"Walk into a middle school classroom and ask 'Who wants to be a computer programmer?' and hardly any hands go up," says Caitlin Kelleher, a post-doctoral researcher in Computer Science & Human Computer Interaction at Carnegie Mellon University. Delivered from the stage in a crowded hotel ballroom, her words sound like the beginning of a joke.
But Kelleher's point is anything but funny. While the number of computer-science grads in India and China is on the rise, the U.S. has seen a 50% drop in CS college majors between 2000 and 2005 (according to a May, 2005, UCLA study), casting doubt over whether the U.S. will continue to be an innovation leader in the global economy.
Kelleher's venue was the Serious Games Summit, held Oct. 30-31 in Washington, D.C. It's an annual gathering of game designers and developers who are joined in researching how their "entertainment" medium can be leveraged as a training tool in areas ranging from the military to health care, education, activism, and the corporate world.
But, according to Kelleher, the situation isn't quite as grim as it might appear. "Walk into the same classroom and ask, 'Who wants to make Pixar movies?' and it's a different story," Kelleher continues, delivering the punch-line that offers a possible solution: Entice kids to study programming by focusing on the "products" of computer science that they're already familiar with—animated movies, video games, and social-networking sites—rather than the process of writing code. Not only does this make the discipline seem less tedious, it gives programming an element of cool.
Entertainment meets academics—this is the thinking behind Kelleher's research. She's part of a team working with CMU professor Randy Pausch on the next generation of Alice—software designed to introduce children to programming through simple exercises in digital animation. In development since the mid-1990s, the tool is based on early 3D-modeling software. Kelleher shows the current versions of Alice, one for middle and high school students and a more involved one for college students. At the moment, they have rather clunky onscreen graphics, though that will soon change.
She shares the podium with Steve Seabolt, vice-president of university and marketing education at Electronic Arts (ERTS), who's there to announce that EA is pouring $300,000 into the Alice project. Moreover, it has licensed the software from the popular Sims franchise, which has sold 70 million copies, to CMU, allowing the researchers to jazz up the Alice visuals and user experience.
Alice isn't the first effort to make computer programming fun for kids. Starting in the 1960s, MIT researchers began developing the children's programming language, LOGO. Now the Lifelong Kindergarten group at MIT's Media Lab is developing The Scratch Project, in collaboration with the KIDS research group at UCLA's Graduate School of Education & Information Studies.
The Scratch Project's toolkit (based on open-source programming language Squeak) allows kids to create animations and computer games by selecting parameters from a point-and-click menu. The tool should be available for public release by the end of 2006.
Researchers at the Media Lab's Lifelong Kindergarten project also worked with Danish toymaker Lego to create "programmable bricks," or building blocks equipped with sensors and other chip-driven tools. The research led to Lego's Mindstorms robot-building kits and the PicoCricket, now used by educators and inventors (see BusinessWeek.com, 9/7/06, "Invasion of the DIY Robots").
In the commercial arena, Massachusetts-based programmer Igor Kholodv released c-jump, an old-school board game for kids 11 and older that requires players to use programming commands such as "if (x = =1)…" to advance their game pieces.
Also available is ToonTalk, a computer-programming PC game for younger children developed by Oxford University researcher Ken Kahn that allows young kids to create simple animations.
Unlike its predecessors, CMU's joint project with EA is a collaboration between academic researchers and a video-game industry giant that pushes the envelope beyond mere financial or equipment sponsorship. Other traditional sponsors of Alice include Intel (INTC) and Microsoft (MSFT). "Randy Pausch says, 'It's like Coke giving away its formula, or Disney giving away Mickey Mouse,'" Kelleher says after the conference.
In a real-time demo of the original, pre-Sims Alice software at the Serious Games Summit, Kelleher chooses three cartoony avatars from a drop-down menu—a boy, a girl, and a stern-looking woman with glasses and her hair in topknot named the Lunch Lady. By clicking on drop-down commands—driven by about 10 or 15 lines of plain-English code—no "if (x ==1)…"—Kelleher programs the simple actions that cause an avatar to walk or turn.
Learning the Language
After the demo, the audience is respectful but circumspect: Is Alice really teaching computer programming? One man suggests the process, especially when the Sims characters are used, is more akin to Machinima, the name for a genre of movies created using video-game technologies.
After all, Alice doesn't require students to actually write Java or Python, only to string together pre-written commands from the menu and occasionally type in parameters, such as how much a character's head should tilt.
"[Alice] just changes the mechanics of how people write programs. It's still programming," responds Kelleher. "The idea is to familiarize kids with what programming is and to introduce them to the key idea of choosing parameters."
The CMU researchers have data that suggest the current, pre-Sims version of Alice has helped unmotivated computer programming students to stay committed. In 2004, the Alice team tracked students at St. Joseph's in Philadelphia and Ithaca College in New York. The schools were chosen because their student bodies generally had median, rather than high, SAT scores and represented a more "average" academic ability.
At each school, the researchers chose two groups of computer-science students deemed likely to drop out of the major. One group learned Java using traditional book and lecture methods. This group (at both schools) had an average grade of C, and only 47% continued as computer-science majors. The second group took a course in Alice, and then took the same Java class as their peers.
The second group's average grade in the Java class was B, and 88% stuck with the major. It's logical to assume that those numbers will be even higher after the slicker Sims graphics are incorporated.
The new version of Alice will be distributed free online beginning in the spring of 2008. EA's Seabolt casts the company's Alice involvement as an "extension of the Sims franchise." He adds that "in The Sims, we have the most widely recognized [PC-game] brand. If we can make math and science relevant by connecting those skills to their favorite game—and their desire to build games—everyone wins."
But the real return for EA will likely come years from now in the form of large pool of software programmers for hire. At least, that's the hope of EA and the CMU researchers. And given the plummeting number of U.S. computer-science students, it should be the hope of everyone in the business world.