XO Laptops To Ship with Fedora or Windows XP


In a shift of strategy revealed in a Q & A from XO Camp, One Laptop Per Child will be dropping the Sugar User Interface in future XO laptop shipments for a version of Fedora or Windows XP. Sugar will only be one application of many. Let's hear Nicholas Negroponte explain the change in his answers to olpc community questions:

Nicholas Negroponte of OLPC
  • In what way will OLPC support the community in this transition? Will OLPC keep improving and debugging the software of the already delivered XOs?
    OLPC will move to a Linux desktop that will run Sugar as an application. Fedora 10, which shipped in November, is a major step in this direction. It's a standard Fedora distribution that will boot on the XO-1 and includes Sugar as one of the desktop options. OLPC is working with the Fedora Project and Sugar Labs to provide a more tightly-integrated and better-performing version of this solution in the Fedora 11 release.
  • Will OLPC open other options for software platforms in addition to Sugar?
    We announced the dual boot Windows in April; alternative Linux options exist already. Fedora 10 allows other desktop options, and Ubuntu and Debian distributions customized for the XO are already in wide usage.

Now Christoph doesn't see a real story there. He sees SugarLabs working hard on getting Sugar upstream into other distributions such as Ubuntu, Debian, Fedora, etc, and onto new platforms with "Sugar on a Stick", so its shipment with XO laptops is not a make or break for its survival.

But I see OLPC's separation from Sugar, its willingness to ship with regular operating systems, instead of focusing on an educational interface, as the last straw breaking OLPC from its original constructionist mission. The One Laptop Per Child organization has truly changed.

When Nicholas Negroponte says OLPC's mission is to

"eliminate poverty through education, by providing the means for children to learn learning"

He's not talking about a clock-topping hot educational device as the means. He's not talking about constructionism as the goal. He's only talking about "providing one connected laptop to every school-age child," even if its running office automation software. His focus is on shipping as many XO's as possible, not on changing education.

His new vision for OLPC? The Dell of the developing world.

Related Entries


I couldn't disagree more with your conclusion: that OLPC is more limited in scope by this change. It expands the scope. If Sugar is still there, then it's still there.

Personally, I found Sugar unusable. We have Fedora 10 on our machine now, and the machine is much better behaved and useful.

OLPC may be getting smarter.

After reading the blogs, reviews and forums NN and crew may finally becoming aware that Sugar's not really ready for prime time, yet. Fedora, Ubuntu and, yes, XP, etc. are further along the evolutionary path and have more and stronger connections to the potential user population.

Sugar can still be an application that runs along side or within one of the existing OSs.

As a stand alone OS Sugar still has a considerable way to go.

Consider all the educational, etc. software available for the new target OSs. Hard to resist. Go for it OLPC!

So are y'all saying that Sugar was a mistake? That the XO should have shipped with Ubuntu (my preference) and offered different applications, instead of Sugar?

Would that have made the XO as revolutionary though? Or just reinforced the impression that its a cheap laptop vs. a learning device for children?

The physical features of the XO, ruggedness, wonderful screen, battery life with rapid recharge potential, etc., remain the same regardless of the OS.

Ship with Ubuntu, add a mesh neighborhood module, a friendly file management/journal module, ebook reader, flash video, and web access via Firefox, etc. all adaptable for education levels over a person's lifespan and you're good to go.

Build "Sugar" apps so they run on Ubuntu, Fedora, etc.

Why reinvent a (not fully round) wheel when there are several reasonably formed wheels rolling along already?

PS: Oh, yes, IMO, if the XO is to have a significant life going forward it needs a faster processor and options to upgrade RAM.

Its limited memory and processor speed are main factors, I suspect, in its poor 2008 G1G1 showing.

But, Wayan, the king has been naked since 2005...

I don't think anyone who has messed with the XO-1 will think its just another laptop, for the simple reason that is RADICALLY underpowered in that respect. Comparing it to the eeePC shows this rather quickly. Once you factor in the rather non-upgradeable nature of the XO this becomes even more apparent.

One must look at the XO as it was intended a low cost -purposed- computing platform, otherwise it is just another poorly spec'd low power notebook.


Ned, I do worry that a Fedora XO would've (and will be) seen as "just another poorly spec'd low power notebook." I mean it takes 8 minutes for the XO to boot Fedora.

OLPC can make it boot faster, but there is only so much you can do with a Geode, which is why I thought Sugar a better fit.

"Ned, I do worry that a Fedora XO would've (and will be) seen as "just another poorly spec'd low power notebook." I mean it takes 8 minutes for the XO to boot Fedora."

You talk about Fedora 10 which aims to support XO out of box as possible and providing bugs report. The interview mentioned about Fedora 11 and olpc started to use Rawhide branch for their development. The goal is to reduce duplication as possible and rely to Fedora which is the upstream project.

"OLPC can make it boot faster, but there is only so much you can do with a Geode, which is why I thought Sugar a better fit."
See http://fedoraproject.org/wiki/OLPC/Packages_for_F11 where is an effort to integrate OLPC effort in Fedora.

Companies usually avoid hiring academics, past a couple of years their PhD because, despite their grate ideas, they can not function outside the academic environment resulting in mess ups (to be nice...). OPLC is a proof of that.
So why would anyone get an OLPC machine by now? Sugar is gone, mesh is "forgotten", power management is still "on the works", price is similar to more capable machines, and what is left is the screen and the dead processor contributing to a limited power consumption.
So when the Pixel Qi screens start shipping with other machines in few months, what exactly OLPC is going to have going for it over any other hardware/software organization, if not the educational package and expertise? Management, manufacturing, marketing, installed base, implementation expertise, local infrastructures, successful pilot programs, grass root support? What?
Maybe a new processor with 1 Tflop/mW is in the works... Who knows...

link that works:


go to feb 5

There's a lot of news in the Q and A.

Just a few things that stood out for me:

-- 500,000 backlog and new 100,000 order from Ruwanda

--2 years til X0-2 is out. I guess Pixel Qi must have run into technical problems.

--less than three months til dual-boot deployments.

--zero-cost connectivity is to be from surplus capacity in a number of forms.

The XO-2 development is separate from Pixel Qi, which is talking about having its screens in devices by the middle of this year

The question here is just what role Pixel Qi is playing in the XO-2. The FAQ at pixelqi.com says

"While Pixel Qi has said that it believes a $75 laptop will be ready in 2010. However Pixel Qi isn't making a $75 laptop. We are making the screens for it."


"Pixel Qi is providing OLPC with products at cost."

But Mary Lou's plan to get the cost down to $75 is to build the computing circuits into the display, and that would be the most innovative and hence difficult part of designing the X0-2, so I am assuming that is where the hang-up is. But maybe I'm wrong about that.

Wayan, I am pretty skeptical of your assertion here. OLPC laid off Erik Garrison, the engineer who got F10 running on the XO, back in January.


Don't confuse Nicholas Negroponte's assertions with my own. He's the one talking about Fedora as the XO-1's future operating system. I have my doubts about this strategy when he seems more focused on XP than any other OS.

Sugar is good riddance, it is a pity that they didn't ditch it 1-2 years ago.

Yes, the XO is seriously underpowered, but an optimized desktop Linux version could run it well.

My Xfce equipped Ubuntu Intrepid works mostly fine on the XO but performance is admittedly lacking. It can be used well for one purpose at a time (browsing/bittorrent/ebook reading/abiword/gnumeric) but running more than one app in parallel makes it extremely slow.

A Puppy Linux-like distro may be able to squeeze more performance out of this.


"OLPC is moving[...]" makes me think that even if you're right, it'll be years before Sugar/Fedora is replaced by [other desktop env]/[linux distro]. Plus, NN's not clear on what he means by "as an application". So I think Christoph's right and you're getting excited over nothing.


Sugar's really good for educational use, and all the feedback gathered backs this up. It's not designed as a new desktop environment for adults, so it's not really surprising that it's not meeting your expectations there. Do you go to Lego conferences with bricks saying "this brick is much better for building than any Lego brick"?

"Do you go to Lego conferences with bricks saying "this brick is much better for building than any Lego brick"?"

Actually, children should not use lego bricks. To be prepared for their future jobs, they should start with real bricks and cement in primary school. ;-)

Actually, I also think Wayan is overdoing it. There is a lot to say for the children and teachers to have the opportunity to switch to another desktop if they feel like it.

I do not see how this is different from the master development key. If a child wants to reflash her XO bios and HD, she can do it. So why should she not have the ability to switch to XFCE if she wants?


Because they are switching to things that are absolutely definitely not XFCE.

"Because they are switching to things that are absolutely definitely not XFCE."


But it is one thing to add more options, like Xfce / Fluxbox / Enlightenment / Blackbox / Openbox / Afterstep / FVWM / WindowMaker.

It is quite another one to remove all choice in one swoop, by using Windows.

I do know that "Choice is bad" is MS' mantra.


And this is my real fear - that Negroponte is actually thinking XP as the Os of choice, but including Fedora as an afterthought for the FOSS crowd. To Bryan's point, they laid off the Fedora engineer and its Teapot, not OLPC that optimized Ubuntu for the XO.

In my opinion, we’re going backward. Sugar is a GUI more adapted to learning contexts than any desktop metaphor. Admittedly, Sugar is under development, but its interface goes far beyond some considerations of user-friendliness. The Laval University in partnership with the Teacher training school of Libreville tried an experiment of use with pupils of 10 to 12 years old whom never used computers before. In less than one hour, they used Sugar easily. Let us not forget the teaching principles behind the design of Sugar. The affordances present in the interface supports the learning by doing and collaboration. Never a desktop metaphor will be able to develop collaboration like the Neighborhood and the Group views.

holy mac and cheese this is not good.

its one thing to be duel booting...you please both sides that way, and get this whole spectrum of possibilities. That is, on the one hand you are able to use both the OS that many businesses and schools and what not use, and on the other hand use an OS that allows for anything to be done to it and is perfect for teaching young children on.

to do away with sugar though...not good. But you know what? this is a zero budget NGO. we need to just keep on trucking, and adapting sugar however which ways are needed in order to allow it to be used anywhere and everywhere. We need to continue developing sugars gui, as well as its reliability and number of quality activities. At the same time however, it needs to be able to run as an emulation or software on the XO under windows...

is that even possible given the low power of the XO?

i gotta say guys...i'm worried.

Wow, now this software discussion is an amazing read:

cjb: we've outlined a plan about how you might be able to keep developing software even after you lay off almost all of your engineers (I consider being part of generating this plan to be the crowning achievement of my career so far, frankly, because the idea of continuining development right now is so manifestly ridiculous.)

mucha: The nutshell: we finish 8.2.1 as planned, and merge into F11 (or F12, if not finished by F11-time) for future software releases; at that point, Software Is Fedora's Problem and OLPC is along for the ride.

"His new vision for OLPC? The Dell of the developing world."

Okay, so why can't we just go to olpc.com and click "Buy"? Or will OLPC be "The Dell of developing world governments that are prepared to buy at least 10,000 units"?

Last year I pretty much recommended this as the course of action:


Put a more conventional Linux distro on the OLPC, so that regular Linux apps will run unmodified, and split off Sugar so that it can run on any platform.

Except, of course, no one is working on that. Fedora barely boots on XO, and it does not leave nearly enough resources to run Sugar. The whole thing is about pushing XP on XO.

I've heard the windows deal fell through, because they weren't willing to share the advertising opportunity with microsoft. Why not run DSL on it?