Special Issues » Technology Issue

Developing a laptop for the children of the developing world

by

comment
Bridging the divide | Goodwill in Raleigh | Teens teaching teens | One Laptop per Child | e-NC Authority | Connecting nonprofits
Greg DeKoenigsberg of Red Hat with a beta model of the XO, a laptop for kids in the developing world - PHOTO BY DEREK ANDERSON
  • Photo by Derek Anderson
  • Greg DeKoenigsberg of Red Hat with a beta model of the XO, a laptop for kids in the developing world

In the conference room at the Raleigh headquarters of Red Hat, approximately 80 members of the Triangle Linux Users Group crowd around a small laptop computer. Its keyboard is built for a child's fingers, and the bright green-and-white device looks like a high-end Fisher-Price toy. But this machine is designed to be cheap, for distribution by the millions in the developing world. After four years, the One Laptop Per Child project has produced a beta model for user testing. And Greg DeKoenigsberg, Red Hat's community development organizer, is one of the few people who has one.

As the geeks take turns inspecting the laptop's much touted innovative, intuitive design, DeKoenigsberg begins to make his pitch. "I'm going to turn you all into XO developers today," he says.

Nicknamed the XO after the icon that appears on the device (it's supposed to represent a kid, with the X as the body and the O as the head), the OLPC laptop is being designed at Massachusetts Institute of Technology's Media Lab. Its creators hope it will serve as a textbook, a library, a telephone (there is a built-in speaker and microphone), a camera (that's built in, too), even a musical instrument, as well as a link to the Internet. The countries who have signed on so far to buy the units are Rwanda, Uruguay, Libya, Brazil and Nigeria. The project is funded through cash and in-kind contributions of more than $2 million each from companies that include Google, Nortel, Red Hat and Advanced Micro Devices.

The mastermind of the OLPC project, Nicholas Negroponte, chose Red Hat over Microsoft because Red Hat's software is open source Linux—meaning anybody can look at the code and improve it.

That's exactly what DeKoenigsberg hopes the TriLUG members will do. While the hardware glitches are getting worked out in Boston, he's inviting open source enthusiasts to take a crack at Sugar, the user interface being designed to run on the OLPC. His first public pitch is to Red Hat's hometown crowd.

Red Hat is a for-profit company, so what's their motivation in undertaking this massive not-for-profit job? "Millions and millions of kids knowing Linux before they know Windows," DeKoenigsberg says to laughter from the TriLUGs. (Microsoft jokes play well to this group.) "That's pretty much it."

And what will motivate Linux users to work on Sugar? When it comes to most open-source projects, DeKoenigsberg says, "it's a fairly common misconception in the world at large that people get involved with open source software primarily out of some altruistic need. There may be elements of that, but for the most part, people get involved with open source projects because they've got to get something done." The sense of an ethical obligation to give back to the community often comes later.

"One Laptop isn't necessarily like that," he says. "As cool as this little thing is, it's not going to be as functional as the machines you and I use every day." So in this case, the urge is not to scratch an itch. It's to become part of something really, really big. "How many opportunities will we have in our lives to get involved with something that can potentially affect so many lives on such a grand scale with an expertise that not a lot of people in the world have?"

That appeals to Daniel Bartholomew, one of the geeks in the audience. Most software developers don't work on commercial applications that the general public has heard of; they work on in-house software, maybe a database that only one company uses. "That's all they're working on, these little no-name apps that no one knows about and no one will ever know about," he says. "This is a way for a software developer to do something that gets them recognition outside of their own company. It's a way to get recognition for what you do that you probably wouldn't get any other way."

Sugar also presents some unique challenges that are likely to entice developers. The XO is not designed like the computers most of us use, and neither is the software. Negroponte is adamant that educational software should be geared toward exploration and communication, not toward solitary office tasks. So Sugar is being designed as a new paradigm: Instead of a "desktop," there's a "neighborhood" that displays other users who are connected through a wireless mesh network. Instead of "applications," there are "activities." One of the most innovative things about the OLPC is that it's designed to be used in a collaborative way. Unless an activity is set to private, peers can peek onto one another's screens and join in. They can even copy that activity onto their own XO. Collaboration is the default.

"It's a design goal that everyone who writes an activity is asked to bear in mind," DeKoenigsberg says. "What is it about the activity that you're writing that is inherently shareable? Figure that out before you write the activity." The most popular activity so far, he says, is called TamTam, "a sort of simpler version of Apple's GarageBand" that allows kids to create and edit various sounds to make their own music. "I hope that scares the music industry to death," DeKoenigsberg says, "because it ought to."

Another design problem is scaling down the software to make it run on very little electricity, with very little processing power. The OLPC doesn't have a hard drive; it has 512 megabytes of memory in a flash drive, which means no moving parts inside. It has a 366-megahertz processor, which won't come near the speed or functionality of the computers most of us use. But it will be far more functional for its target audience—the children of the developing world. It can be set to use as little as 1 watt of power, with a screen that can be seen even in direct sunlight. MIT is working on ways to make the OLPC "kid-powered"—replacing electricity with a foot pedal or yo-yo device, to deliver 10 minutes of working time per 1 minute of power-pumping.

"This is beta hardware running pre-alpha software," DeKoenigsberg warns several times—in other words, it's a very rough draft. And unfortunately, at this point, volunteer developers won't have an actual XO to work on; they'll have to create a "virtual environment" on their own Linux-running machines.

Even as he's recruiting volunteers, DeKoenigsberg is frank with them about the criticism and skepticism the OLPC project has met with. The OLPC can't address situations of extreme poverty, in which people live on less than $1 per day, he says. "There are other priorities for those people"—namely food and medicine. But the hope is that it will help with the "information famine" faced by communities with no libraries, no books, no telephones and no real way to connect to the world and improve their prospects.

Questions from the TriLUG members are mostly about the hardware and the massive logistics of the project as a whole. How are millions of laptops going to be distributed? How will the organizers prevent them from being stolen and sold on the black market? Why is the notoriously corrupt government of Nigeria one of the first partner nations? How will kids be protected from Internet porn—or from adults who use the laptops to exploit children?

"These are all fair questions," DeKoenigsberg says. "Terrible things are going to happen when these things go out into the world, you know? I mean, that's just a fact. The question is, are those terrible things that are going to happen enough of a deterrent to prevent us from getting all the potential benefits of having kids networked?

"You have to measure this stuff against the potential," he says. "It's not a reason not to do it; it's a reason to be careful."

To find out more about software development for the OLPC project, visit fedoraproject.org/wiki/OLPC.

Add a comment

Quantcast