For those of you who were present as well as for those of us who would like to have a reference and remember what happened in class today, below are my notes from today's follow-on discussion on the "One Laptop Per Child" project. I thought that the discussion was quite balanced -- with people pointing out both the strong and weak aspects of the various issues. We didn't vote in the end :) -- "go/no go" -- but my sense was that a vote would not have been unanimous. Also, in class I didn't get a chance to mention some of the arguments "for" that I had prepared, but they are conveniently included below. A few references are also included at the bottom. Enjoy! Valentin ********************************* One Laptop Per Child (discussion) November 15, 2006 ********************************* Power issues: - There are apparently companies that use hand cranks for generating energy (FreePlay) -- in portable low-tech devices of various sorts, e.g., flashlights, radios, phones. - OLPC plans for a 7.5" screen, which is smaller than the usual laptop screen (that takes much less energy than the 60W of conventional machiens). Additionally, when backlight displays are replaced with non-backlight displays, yielding extremely low energy consumption. Another invention in low power that is relevant is using NiMH batteries that can be charged quicker and with less power dissipation; in contrast, Lithium batteries have a rapid charging curve at first but then to fully charge only small portions of the actual energy expended is converted to a charge (which wastes human power). - There would be a range of deployments -- some of which are power-enabled and some of which are "power free" (i.e., no electricity). In some places, e.g., rural Cambodia, although there are no power lines, people regularly watch TVs on car batteries and there is a whole infrastructure around charging those batteries (i.e., someone driving around and doing service while charging people's batteries). Environmental impacts: - Relatively little information is available yet. - OLPC has committed to conform to ROHS (Regulation on Hazardous Substances), a European standard. - Many recycling plants for electronics are in the Third World. The process of recycling produces toxic chemicals. Still, it's better than burning or simply throwing away in the garbage. - In villages (e.g., in China) close to recycling centers people have been found to have elevated levels of lead up to ~200x what is considered a safe dose. - The logistics of recycling are not trivial, especially in a Third World scenario. - There is a movement toward using bioplastics for electronics. This is still in the experimental stage. - The design is intended to last 5 years at least before it breaks. The expectation is that it won't be obsolete, but that the machine will break first. (The definition of "obsolete" is different in the developing world, where people use and fix things until they possibly can and only then dispose of them, especially when those are expensive things.) Other issues (content, deployment, etc.): - May increase the Digital Divide (esp. in countries like Brazil that would otherwise need several billion dollars to pay for all kids). - Curricula don't evolve as quickly as we would like them to be, especially in some developing countries without formal processes. - Localization of software is costly. Often, there are computer labs in which machines idle because there is no content that is appropriate for them. It may be possible to have expatriates help with localization and content creation, but this is not part of the OLPC project. In Cambodia, there are resources, but Unicode issues preclude them from searching in their native language. - In many developing countries, kids usually work at home, so there's not really much time that they have available for tinkering with computers. On the other hand, computers change the way people do things, so they may enable new ways in those kids' lives. - In Cambodia, Internet is "weather-dependent" -- half of the year during the rainy season electricity would go out, then it would take a few hours to bring up the Internet (by the ISP), so half of the day is routinely lost and no one is using the computer libraries. - The deployment needs a large initial subscription, because that's what will get manufacturers interested in investing into it (even if all the design work has already been completed). By nature of how they do business, they need to see a positive effect on their bottom line, albeit possibly not all of it immediately. This would simultaneously serve to drive down the cost per unit, due to economies of scale. Low-cost deployments (and pilots) would cost dramatically more because of the (lacking, in this case) economy of scale. - The biggest issue with OLPC seems to be the large deployment without a prior pilot. It may be cost-effective if a company would agree to manufacture devices even at the price tag of $5000 per unit, but much fewer in numbers (e.g., 10000), with which to experiment in classrooms and to discover what works and what doesn't (and whether it's worth scaling up the effort from the point of view of local governments). Strong points of OLPC: - The one-sentence pitch of the OLPC project (i.e., the problem they're trying to solve) is, roughly: "At the cost of a few textbooks, one can buy the textbooks (in electronic form) as well as a laptop, which enables constrictivist learning." - The project is a bold effort that puts a stake in the ground and pursuing it -- engages the technological industries and challenges them to come up with fundamentally new cost-effective components and designs. Novel are not only the technological aspects, but the "business" model as well. It's okay if not all of the original visionary objectives are reached, but we owe it to ourselves to explore the possibilities. - The most popular uses of new technology emerge usually after some time and usage, rather than immediately, so counting on predicting what that usage model would be is bound to fail. Since we can't know ahead of time exactly how students will use the devices -- and this usage would differ from country to country -- the idea is to build a robust, general-purpose device aimed at low-power, durability, etc. The rest will evolve. - Governments need to be involved with paying at least some portion of the cost, so that they would have a vested interest in the project. Otherwise, corruption would be rampant, as is the case with almost any development aid for Third World countries. (Carnegie did this too with involving local governments who would agree to pay 10% of his donation annually for sustaining local libraries.) - Moving quickly has a virtue, since waiting (until we agree on something better) would mean that the current generation of children in the developing world would not get access to resources that we could realistically provide them now. A delay now would endanger the prospects of entire countries that would miss an opportunity to get ahead through technology in the next generation. To quote John Dewey, "If you think education is expensive, try ignorance." - The LCD improvements are patent-free, so they will percolate to other configurations, affecting machines broadly. - The technological aspects behind OLPC are strong. The rest gets progressively "cloudy." :) - The out-of-the-box experience for laptops in OLPC is connectivity via mesh networks to each other. Then, only one of the laptops will need to be connected to the Internet (in the spirit of the 802.11s protocol for mesh networking). Interesting, relevant references: - "Philanthropy's New Prototype", by James Surowiecki (Technology Review) Part I: http://www.technologyreview.com/printer_friendly_article.aspx?id=17722 Part II: http://www.technologyreview.com/printer_friendly_article.aspx?id=17777 Part III: http://www.technologyreview.com/printer_friendly_article.aspx?id=17778 - "Design Reality Gap", by Richard Heeks *************************