It looks like the interview that I did yesterday with Jim and Walter is up and it’s one of the better articles I’ve seen lately. We really are going to be able to do an OS that takes up less than 100MB of flash. We’re really close today. John and Dan have been working hard to make that happen. (One package that keeps rearing its ugly head is perl. That sucker is really tenacious. Any time you move packages around it keeps showing back up like a bad check.)
But that’s besides the point. Jim likes to stress that we’re interested in taking what we have and making it faster. There’s a lot of low-hanging fruit there. But I want to not do something that’s only faster, but something that flat out better. That’s an important distinction that I think a lot of people don’t understand.
I just got back from Nat‘s keynote at this conference where he spent time talking about the history of the OED (fascinating story!) and how it’s one of the earliest examples of a collaborative open source project. At the end he was showing off the features of SLED 10 and proving how well integrated and polished it was compared to where we were in 1992 or 1998 or even 2002. And it is – hats off to those guys. I hope they do well with it. Were it not for issues of hardware integration and multimedia, I think we could beat windows out of the box. There’s little question of that in my mind anymore.
But I think one of the things that I’ve been saying inside of One Laptop for a while is that we do have a chance to change that model. Why do we let Microsoft, or Apple for that matter, define what computers have to be? Why can’t we move ahead with the knowledge of the past and create something really new and interesting? Something designed for the laptops, something that solves the specific problems of learning and the environment they will be deployed in. That’s where we’ve been going, and that’s what it takes to really move things forward, both for the kids and the industry as a whole.
It’s not about putting a “Linux Desktop” or “Free Software” into the hands of kids. That’s just a means to an end. We need to solve real problems for real people, and in this case it’s kids. And we can do better than we have. That’s where I think the focus needs to be when we think about how to build our software, how we take existing code that exists in the open source world (of which there is some great stuff!) and move it onto the laptops where appropriate.
Context matters.
I have tried the olpc software 20061026 edition through QEMU. I am impressed with the progress with the inclusion of software. What is only left is the polish. Keep up a good work.
Chris, thanks for doing the interview yesterday.
Alan Kay said, “Now you’ve got millions and millions of people who think that doing even the most trivial things on a computer is a sign of computer literacy. This includes parents, teachers and the kids themselves. But most of what is done is about as worthwhile as playing an air guitar.”
So I’ll keep reading the OLPC blogs to see what happens when the kids get a hold of these things.
Many cellphones running on Linux have a rather small footprint. Sure, they lack most of the applications, but I was wondering: would it be easier for you to borrow many of the specifications of the kernel for those devices and bring them back to the fedora/olpc project? How about the maemo project?
We’re more like a laptop than a cell phone and we’re running a stripped down kernel that’s still pretty close to a standard desktop one. We haven’t made any VM changes, for example.
We’re probably going to end up overlapping with maemo a bit. But we’re not the same kind of device as they are and have very different target audiences.
Thanks for your answer. I was wondering how much of your work is going upstream, either in the Fedora fork of the kernel and in the main kernel itself. What I wonder is how much we are going to see of the olpc implementation on regular kernels to be use on laptops. Most distributions now have kernel for the architecture used, and for desktop/server, but not much for laptops. It’d be great if custom kernels for laptops were developed, benefiting from your great work on the olpc.
All of the kernel work we’re doing for OLPC is going upstream. Right now that largely means drivers for the platform. Our config isn’t that different than a standard kernel, it just leaves out a pile of drivers. When we sit down and work on the suspend code and work on the memory compression that will also end up upstream and might be useful in other contexts.
What’s the chances for something like Squeak and Croquet working on it? I know Alan Kay is one of the prime movers in Squeak and would be interested if it’ll run within the OLPC scheme of things? I have my own program, OAR, which I’m trying to get running on Fedora Core 6, using Squeak and Croquet for reducing the recidivism rates within our nation’s various DOC’s. With a really inexpensive laptop, I’d have a better shot at success for it’s adoption. Our prisons are third world countries with 2.7 million men and women living in them. How about one laptop per convict? Ric