Chris Blizzard has been involved in the open source software community for over a decade. He has worked extensively on Mozilla's Firefox browser from the time it was open sourced by Netscape in 1998. He then joined Red Hat in 1999 because he "thought it was going to be interesting". At Red Hat he has been in a number of roles - initially building its next generation of support tools, then in its desktop group supporting and promoting its web browser solutions through to today's effort where he has worked on the customised operating system and Sugar User Interface for the One Laptop per Child project.
Before he arrives in Australia for linux.conf.au in January, Blizzard took some time out to share some thoughts with Howard Dahdah.
You are working on making children's lives better through technology. Tell us how you personally, and Red Hat, got involved in the OLPC project.
I think that answering that requires a bit of background. One Laptop per Child, as I'm sure you know, is a humanitarian effort. It's not a business and I don't think that they went out of their way to just build laptops -- it's just that no one was building laptops that could reach that huge underserved part of the world.
Red Hat as an organization participates in humanitarian efforts when we have the chance, but we're not the largest company in the world. We're not a multi-billion dollar company that has the resources to rebuild a city or buy vaccines for the children of an continent. Instead we have to pick and choose the places where we can have the most impact. One Laptop per Child is a place where Red Hat's unique engineering talent -being the largest and most successful open source company to date - can be mixed with our willingness to have a positive impact in the world. Our core competency can be used to positively affect the lives of millions of children all over the world. How many companies get the chance to say that? Or, for that matter, individuals?
My personal story isn't as grand or as interesting. My good friend Jonathan Blandford and I were running the desktop group inside of Red Hat. My skill set, ability to communicate and collaborate, long history of working with open source communities and being a part of building successful organizations made me a good fit to lead the efforts for Red Hat in this endeavor.
OLPC uses a cut-down version of the Fedora core. Can you tell us what you have done to that core and what your aim is for the final OS?
We're using Fedora Core 6 as a base for our operating system for the moment. We might move to Fedora 7 when it comes out because it's going to include a lot of things that we're going to need: faster startup time, better init scripts, updated libraries and the support of the community as a whole. In fact, a lot of what we're going to be doing in One Laptop per Child will hugely benefit Fedora 7 when it's released. As an example, we've already seen some changes in yum (an rpm-based update tool for Fedora) that should double its performance as a result of testing it on this platform.
What we've had to do is to pare down a lot of the packages that aren't really required. We don't need support for RAID, we build our own slimmed down kernel, we don't include a lot of random server-related utilities, etc. From there we've been able to build back up into something that will look like a true desktop instead of just putting a desktop environment on top of something that still at heart acts like a server. What we end up with should be pretty interesting, and I hope will lay down the model for a large number of other Fedora-based derivatives down the road.
You have been working on the Sugar UI for the OLPC. Can you give us a brief rundown on what you have done with it from the start till today?
I have to first say that I am only part of a team that has worked on the UI. A lot of the original ideas come from Seth Nickell, who will probably kill me for using his name, and Bryan Clark. Both of them did a huge amount of thinking and design work back in 2004-2005 along with Havoc Pennington to ask the question: if you were to design a client experience with a particular audience in mind, what would it look like? A lot of the ideas that came out of that are in our UI.
Walter Bender, who is the President of OLPC, along with an excellent design firm out of New York called Pentagram, and Marco Gritti, one of Red Hat's engineers, also contributed heavily to the design. The zoom metaphor, the idea of the frame and the models for how to represent mesh networks came out of sessions with those folks. This has been a collaborative effort.
That being said, we've gotten a good bit done in the short time frame we've had to work on it. The frame basics are working reasonably well and you can launch and navigate through activities. The mesh view, the friends view and the activity view are working, but not yet quite up to production quality. We've got a few sample activities to work from, including a Web browser (based on Gecko, the underlying engine in Firefox) a music player, an RSS reader, and the eToys environment.
But we've got a lot of work ahead of us. The journal, which is one of the most critical and interesting pieces of the environment that replaces the desktop, files and folders metaphor, has just now been started. We should hope to see that up and running in its most basic form over the next month or so.
Performance is our largest problem right now. There's a lot of low hanging fruit that we need to go work on. Over the next couple of months that will be near the top of our agenda. Boot performance is slow, individual activities take too long to start up and there are parts of various activities that are too slow. But we know where some of those issues are and we'll be making progress along with the rest of the open source community to solve them.