular programming courses at CCAC, we believe that it is rare in such courses for students to write programs purely for enjoyment."  Their opinion may be right or wrong. Novelty might be the issue here, too, or community college kids who have a toy in hand. Indeed, there may be many variables at play here that led to greater retention in one of the two semesters, variables not studied by the authors. For example, what about the course redesign? students' feeling of importance by being surveyed before, during, and after the semester, including after each assignment? were the students similar in that one semester or could there be something different about the students who took the class? there's no way to answer any of these, because this was not a rigorous study.

Retention in a single course is also quite different from retention in a discipline or career.

The worst thing about making unsupported claims of an expensive educational program is that people may spend a lot of money to replicate it. In this case, faculty may go out and start buying and distributing robots as a way of retaining students in CS, a very important issue for our economic future. Or funding agencies like Ed or NSF might supply grant money for buying robots. I prefer evidence-based programs for my tax dollars.

Reply

Bio

Christopher Mims is a journalist who covers technology and science for just about everybody.

Subscribe to the Mims's Bits RSS Feed

Advertisement
Advertisement

Facebook

Advertisement