The XO Files: I Want to Believe


Reading WorldChanging's "editorial retrospective" by Ethan Zuckerman on the article on the OLPC XO he posted in June, 2006 really made me wish that the project had, well, worked out better. Articles like Ethan's remind me of the good work and ideals that have gone into the OLPC XO, which both refreshes me and frustrates me further. This post begins a four-part series on the One Laptop Per Child project, some of the key problems it has faced, and the amazing promise that it still holds for international development and global education.

The XO Files: I Want To Believe
The XO Files: I Want To Believe

I Want To Believe: Part I: Laptop project or education project?

The XO, despite anything else, remains an amazing chunk of technology and an exemplary model of innovation for change. Ethan Zuckerman focused on the amazing feats of engineering possible with a bit of foresight, some tight cost and design constraints, and some really smart people:

It's got bunny ears - antenni for the 802.11s wireless radios, which are designed to self-assemble meshes with other laptops. The ears fold down to cover the USB, power and mic ports, an excellent design for the sorts of dusty environments I can imagine the device used in. [...] Ever since Nicholas outlined the engineering challenges of building a good hinge, I've been fascinated by the different ways people attach screens to laptops. [...] The machine still needs to be miserly with power to be usable as a human-charged device. And this is where the team have worked some serious magic. [...] The machine draws [0.5 watts ] in ebook mode, using a black and white display. The display [...] can store a black and white image and display it without any assistance from the CPU, again allowing the CPU to shut down and save power.

Technology paired with idealism does not a development project make; there is a huge gap between creating and pushing out a technology and creating a vibrant, community-led, sustainable change on the ground. This canyon between Boston and the Global South is the leitmotif of the many obstacles the OLPC project has faced. With his nigh-trademark prescience, Ethan Zuckerman pointed out an aspect of this disconnect that has proven to be a serious problem in the current "marketing" of the machine -- and why Windows XP is now the default OS, driving away many of the early XO software developers:

Getting across the distinction that this is a children's laptop, not just a cheap laptop, is a surprisingly difficult task. When I last wrote about the laptop on Worldchanging, a number of commenters mentioned that they'd like one of the computers as a backup or travel computer - I suspect they might feel differently after playing with one of the current prototypes.

The difficulty of course is that the XO is a great travel laptop; it's lightweight to make your shoulders happy, rugged enough to toss in your daypack for a hike, the sun-readable screen means you can use it outdoors, the SD card slot makes it a great way to review digital camera photos, the antennae can pick up even a feint wifi signal, and the USB slots and headphone jack make a handy movie-viewing box. I've even seen people use it as a VOIP/SIP phone to make low-cost international calls. The low-power possibilities (especially if you deactivate wifi) means it will last for hours on one charge, and can be topped up using a portable solar cell.

Without diving further into the amazing technology (which we'll discuss in Part II), you can see how hard it is to "sell" the XO laptop without gushing about the amazing laptop features it sports "under the hood". The failure to drive home that the XO is a children's educational device, not a laptop, is why people ask questions like "Does it run MS Office?"

"When I went to Egypt for the first time, I met separately with the minister of communications, minister of education, minister of science and technology, and the prime minister, and each one of them, within the first three sentences, said, 'Can you run Windows?'" Negroponte says. [Technology Review, May 2008, emphasis added]

Would you ask if a Speak-n-Spell ran Windows? No, but the Speak-n-Spell was still a reasonably powerful "computer" in many ways, as Wikipedia reminds us: "The Speak & Spell used the first single-chip voice synthesizer, the TI TMS5100, which utilized a 10th-order linear predictive coding (LPC) model and the electronic DSP logic," and it even was able to interface with early desktop computers.

A TI Speak-n-Spell
Speak-n-Spell: The 1980s XO?

The Speak-n-Spell was branded, promoted and purposed solely for early childhood education - not business (or communication with alien races, but that's why it's a good platform). The OLPC's XO Laptop is branded and purposed primarily for education, but the "$100 Laptop" moniker, and every branding step that focused people's attention on the fact that this thing was a laptop, created unrealistic and damaging expectations to the project goals. If you call it a laptop, even a children's laptop, presumptions on what it does and how it works are made. The XO is a collection of path-breaking technology that shouldn't be constrained by the concepts of a "laptop."

But that's how the OLPC Foundation treated it, talked about it, and it's how the world saw it. Despite the repeated (but now defunct) mantra by the Foundation that it was an education project, not a laptop project, their actions focused on the engineering and distribution of the technology, and not on the custom, per-country curricula integration and teacher training that an educational project would need. It should have been an education project, but in design, execution, and promotion, was a technology project, and now it must compete, not as a selfless work to support global education, but as a low-cost laptop in an increasingly crowded market that it itself helped to create, which will be the topic of the next entry.

Related Entries


Repeat after me:
Windows is not the default OS
Windows is not the default OS

The public perception has been negative when NN announced that OLPC would be able to run windows. But a large portion of OLPC's employees are programmers, linux programmers, and GPL programmers.

As recently as two days ago there is a quote here from Kim Quirk talking about how the XO won't be dual boot in the US, but will run Sugar:

I realize that this false impression is probably due to OLPC not shouting Sugar Sugar Sugar as loudly as possible. But AFAIK, Sugar on GNU/linux isn't going away.

You say the XO shouldn't be constrained to the limits of a "laptop", but at the same time seem to call out the "Speak and Spell" as an example they should have emulated.
The S&S becomes a doorstop for the owner after the owner has learned to spell.
On the other hand, the XO can grow/change with the needs and abilities of the user.
It is a computer and unlike almost any other product man has ever produced, it has this ability. (Note: Yes, all computers have this ability, that is my point.) That flexibility is the most important aspect of the XO and without it, any discussion about its use in the developing world would be lost due to issues of cost effectiveness. On the other hand, I agree that using the word "laptop" to describe it was probably a mistake. Perhaps they should have gone with "dynabook" or coined a completely new name...

The perception of any computer is governed more by the available software for it than by anything else about the machine. At G1G1v1 time, the list of available activities was clearly education-oriented and clearly aimed at elementary school level. This is no longer the case, in my opinion. Which may be why the perception of the XO as a "children's machine" is getting blurred. (That and the fact that an experienced linux user *can* make it do a heck of a lot more that first appearances would lead you to believe.)

Looking over the current list of activities that are "ready for prime-time" (i.e.: at least "beta"), I see a lot of recreational activities that are only tangentially justifiable as educational, a lot of programming languages or environments, a bunch of what seem to be adult-level utilities but not a lot of grammar school level education-specific ones. (But maybe that's because those are being wrapped in big bundles like ePaath, GCompris and MaMaMedia?)

The software suite for the XO (and here I'm talking about activities, not Sugar) is developing more slowly than we'd hoped, I think, but probably not slower than we could realistically expect, given the dependence on the community of volunteers to provide these activities, rather than having OLPC have a staff sufficient to tackle the job. The initially poor (but now improving) developer-level documentation didn't help either.

While many of us are programmers and some of us have expertise in subjects that would make for good elementary school education topics, few of us are also educators. Which means that the community of volunteer developers is not likely to write good educational software -- or even know what that means! -- without some direction and hand-holding. (It's probably also true that very few of us are also capable of translating our own program's UI into any other language, which is another place where depending on volunteer help is failing.)

Just my opinion.


"The S&S becomes a doorstop for the owner after the owner has learned to spell."

Or even before that. Since for kids in most of the targeted countries English is not the primary language and most languages use phonemic orthography (i.e.'spelling' is phonetically based ), then teaching their own language (surely a priority before anything else) doesn't involve teaching 'spelling' at all...

For those who do want to learn English, the S&S functionality could be (since XO is not just another S&S-like 'gadget' as Intel and MS was trying to portray it :) implemented, thanks to its computer-like functionality, as just another XO activity....

The main difficulty is that the XO is not AVAILABLE. Only a couple of hundred thousand were sold to G1G1 users and to a couple of educational projects in developing countries. The possible user & developer base is still VERY small.

If anyone could buy an XO for USD250, extreme amounts would be sold, and the user & developer base would grow huge quickly. Then, software development would be a smaller problem, the open source community and commercial vendors would quickly fill-in the gap left open by OLPC's development incompetence.

I don't think this is only an education project:

Today OLPC is the ONLY one laptop with outdoors screen mode.
Today OLPC is one of the best feature/price ratio of all laptops.
Today OLPC has one of the cheaper batteries of all.
Perfect ebook.

So they want to give kids what adults don't have. Don't let me wrong, I'm with kids having access to tech as much as anyone here, but give adults the same possibilities (they are a lot of poor 3erd world adults that can use it for making their lives better too)

@Seth - I seriously doubt that MS will allow the sale of their customized XP in the US, especially for the "Unlimited Potential" price of $3. The dual boot option being discussed for G1G1 is between Sugar and a desktop Linux

@Bill, et al

Obviously, the XO is a laptop, for all intents and purposes, under the hood. But for the marketing of the device, it should have been presented as an educational tool and NOT as a cheap laptop. The minute you say laptop you start down a different path in who should be using to do what kind of things.

@sola - That's exactly the point I make in different ways in the next two parts.

As a couple of people have commented, the key difference between a Speak and Spell and the XO is maluability to a task. Unfortunately, it is not very maluable from the perspective of those who should have the ability to develop resources for the XO: teachers and children.

Now I know that people claim that Python is a fairly good language for teaching programming skills. And, to a degree, they are correct. But there is a problem with saying that an environment is good for teaching programming skills, and saying that it is good for novice programmers who just want to create something without spending an enormous amount of time learning how to program. For those people, environments like HyperCard and HyperStudio were much more valuable. And, quite frankly, it's a shame that similar resources have not been developed for a computer that is about 100 times more powerful than the machines that HyperCard was originally developed for.

@sola: as much as you would like a free laptop, it just ain't going to happen. I'm saying that tongue in cheek since your $250 (was $300 a couple of days ago) laptop essentially amounts to giving you something for nothing. Think about it this way: let's pretend that you sold a million of these for $250 a piece. That represents $250 million in revenue. The last number I heard for manufacturing costs is $203/unit. How many companies are willing to consider a 20% return. Now factor in other costs: getting it to the customer, replacing defective machines, providing an acceptable level of support. The return is now substantially lower than 20%.

Yes, yes, and more yes... but:

You know, the more I think about the "can it run windows?" thing, the more I think about it from a developing world perspective. People aren't dumb and they know that in order to get a job their kids will need to know how to use windows and windows tools (MS Office anyone?). We may loathe it, but it's out there. Especially in developing countries where there is no other alternative. So, what does that leave us with? Well, there is a need to give users/clients/customers what they want to be successful -- or at least that's what they told me in economics class. Maybe the XO needs to take a page from the Intel Mac and be a dual boot computer? I mean, then the teacher can harness the power of both the XO as a learning space and Windows so the kids can get a white collar job in a McOffice down the line.

Just a thought.

"People aren't dumb and they know that in order to get a job their kids will need to know how to use windows and windows tools (MS Office anyone?)."

Yes, I heard that before: "The children should learn MS Dos, that will be their future OS when they have to get a job. What can they learn from Macs that they could ever need?"

And so it was from 1985-2008. Whatever you learned from MS products has been less than useless 10 years later.