Nicholas Negroponte is about to meet the goal he set in 2005 in Tunis of an under-$100 laptop computer designed to give the world’s children access to the information society.
His One Laptop per Child Foundation revealed prototypes of its third generation design, the XO-3, in January. It hopes to find manufacturers who will allow the design to be available in quantity by the end of the year at around the $100 bar. (There’s a delay; the symbolic price point is perhaps more sacred than the deadline.)
OLPC xo-3 prototype
More photos here.
Mission accomplished! Or is it?
The specifications of the reference design are here. Unofficial summary: it’s a semi-ruggedized tablet with an 8″ touchscreen, a Marvell smartphone SoC including a low-power ARM-type CPU, 2 USB sockets, 4 to 16 GB of solid-state memory only, and wifi. It accepts a wide range of voltages for recharging from solar panels, hand cranks, or car batteries. The software is built on a Linux kernel and Android or Fedora as operating system, and runs a fairly standard suite of browser, word processor, pdf reader, drawing program, calculator, and media player. A spreadsheet is in the works, but as a second best OLPC laptops already run the Google Docs cloud-based one.
It would be churlish to deny Negroponte his laurels on the grounds that the XO-3 is a tablet, not in current usage a laptop. You can always plug in other peripherals, including a standard USB keyboard, which you can get for under $5 a piece in bulk from China. With greater difficulty, you can probably connect an external screen or projector; though these aren’t likely to be affordable yet in poor countries.
It would be a mistake to think of this as just a cheaper commercial tablet or netbook. Netbooks are designed to run just like conventional laptops, which in turn emulate desktops. The iPad and competitors are primarily media players and readers, with data entry and work creation a secondary consideration. The OLPC project is a specifically educational one – and of a utopian flavour.
On the Rousseau scale between Mr Gradgrind’s academy at 0, with children seated in rows chanting the multiplication table, and a Maurice Sendak garden of undirected self-discovery at 10, the OLPC philosophy comes in at around 7:
..With mobile, connected laptops the walls of the classroom open and the entire community becomes the classroom and virtually the whole world enters on demand. The children carry the classrooms and teachers of the world with them through the community and into their homes. Children can participate in the study of global issues while simultaneously using local context for understanding. They can fully participate as producers of knowledge and not just as consumers of materials produced by others.
In pursuit of this philosophy, the OLPC has written a bespoke child-friendly Linux-based GUI called Sugar. (You can download a bootable version to play with.) It was designed from the outset for easy collaboration; Sugar supports wireless mesh networking, and available “neighbours” are visible on the home screen. If there’s one Internet access point, any one machine can act as the router. Of course, any user needing to use the traditional software resources of browser, spreadsheet and so on will have to adapt to the conventions of the industry-standard GUIs, but I suppose that’s manageable.
About 2.5 million OLPC laptops have been sold and given away. Two countries have mass deployment: Uruguay with 380,000 – the entire primary school population -, and Peru with 870,000. Around 15 other countries – including the USA – have pilot deployments in the thousands.
What to think of this? Clearly Negroponte deserves congratulations for bringing off his self-imposed challenge. Market forces would no doubt eventually have produced $100 netbooks and tablets. The big drivers in cost reduction, especially the ARM ecosystem in mobile phones and cheaper solid-state memories, have been available to any vendor. But Negroponte has brought this forward, and in a way that creates a lot more educational opportunities. It’s not a valid objection that $100 is still too much for the very poor, in Indian villages and sub-Saharan Africa; the OLPC (and its emulators like the Indian Aakash) will surely continue to work to bring costs down, riding and pushing the predictable advances in the technology.
It is, of course, a technologist’s vision. Give people the tools, and they will find ways of using them. The project does remind me of the starry-eyed top-down educational computing projects of the1970s: as if putting An Actual Computer in every classroom would magically generate a generation of engineers, and language laboratories obviate the need for teachers who actually speak foreign languages.
Cheap computing is changing education in rich countries, though you’d hardly know it from curricula. Spreadsheets should for example have revolutionised learning statistics and the representation of quantitative information in charts and tables, and conversely spell checkers have made orthography a trivial skill. It’s also become much, much quicker to look stuff up, from the dates of the Kings of England to the Latin names of birds to the etymology of foreign words, so learning information by heart is less important. The general rise in raw examination scores may be in part down to mass computing access.
I think I have two minor beefs with the OLPC. One is the over-emphasis on collaborative projects. The aim is sound; real life is almost all teamwork not solitary effort. But the reason team projects don’t happen much in schools is the conservatism and lack of ambition of teachers, reinforced by the Gradgrinds of officialdom with their test scores and business-speak of accountability. It’s very little to do with computers. My mother was lucky enough as a child to be sent to Miss Gilpin’s Hall School in Weybridge, Surrey, in the 1920s. The pupils carried out what would now be called multimedia projects, on say the Chanson de Rolland – a play, music, writing, art, brought together at the end in a professionally printed book. Now even today most teachers in rich countries aren’t up to this. What chance is there in Peru or Rwanda? Occasionally the Rwandan equivalent of Eva Gilpin will seize the opportunity, which is worth having, but don’t expect too much.
The second is that the revolutionary social impact of the computer is not as stand-alone golem but as information portal. The full potential of the project requires a solution to the Internet access problem. OLPC stresses the lesser but more soluble aspects of local networking, and offline resources accessed through school servers or USB memory sticks. I’m not convinced they are right. The mesh technology greatly lowers the bar already. One landline or satellite dish, plus a wifi router, per school is a perfectly achievable objective in middle-income countries like Brazil.
That’s where the project becomes truly disruptive. One objection to it is that the money for 100 laptops would be better spent say on a school library. Possibly. But what are 2,000 books compared to the Library of Alexandria that is today’s World-Wide Web? Again, most pupils and teachers will only use a tiny fraction of the resources that could be useful to them. Some will sadly find obscurantism, cruelty and obscenity. But for a tiny few, the Shakespeares and Darwins and Einsteins of this century soon discovering the world under banyans and baobabs, it will prove a revelation.
A school under a tree in Sudan
Photo credit
As long as spreadsheets focus on bullsh*t graphics like tilted pie charts and ignore the work of Bill Cleveland, Lee Wilkinson and others they won’t revolutionize the visual display quantitative representation. Look at the default graphics in the latest version of Excel in the light of Wilkinson’s work on visual grammar, please. Excel doesn’t produce dot charts or conditioning plots. Excel hasn’t reached the level Edgar Anderson was at in the 1960s, much less Tukey’s work on Exploratory Data Analysis in the 1970s. Just to be clear, Open Office is no better.
Until recently, Excel’s numerics were unreliable too. Microsoft had a penchant for choosing to implement algorithms that were less than state-of-the-art. As recently as five years ago, I told students that Excel was a useful tool for data entry but unreliable for data analysis. If you have to run the data through a serious statistics package (e.g., Epistat, R, S, SAS, SPSS, Stata) to check on Excel’s calculations why bother doing the computations in Excel at all?
Microsoft seems to have fixed the numerical issues in Excel. Open Office has always had thousands of geeks fiddling with it and it has always had pretty good numerics. But the choice of analytic methods is still limited to things R.A. Fisher would recognize intimately. No robust methods, no non-parametric methods, no bootstrapping and no Bayesian inference. I’m not certain that any of these ought to go into Excel’s tools, but until they do spreadsheets won’t revolutionize statistical presentations, because Excel’s tools are the same old stuff.
Tilted pie charts is a new one on me, but you’re right, Excel 2007 does offer the gimmicky thing. At least the distorsion is upfront and the mind sort of compensates. Far worse are the cone and pyramid charts it also offers; the data are proportional to the heights, but the eye reads as area or volume. Truth fail. But to get things really wrong you need to quit your spreadsheet for a graphics program, which has no statistical principles at all.
I was in truth thinking of secondary school students rather than higher education. You can use standard spreadsheets to teach, learn and practice basic chart types and statistical methods, and seeing through their dishonest abuse, though I’m sure these applications could do with improvement for advanced work.
I have used that chart as classroom example of how not to do it. Thanks for bringing it to my attention.