I wuv you.
You are currently browsing the daily archive for March 27, 2007.
A few weeks ago, Julie Bryce came up to OLPC from Red Hat in North Carolina and did a series of interviews with people inside of the OLPC organization that are doing the on the ground work to make the project successful. The first part of that story can be found here. You get to hear from Mary Lou Jepsen, Owen Williams, John Palmieri, David Woodhouse, Marcelo Tosatti, Walter Bender, and the wonderful Eben Eliason. It’s just a taste, of course, but a good one.
Greg DeKoenigsberg, based on a lot of good work by Owen Williams, put up a Red Hat Magazine article about how to build a simple activity for the OLPC XO laptop. The article itself is good, of course, but I want to point at something deeper here. Something that been pretty fundamental to how we’ve been building out the infrastructure for the interface on the laptop. Notice how incredibly simple it is to write an activity. Set up one descriptor file, one icon, and a source file. Copy those files into a directory on the laptop and you’re ready to go.
You don’t have to learn the entire RPM build system to be able to get an activity on the laptop. You don’t have to put random files all over the place on the file system nor do you have to know the details of %configure, or how the macro system works or how you need to re-run ldconfig if you happen to use a library. The cognitive load is shifted from learning everything about how we package software to actually writing the software itself. You also don’t need root access to install software and, holy crap, it’s trivial for me to give you a piece of software that I have on my machine.
The choice of Python + Gtk is a part of this story as well. Python and Gtk are about as close as we can get today in terms of being able to make something that’s super-easy to get started with. I think we could get better. It seems that squeak and tools of its ilk are on that path, but they march to a different beat in a lot of ways and aren’t integrated enough with our environment to be considered as of yet. But that’s not really the point. The point is, how low can we make the barrier to entry? How can we get people kicking ass as fast and early as possible?
Over the next few weeks we will be continuing to make it easier. We’ve gotten to the point where sugar is stable enough where other people might be able to hack against it and develop more activities. To that end, we’ll be releasing some livecd-like utilities on a regular basis so people can get some hacking done on top of it. It’s not just about sugar the environment anymore, now it’s about the ecosystem around it. Curious and adventurous folks (like Owen Williams, for example) have already been discovering what it’s like, with all its up and downs. But now we want to open it up even more.
More to come!
interesting stuff