June 3, 2008
[berkman] Berkman lunch: Walter Bender, Sugar Labs
Walter Bender, who was president of the One Laptop Per Child project, is now the founder of Sugar Labs. [Live blogging. Getting it wrong. Missing stuff. That’s just the way it is. Also, this will be much choppier than the talk actually is.]
The aim of OLPC, says Walter, was to transform education around the world. Laptops aren’t the solution the world’s problems, but learning is. And laptops can help with that. “It’s all in service to learning,” he says. He refers to the book Predictably Irrational. Chapter 2 is on “anchoring.” Walter says that he’s anchored to the idea that a “connected computer is the most powerful tool we have” for knowledge creation and sharing. It’s not a panacea, and you can’t just throw the laptops over the wall and wait for the magic to happen, but laptops can be a catalyst.
School reform is impossible, he says, if it’s top down. But it will be accomplished by students bottom up.
Walter’s talk to us is titled “Confessions of a Fundamentalist.” His Open Source fundamentalism was taken as “distorting OLPC in a way that distorted its mission.” He is a fundamentalist about what are the best ways of enabling learning, of planting seeds for learning. He’s willing to bend his principles about Open Source but not about learning.
He subscribes to constructionism, a theory of learning developed by Seymour Papert. Papert was a student of Paiget. “You learn through doing. If you want more learning, you want more doing. And what’s a better tool for doing than a computer?” The corollary is that “love is a better master than duty, so you want to engage people in things they’re passionate about.” Computers, as Turing machines, can do anything. Computers are a “damn good” vehicle for working on what matters to you. He gives an eample of a Thai village where the children used computers to figure out where the reservoir should be built.
Everyone is a learner and everyone is a teacher. “It’s just inherent in our being.” And we’re expressive and social. The teacher-student dichotomy is false. We should instead by learning centric, Walter says.
“Proprietary tools are often associated with the delivery of knowledge,” he says. The criticism is that we’re trying to turn every kid into a Linux kernel hacker. “Yeah. we are,” he says, to chuckles. But they don’t expect every kid to become one. They really want kids to appropriate rather than merely access knowledge. “Open Source has a culture around appropriation that’s important to the culture of learning.” (Of course, he says, you can do constructionist learning with proprietary, or service-oriented with Open Source.)
When OLPC designs its gen 2 that’s more like a book, they’ll be making the mistake of forgetting the dyna in the dynabook (Alan Kaye’s idea that a ebook is more than an analogous), he says, in response to a question. We want books that make it easy to insert comments, for example. “You want to build in affordances that encourage the type of behavior you want to see.”
In a digression, he says that when he headed the MIT Media Lab, he had a three part process: Build, critique, iterate. “Use your time at university to make really big mistakes.
“Learning wants to be free.” There’s a difference between governance and the engagement of the ommunity The Open Source community has developed a “number of very powerful tools around engaging in collaboration and engaging in critique. Those tools are for the most part lacking in the world of education. Certainly in primary education.” They started a pilot in Nigeria where there are 300-500 languages. They were in a school where the primary language was Igbo. The OLPC’s dictionary was only in English. So the kids wrote their own. The kids discovered they had the power. “To me, that’s a real game-changer.”
Walter now talks about Sugar, the user experience that has come with the OLPC by default. It’s available on Ubuntu. Sugar is based on the first principles: be a learner, be a teacher, be expressive, be a learner. Three things abbout Sugar:
1. It wraps applications in “activities,” adding sociality: everyone is right there with you. E.g., when you’re writing a doc, anyone is one click away from “putting their cursor in your document.”
2. The Journal makes sure that everything is preserved, but the importance is that it creates a diary, a portfolio of your work. You can there have a conversation with a teacher or parent about your progress. That march through time “is an important feature of learning.”
3. The framework is simplified and transparent. The transparency means there’s no upper limit. E.g., TamTam starts out as a “busy box”: choose a sound and slap the keyboard. But you can progress to TamTamJam, which is more network-centric and lets you layer instruments. From there you can go to TamTamEdit, where they can compose music. Then, in the SynthesizerLab you can create your own instrument. Then you can edit the Python code underlying the instruments, or hack C-sound (“midi on steroids”). “No upper bound on complexity.”
Sugar is now reaching out to be a general-purpose environment in the Linux “and perhaps even in the Windows” world. “I don’t know how to do it in the Windows world,” he says, but …
“Sugar is pretty raw. It’s alpha. It’s flaky. And it’s in the hands about about 600,000 kids…which is pretty good!” The kids are giving feedback and making improvements.
“Now I come to David Hilbert.” In 1900, he posed 23 problems to mathematicians. Walter has 23 problems facing people interested in technology and learning. He’s going to blog them. They include: How can we make the damn network work? Create malleable code that doesn’t turn into malware? How to get localization/internationalization tools that are two orders of magnitude better? How do we a better job of using more wisely a very scarce resource: power? Does constructivism scale? We need better tools to introduce change. How to transplant the culture of freedom and critique from computer science into education? Economic challenges. Research correlating learning and economic development?
Q: Learning should be free? Which senses?
A: Not free as in beer. But you learn to program by copying code.
Q: OLPC has inspired a bunch of commercial tiny laptops. Will this help?
A: Five were announced last week. It’ll help.
[me] Is the constructionist theory cross-cultural?
A: Constructionism is built on first principles that are not culturally dependent. It’s no more culturally tied than Piaget. What children love — what matters to them — is culturally dependent. And what’s the role of the teacher? The teacher is unleashed. They have a lot more fun.
[clippinger] Constructionism has implications for authority, which have dramatic cultural implications.
A: The finance minister is always interested because they see that that’s how they’re going to get entrepreneurs.
[roger] Have the proprietary software companies gotten there first? How does that play out?
A: That will be one of the big social-economic battles over the next 20-30 years. The ones who go with Open Source will do better.
[harry] That’s part of your fundamentalism. For me, the question is how many types of cheap laptops there will be in five years. Will there only be a couple?
A: OLPC tries to keep the pressure on. The market will be big enough. But I worry. If these things are used to replace chalkboards, it’ll be a drag on the process.
What about when you look at college students? Some of these principles are not being taught.
A: Part of one’s education should be getting dirt under your fingernails and building stuff.
Q: [ben] We changed so much after Nigeria. Do you think the trial was successful?
A: What we have today is much better, but those kids were learning and constructing.
Q: Much of what was broken were the social affordances. Maybe the lesson is that we can achieve success without the fancy features…
A: There’s a bit of a placebo effect, sort of. You say “This is yours. It’s about sharing, expression, communication.” That in itself was enough to make the change. Those early systems couldn’t support the growth, but they had enough to introduce the change in culture.
Q: The idea of glossiness. If you present an interface that looks a bit broken, people are much more willing to get their hands dirty and play with it, and think they have something to contribute.
A: You don’t want things to break, but you do want people to explore. Rather than trying to make everything hard to break, we’ll make it easier to repair. As long as it’s easy to get back to where you were, people will try things out.
