Wired:The headlines this week haven't been cheery for the One Laptop Per Child (OLPC) project. Masterminded by Massachusetts Institute of Technology Media Lab guru Nicholas Negroponte (who long-time tech watchers will remember as the advocate of the failed Media Lab Europe), the project has never ceased to be controversial.
But even its strongest advocates must have winced at the double-whammy of a New York Timesheadline that hinted at "the demise of One Laptop per Child" and a review of the group's hardware in the Economist that was titled, "One Clunky Laptop Per Child", and subtitled, "Great idea. Shame about the mediocre computer".
The OLPC project is so simple that I think most of us assume it must be wrongheaded from the get-go.
Negroponte's vision was to produce a $100 (€68.20) laptop, designed for schoolchildren in the developing world, to be given to them by those countries' governments to aid the next generation in leapfrogging past their current educational limitations - and perhaps even those of their contemporaries in the developed world.
Almost everyone except Negroponte has a kneejerk problem with this plan. Those who love technology hate the involvement of governments (Negroponte refuses to sell the XO-1 laptops on the open market; eager US geeks can only get their hands on them by donating enough money to buy one for a developing country's schoolchild). Intel took great offence at an educationalist entering into potential new territories and created its own competing product, the Schoolmate.
That New York Times headline described the collapse of the uneasy truce between the OLPC and free-marketeering Intel, which broke down this week when an Intel salesman persuaded a South American country to dump the OLPC for the commercial alternative.
Meanwhile, those who distrust technology as a panacea find it shocking that children who can barely afford textbooks are being handed such overpowered tools.
But perhaps the harshest criticism this week has come from Negroponte's fellow technologists, who have actually got the machines in their mitts.
"A nightmare," said one, reeling at the bugs and slowness of the first generation of the XO-1's operating system. "Confusing," said others, thrown by its designers eschewing both traditional hardware (adults find it impossible to open the laptop, let alone type on its kid-sized keyboard), and software (the operating system is neither Windows nor Mac, but a radically rethought Linux desktop).
Playing around with the XO-1, it's obvious that the software is half-done. The laptop does not hibernate properly, crippling its incredible potential battery life. Its innovative track pad, which can work as a stylus-sensitive graphics tablet, is only used as a simple mouse pointer. It's frustratingly hard to get books to read while running the machine's amazing "e-ink" monochrome mode, which gives children the same resolution and readability in direct sunlight as cutting-edge e-book readers.
But those who criticise this first batch of code forget two things. First, software is eminently updatable, and the OLPC architects had the sense to build in simple ways to update the software. Second, they forget (or perhaps were not around to remember) quite how execrable the first IBM PC software was, or indeed the first versions of MacOS X.
All these abilities lie latent in these boxes, and will be brought out with the hopefully improved software that can now be written for it. Because of Negroponte's demand that the hardware and software be open, you don't need an OLPC XO-1 to write or run OLPC software.
And the OLPC is (supposedly) built to last. It's as near indestructible as possible for the price. The idea is that the OLPC should last 10 years, amortising its costs to less than $20 (€13.60) a year. That's the cost most of its targeted developing countries spend on textbooks for their schoolchildren. If the OLPC works as minimally as intended, as a dirt-cheap, unusually hardy e-book reader, it will have replaced a decade's worth of fragile, soon outdated educational literature with the digital equivalent.
Staring at the bright green, plastic XO-1, tiny yet tough, I admit to finding it more unnerving than contemptible. In some ways, it's because I can foresee so many ways in which it might fail. Children and teachers may not understand it (a failure of the teaching philosophy of learning by exploration that underlies Negroponte's plans). The audacity of its hardware decisions may well have exposed it to a Titanic-like fundamental flaw (will those screens shatter in the first Peruvian frost, or fade to black over six months' use in the equatorial sun?).
But what makes me more nervous is if it succeeds. Playing around with eToys, a seemingly innocent game on the laptop, I stumbled, as a smart child would, on a toolkit of amazing, bogglingly advanced utilities: frequency analysis systems; a voice synthesiser, a video editor and a gesture recogniser.
Just like the XO-1 itself, eToys has the veneer of a child's toy, but hides a machine as powerful and as repurposable as any grown-up's laptop - or Bill Gates' development environment. This computer rewards exploration, all right: if you're a Gates-style genius.
And what would a genius kid, with 21st-century tools, stuck on the wrong end of our violently lopsided global economy, do with this? Bring wealth to his or her country? Exploit his or her less smart fellow students for their underused CPU power? Or use this toy as ticket out of the global ghetto to locate, and violently and perhaps rightfully seize the other secret tools of the West that have been denied them for so long?
