On a desk in a messy office on the eighth floor of a building opposite the Massachusetts Institute of Technology sits a circuit board that might just transform education for millions of kids around the world.
The board is the first prototype hardware for the ambitious One Laptop Per Child (OLPC) project, a nonprofit effort announced in January 2005 and launched by a group of MIT Media Laboratory faculty. Led by Nicholas Negroponte, chairman of OLPC and a co-founder of the MIT Media Lab, the project seeks to develop a $100 laptop computer for use by children in their studies. The machine, targeted at children in developing nations, will be offered in bulk to governments and other organizations.
The prototype's arrival in the United States from Taiwan early this month represented a milestone. However, the project is running late, and several obstacles remain before computers get into kids' hands. (And we're not even talking about Microsoft's Bill Gates recently irking Negroponte by publicly mocking the $100 laptop idea.)
The price of materials is one issue. The computer will use energy-saving flash memory in place of a hard-disk drive. Negroponte originally called for 1GB of memory, but this has been cut in half. Analysts were skeptical that the team could afford 1GB of memory, given the machine's target price.
"I guess they were right," says Walter Bender, president of software and content at OLPC. Bender joined the project earlier this year after serving as executive director of MIT's Media Lab from September 2000 until January 2006.
The project, which has changed its name to de-emphasize the $100 target price, says it expects the initial machines will cost about $130 and will come down to about $80 in a few years.
Other changes include a cut in processor performance. Initially the units were to feature a 500-MHz Geode processor from project supporter Advanced Micro Devices (AMD), but the current specification calls for a 400-MHz chip. Bender says this slower chip was easier to obtain in quantity.
Perhaps the most significant change is in the screen. A projection screen was to have been used but a type of LCD still under development that can be switched between a low-resolution color mode and high-resolution monochrome mode is planned. Bender says the new screen will be "kick-ass, cheap, superefficient and beautiful."
Getting the low-cost laptop into the hands of kids in developing countries won't mean anything unless the machine can be powered. That could be a problem in remote villages where not every home has electricity. The current design calls for the laptop to have a conventional power jack.
If a home has electricity it can be connected via an adapter, like a normal laptop, but if there's no power supply it can be hooked-up to an alternative energy source. This could be a hand-crank or foot-operated generator or something like a solar panel array, Bender says.
A year ago, Negroponte said he expected to receive the first order in June 2005 and to have gathered orders for about 6 million machines by the end of that year. But the OLPC is not officially accepting orders until it has completed development of the prototype and set production plans, Bender says.
Bender recently showed the board in Cambridge, Mass., running a version of Red Hat's Fedora Linux. "What you're seeing is close to the final design," he says. "Very little is going to change on this board."
He expects to finalize the design late this year, and to begin production in the first quarter of 2007. A year ago, Negroponte said production would begin in 2006.
Still, the arrival of the first prototype, the product of several months of work by Quanta Computer, marks a strong step forward for a project that many doubted would make it this far.
The Taiwanese company is the world's largest maker of laptop computers and one of a number of supporters that has pledged to assist OLPC.
Other supporting companies include AMD, Brightstar, Google, News Corp., Nortel and Red Hat, and wireless network chip maker Marvell Technology Group recently joined the list. Marvell will work on Wi-Fi networking for the computer.
OLPC isn't the first effort to make computers significantly less expensive. Oracle, Sun and others have pushed the thin-client concept, and recently Intel partnered with a Mexican telecom company to get inexpensive PCs into the hands of first-time computer users in developing countries.