Michael Diodato, writing for the HIR, finds the OLPC technology disappointing - the much touted “mesh network” is unlikely to be useful in rural settings where people may not be within range of each other, and certainly not within range of an internet connection; the crank is unlikely to provide sufficient power for ordinary use, including use of the wireless network; and the restricted configuration requires proprietary software with limited applicability elsewhere.

It is not technology, however, that kills the OLPC’s value proposition - unfortunately for Negroponte, the HIR finds the OLPC short even on learning. Amongst the practical problems mentioned are the challenges of distributing - equitably - 1.2 million computers, and the cost - to families - of keeping the laptop in and their child out of work: “Considering the opportunity cost of keeping the laptop and the option selling it for approximately $100, the family is likely to choose the latter.”

Testing the “Learning” Proposition

There is one final nail in the OLPC coffin: the efficacy of learning with the OLPC - the basis of Negorponte’s argument - has never been tested. Negroponte, himself, is a serious critic of any “pilot tests”, claiming at the IADB that “to do a pilot project is ridiculous,” and suggesting at TED 2006 that “this is not something you have to test, the days of pilot projects are over.” That is a rather odd stand for a “scientist.” As Robert Kozma points out in his piece, “empirical data are the sin qua non of both scientific research and engineering design.”

The OLPC is gaining serious traction now. Recently, Intel joined the OLPC bandwagon, perhaps as a hedge against its own Classmate PC project - which has been shipping in volume, at prices of around USD 225, since March. And the laptop may finally ship this year, maybe as early as September. But for countries - particularly developing ones - to proceed without testing the “learning” hypothesis is irresponsible.

Countries that adopt OLPC without pilot testing are in effect conducting a nation-wide experiment. It is a roll of the dice. If the OLPC predictions are correct, the nation and its children win.

On the other hand, if the grand national experiment fails, it is developing countries and their children that are least able to manage the consequences of this failure or recover from the expended costs.

Alternatives: One Laptop Per Family? Basic Education?

Neither of these criticisms is dogmatic, which makes them all the more credible. Technology is a good thing and its use in education to be encouraged. But how is that to be best achieved? And what is the goal of technology in education, anyway?

The HIR, for instance, offers some alternative distrubtion models that tackle some of the social shortcomings of the OLPC. How about setting up OLPC cafés (shared laptops in communities), or giving laptops not to each child, but to each family?

Fewer laptops would be needed [and] more money could be spent on individual laptops at the same total cost. …It also reduces the chance of a family selling a laptop. If the entire family is benefiting from the machine, as opposed to just the children, then the family may decide to keep the machine. If software is designed that would enable the parents to use the laptop for their own purposes, perhaps for growing crops in an agricultural area, then the parents would have more of a reason to keep the machine.

But in the end, there may be an even simpler alternative to the OLPC - so simple that nobody really thought of it. Or perhaps that is why the OLPC is the preferred option, being so much sexier and providing much better photo ops for the world’s politicians and technologists:

One simple solution is to invest money into traditional methods of education as opposed to those requiring technology. Although one of the main purposes of these advances is to allow all children to access the Internet, such a goal is not essential to learning…US$250 million being spent by Libya could instead buy millions of books or pay salaries for thousands of teachers.

When looking at the use of technology in education, it is difficult to dispute that it is helpful. However, with only a limited amount of resources that can be devoted to a developing country, it is imperative these resources are distributed in the most efficient way possible.

In the end, India was right to reject the OLPC. Whatever the real or imagined reasons for the rejection, the money is better spent elsewhere and the project is, after all “pedagogically suspect”.

In choosing not to join the loud debate on Technology for Development I’ve held that technology for its own sake is useless, a view often ignored by western commentators, with the OLPC a perfect example. In the context of development, what is useful about technology is not what it is but rather what it can do.

Why the Digital Divide

I have, however, given substantial attention to why technology hasn’t reached the poor, and why the proverbial ‘digital divide’ exists. My brief experiences show that it is both a chicken-and-egg problem, and one of multiple dependencies.

It is a chicken-and-egg problem because providing service to the poor is a precondition for making money from it. And making money is a precondition for someone creating the service or developing the underlying infrastructure. Service provision to the underserved has not yet proven to be profitable, and private capital has thus been shy of funding technology provision. That is why telecom coverage is only 35% today, and why Microsoft Windows is available in Estonian (with only 4 million speakers), but in only 14 Indian languages.

The problem is further aggravated because success in this segment requires many players to come together. Consider market information systems, few of which have worked partly because they operate under systemic constraints (e.g. commodity pricing) that reduce customer satisfaction. However, another reason for their inability to get critical mass is that they require many providers to be simultaneously available on the system.

The needs of the poor are diverse and include insurance, healthcare advice, credit and weather forecasts. Getting high volumes requires many clients, but that requires getting all these services in on the act simultaneously.

Even infrastructure, a prerequisite to service provision has this problem. Here too, many pieces of a workable pie have been traditionally missing. Successful service requires a delivery platform. That platform requires some network provider or vendor. But rolling out that network and operating end-user devices requires electricity.

Solutions?

Its not a lost battle, however. In fact, some initiatives have shown the way, in particular ITC’s eChoupal. But those projects are notable for going against the logic of short-term financial returns to focus on building critical mass over the long-term.

Elsewhere, a lot of other missing pieces are being put in place. Techies4Dev point to Green-Wifi, an interesting project that solves two infrastructure problems together. Then there’s the Ultra-cheap PC from China, that may not help the farmer, but will certainly expand PC adoption, helping at least urban users expand Internet penetration.

For rural users, there are even wider initiatives going around. The Indian government’s rural electrification scheme while plagued by quality concerns, at least hopes to remove the lack of power. Add to that the TRAI’s Universal Service Fund (USF) and impending use of WiMax, and we may finally eliminate most infrastructure barriers.

All that will remain then is to figure out what people can do with technology, and then communicating with them.

I usually avoid repeating news found elsewhere, but this one is important. India has refused to be part of the suspect “One Laptop Per Child” (OLPC) project of Negroponte and MIT Labs.

News of this came to me from the blogosphere, and I have seen two reasons for India’s rejection.

• The World Bank’s PSD Blog says the Indian Ministry of Education dismissed the laptop as “pedagogically suspect.” Education Secretary Sudeep Banerjee said: “We cannot visualise a situation for decades when we can go beyond the pilot stage. We need classrooms and teachers more urgently than fancy tools.”‘
• The Indian Economy Blog says the Indian HRD Ministry explained that according to some American psychologist “any sustained use of computers may lead to a disembodied brain and bring about isolationist tendencies in social behaviour.”

As Atanu Dey points out, the psychological reasoning is wrong. The pedagogy behind it, however, is certainly suspect. Regardless, the project is egregious at best, transfering scarce funds to a project cooked up by the digital-haves to tell the have-nots what was wrong with them and how to fix it.

Atanu Dey calls it a Formula for Milking the Digital Divide - a must read if you know anything about OLPC.

Beyond the problem of opportunity cost pointed out by Atanu Dey, I believe it is based on the wrong idea - that lack of technology (or the digital divide) is a problem.

As an ex-software engineer I know how engineers work - they are excited by a technical challenge and immerse themselves in it. The desire to innovate can be so strong that questioning the project is unthinkable and the work sustains itself purely on the beliefs of those who do it. What gets lost in the process, unfortunately, is the big picture, and the understanding that technology is only a means to an end - one that can often be achieved in other, simpler ways. I get the same feeling about the OLPC.

Technology exists not for itself, but in order to empower and enable. The reason mobile phones are so popular in the developing world is because they do that well. What is the OLPC supposed to enable that cannot be achieved otherwise, and cheaper?

Update: Another reason, reported by ZDNet UK, is that Negroponte is ‘persona non grata’ in India after the failure of MIT Media Labs Asia, which he had helped setup. Thanks to wayan at OLPC News for pointing this out (comment below).