In January 2005, MIT Media Lab cofounder Nicholas Negroponte announced the One Laptop per Child (OLPC) program, a utopian attempt to improve education in poor communities through the design and global distribution of cheap, low-power laptops. Eventually, Negroponte said, the laptop would sell for a hundred dollars. The program was conceived on a grand scale: Negroponte initially claimed that the laptop would not go into production until governments worldwide had placed a total of five million orders.

But the million-unit orders never materialized. To date, Peru is the program’s largest customer by a large margin, having ordered about 270,000 laptops. So in November 2007, the laptop, dubbed the XO, went into production anyway, at a cost of roughly $188 a unit. At about the same time, OLPC began its holiday-season Give 1 Get 1 drive: any donor who contributed $399 to the project would receive a complimentary XO, and a second XO would be sent to a poor community.

Some observers considered the drive a desperate attempt to inject cash into a floundering endeavor. Then, last week, Intel walked away from a tempestuous six-month partnership with OLPC, scotching the planned unveiling of an Intel version of the XO at this week’s Consumer Electronics Show in Las Vegas. The main point of contention appears to have been Intel’s attempts to sell its own cheap laptop, the Classmate PC, to governments that had already made provisional commitments to OLPC.

OLPC claims that Intel violated a nondisparagement clause in its contract; Intel claims that the clause bound only the company’s officers, not its sales force. The New York Times greeted the news with a headline announcing “The Demise of One Laptop per Child.”

Earlier this week, Technology Review senior editor Larry Hardesty sat down with Walter Bender, OLPC’s president for software and content, to discuss both Intel’s withdrawal and the overall health of the initiative.

Technology Review: What effect does Intel’s departure have on the program?

Walter Bender: Zero. Intel had contributed nothing. They contributed nothing to our current product, the XO. They contributed nothing to our learning models. They contributed nothing to the software. So their going away, so far, is a wash for us.

TR: Isn’t this just the latest blow to the program?

WB: After what?

TR: After large contracts not materializing. Originally, wasn’t there a minimum requirement for a government order?

WB: Originally, there was. We certainly made some mistakes along the way. And one mistake was to be a little bit too rigid in our model. Part of it was just based on some false assumptions on our part in terms of what kind of volume we needed to get things launched. And we thought that going to a few large orders was the best way to jump-start things, to prime the pump.

Some of us, our instinct was quite different. And that was to try to get a broad base and try to make this a grassroots, bottom-up launch instead of a top-down launch. Now, it turns out that we have both. And really, what we’re after is any good idea. So on the one hand, we actually do have some large orders. Maybe not as large as we had originally hoped for, but we’re going to do a quarter of a million laptops just in Peru. And we’re doing something on a similar scale in Uruguay.

Those are examples of top-down. But then there’s a lot of bottom-up. We just did about 100,000 bottom-up machines that we’re going to be distributing through the “give” part of the Give 1 Get 1 program.

TR: What was the purpose of the Give 1 Get 1 program?

WB: Our purpose was twofold: one was to enable us to jump-start laptop programs in places that couldn’t afford to start them themselves. So we’re trying to jump-start Haiti, Rwanda, Mongolia, Afghanistan, Ethiopia, Cambodia.

The second point is that we want to broaden the base of participation. There are a lot of people who want to participate in this program, who want to be part of this global-learning movement. So the number of people who are engaged in our mission has increased dramatically over the last month. We’re finding that the community is really jumping in in ways that are beyond our expectations. So for example, now we’ve got 40 volunteers manning a phone bank, around the world.

TR: It’s customer support?

WB: It is customer support. But it’s customer support from the community instead of from us. Part of the reason we can make the laptop inexpensive is that we’re not building those kinds of things into the cost structure. We’re cutting all those corners. And the way that we can cut them is to design this so that people can have local ownership of the problem. And so, for example, quite literally–you can go to YouTube and see this in action–a nine-year-old can replace the motherboard on the laptop.When the backlight in my Lenovo laptop dies, I have to send it back for factory repair, and they replace the whole display. And if it wasn’t done through warranty–and the warranty costs me more than one of our laptops–I’d probably toss the laptop and buy a new one, because it wouldn’t be worth it. If the backlight dies on our laptop, it is ten screws and a two-dollar part. And not only is it ten screws and a two-dollar part–that a nine-year-old can do the field repair on–but even without the backlight, the laptop still works.

TR: With natural illumination?

WB: Yeah. And that broken display that someone’s going to toss in a landfill somewhere–the one I have from Lenovo has mercury in it. The one that we make doesn’t. So we’ve thought about this stuff. This is not a hack. It’s not an academic exercise. It’s serious stuff, and it’s stuff that we’re doing better than anybody else right now. And we hope that the rest of the world learns from what we’re doing and does better than us. But right now they aren’t. But they will. And that’s part of the plan.

TR: Does that mean you plan to license your technology to other manufacturers?

WB: That’s something we’ve been struggling with. We need an economist to help us figure this one out. It’s not clear to me that we wouldn’t be better serving kids to make everything we’ve done be available to anybody for any purpose. And that might get more laptops to more kids faster.

TR: So from your perspective, this could still be a success even if you stopped manufacturing laptops and the technology found its way into a dozen different laptops …

WB: Yep. But I think it’s premature to do that. And the reason is quite simple: unless we keep the pressure on, the prices are going to go up, the efficiencies are going to go down, and we’re going to be right back to the same “bigger, faster” model. We’ve got to keep the pressure on and keep the industry honest until we’ve really proven that this other way is viable. Because otherwise, next year’s Intel machine will be more expensive and more power hungry, and that’s not going to serve the needs of these kids.

TR: Okay, I have to say, I’ve played with the laptop, and it seems slow.

WB: Well, it’s certainly slow compared to the laptop you carry around. But the metric you have to measure things by is not Grand Theft Auto III. The metric you want to measure things by is learning. The word processor keeps up with my typing. The video camera works just fine. The music programs work just fine. It’s a perfectly adequate platform for kids for learning. Every decision we make is, How does this enhance the learning? And the bottom line is, if you can’t turn it on since you can’t power it, a fast processor doesn’t do you very much good.

TR: There’s also the question of whether laptops are really what governments should be sinking resources into.

WB: The way Nicholas [Negroponte] likes to put it is, substitute the word “education” for “laptop.” And then ask, “Should we be giving these kids education?” “Nah, they don’t need education! Education is a luxury. Why should we give them education?” What we’re advocating is that the laptop is the most efficient way we know of of giving them an opportunity for real learning. It’s not that we’re interested in laptops; we’re interested in learning. And it turns out that almost 50 years of research by people like [computer scientist and educational theorist] Seymour Papert has demonstrated that computation is a wonderful thing to think with. It’s powerful stuff. And it’s going to change these kids’ lives dramatically for the better.

————————————————————————————-

For more info on the OLPC, check out these links:

1. http://en.wikipedia.org/wiki/$100_laptop

2. http://wiki.laptop.org/go/Home

————————————————————————————-

And for you techies out there here is a rip up of the XO-1 from bunnie’s blog

I got an OLPC XO-1 a few days ago in the mail as part of the give one, get one program. Hopefully some child out there is enjoying their new laptop–there’s a certain amount of opacity in the process so I have no idea even if this laptop went to some needy far-flung village in a developing nation, as most of the propaganda would have you believe. I don’t mean to be too critical; I recognize that proper allocation of charitable resources is one of the most difficult tasks anyone can attempt, but a donation the magnitude of the laptop, even if it is only $200, is big money in the most needy countries. I somehow hesitate to think if I shouldn’t have just given a check to UNICEF instead to provide medicine to newborns. Then again, as long as one laptop is being delivered to somewhere that’s needy, I’m not doing bad compared to some other charities, where often more than half the donation is retained to cover management fees and fundraising costs. I also got a laptop out of this…so for whatever reason, Negroponte’s logic was successful in convincing me.

Well, the point of this blog isn’t to discuss the merits of charities. We like hardware, and the OLPC XO-1 is an interesting piece of hardware. There are plenty of teardowns for the OLPC XO-1 (including one on the OLPC wiki itself), so I won’t repeat the tedium of what screw comes out of where and just cut to what I thought were interesting highlights.

If I were to make one general comment about the OLPC XO-1, it’s that its mechanical design is brilliant. It’s a fairly clean-sheet redesign of traditional notebook PC mechanics around the goal of survivability, serviceability, and robustness (then again, I’ve never taken apart any of the ruggedized notebooks out there). When closed up for “travel”, all the ports are covered, and the cooling system is extremely simple so it should survive in dusty and dirty environments. Significantly, the port coverings aren’t done with rubberized end caps that you can lose or forget to put on–they are done using the wifi antennae, and the basic design causes the user to swivel them back to cover the ports when they are packing up the laptop to go. That’s thoughtful design.

The OLPC is one of the first devices to use the Li-Fe-P battery chemistry. My understanding is that it’s safer than normal Lithium Ions (and those who have been reading the chumby fora are aware of my liability and safety concerns around putting a Lithium Ion battery into any device), but it has a lower capacity.

Unlike most other laptops, the OLPC is easy to take apart and service–something important for a device you are sending to a remote region of the world. The LCD, which is shock mounted, was clearly designed to be easy to replace if it was broken–a few screws to remove, some panels to slide off, and you’re done. This is unlike my T60p, where even the trained tech who did the warranty repair on it couldn’t quite put humpty dumpty back together again.

The display is made by Chilin (which is related to Chi Mei, one of the major LCD glass manufacturers–the current chumby’s glass is made by them as well). I had the fortune of meeting Scott Song from Chilin at China FOO, and he was telling me some of the interesting yet lesser-known features about the LCD. In addition to the celebrated dual-mode monochrome/color transreflective layer stack used by the panel, it also features a removeable backlight LED bar. Aside from mechanical abuse, the most common failure mode for an LCD is the backlight burning out. Most LED backlights are rated for only about 10,000 hours–if that–and it’s expensive to trash a whole LCD for the failure of a fifty cent component. Therefore, the backlight bar was designed to be easily replaced.

I gave it a whirl on the OLPC, and it certainly lived up to its expectations. Usually, to get to the backlight bar, you have to risk destroying the LCD panel, but in this case, two screws was all it took.

I decided to fire up the backlight outside the display, and let me tell you, that sucka is bright.

Note that the photo above wasn’t taken in the dark–it’s just that the camera shutter had to tighten up so much so that the background looked dark. If I were in the third world, I probably would say screw the laptop, I’m taking the backlight bar out and lighting my home with it!

Here again is a close-up of the pixel array used by the transreflective display (click on the image for a much larger version).

One very slick thing about the OLPC software is that I can hit a button on the lower left hand corner and it will rotate the display rendering by 90 degrees each press of the button. This allows me to easily fold the display back and use it as a sunlight-readable eBook device.

Thanks in part to the low power of the Geode CPU, the design mounts the motherboard against the LCD, unlike most laptops where the motherboard is under the keyboard. This configuration has some simplicity advantages, especially considering the flexibility required by the display unit that can be flipped 180 degrees in either direction. The heatsink for the CPU consists simply of a thin metal heat spreader, which is in close proximity to the plastic casing: no cooling holes, fins, or fans to collect dust and break (notice how the Intel Classmate PC features prominent cooling holes for its under-keyboard CPU module). It also doesn’t burn your lap up while you are using it (although the display still does get quite warm when you use it–I tucked it into my jacket once to keep me warm while running around outside in the cold).

Here’s a photo of the motherboard with the heat spreader on:

And here’s a hi-res photo of it with the heat spreader off (click on the image to access the hi-res version):

Notice how both of the large BGA chips are underfilled to provide better shock and vibration robustness. I actually have never seen an underfill like this before–it seems to be oozing out of the edges–and it also doesn’t seem to be very uniform (some spots seem to have a little underfill missing). Most underfills I’m familiar with to attempt to cover every gap and void underneath a chip (which is actually a very hard process problem); maybe this is some new kind of underfilling technique that expands a little bit upon cure to help cover voids and its robust to a few missing spots. If a reader is familiar with this type of underfill technique, I’d appreciate a link to it.

Here’s the backside of the motherboard, nice and simple (again, click for the hi-res version):

And here are some detail shots, first of the camera:

And of the wifi:

A little trivia here–this is the same chipset used inside the Microsoft Xbox360 Wireless Networking Adapter.

Presumably Microsoft doesn’t use the programmable ARM core in the Marvell 88W8388 to do something as magnanimous as bringing mesh wifi networking to the third world–probably it’s used to help implement security authentication to support their high ASPs for their closed accessories and to implement Microsoft-enhanced protocol tweaks that helps lock out competitors. Engineers make guns, they don’t tell you where to point them–that’s up to the marketing guys. Hackers take these guns apart and use the barrels and stocks as tubing and lumber to build functional art. Sometimes the art includes a bigger gun or a barrel that points backwards, and often it’s up to the media to interpret what this cryptic art means. Really, most of the time, it was done just because it’s fun to make art, but that doesn’t sell copy.

The OLPC’s wifi sensitivity is really quite excellent–the dual antennae clear of the laptop’s body do wonders for sensitivity. On the other hand, I had some trouble getting the native UI to associate the OLPC to my access point, and to get it to stay there. Overall, the software on the OLPC is clever but very “appliance-like”: there are some pre-loaded applications and it’s not immediately obvious how to add new applications using the native UI (it’s hackable from the command line but that’s not very beginner-friendly). Then again, it does include some education-oriented scripting languages that kids can use to write programs, even if it does lack a local gcc installation, and it includes the basic infrastructure for chat, video, audio, and photo sharing functionalities. If anything, I think the OLPC could play an important part in helping people keep in touch better in remote parts of the world, with some software improvements in that area (I don’t have two of these so I wasn’t able to test how easy it is to share files or connect to each other over other protocols). It also didn’t have an easy way to reconfigure the keyboard to the dvorak format that I type (I probably have to go in and hack the X configuration files via the commandline) and the keyboard is not at all comfortable for adult hands to type on–maybe I will have to plug a full-sized keyboard into the USB port to do anything extensive with the device, which sort of voids the portability aspect of the laptop. Then again, the laptop was designed for children with smaller hands, and not adults like me. Wait, I’m an adult? Crap!

Finally, some screenshots of the OLPC XO-1’s OS configuration:

Interestingly, the OLPC ships with a hardware AES unit. Great for privacy, but presumably this is going to cause some troubles for the US government export controls when it comes to shipping the OLPC to certain third world countries on certain blacklists…I know I had to fill out parts of a somewhat thorny encryption questionnaire for the chumby lawyers as part of preparation for chumby’s sale.