Halifax native Ben Armstrong is the project leader for the Debian Eee PC project. The goal of the project is to prepare a version of the Linux distribution Debian that is fully compatible with all models of the ASUS Eee PC netbook line. Armstrong began the project when he purchased the original model of the Eee PC a few years ago and has since grown the project globally.
ACOSS recently interviewed Armstrong by email. View the interview below.
Please give some background on yourself and your use of open source software.
As a kid whose dad had been fascinated with computers since the fifties, it may have been inevitable that I should grow up liking to play with them myself. When we puzzled together how to implement Conway’s game of life in Pascal on the university mainframe, and then took a trip to campus to pick up several generations of line-printer output from the game, I was hooked. After that, I was a sponge for any material I could get my hands on, so by the time the eighties rolled around, Linux and free software were quite naturally my new playthings. I’ve been a Debian developer now for nearly as long as Debian has been around, and have raised my whole family using it. Every opportunity I can, whether at my job as a professional software developer, or in casual conversations on the bus, I look for ways to share this passion for Linux and free software.
Why did you pick the ASUS Eee PC line of netbooks for this Debian project?
Quite simply, it was the first. For years before it was announced, I knew I wanted a device like this, so when the model 4G finally hit the market, I snapped up two: one for myself, and one for my wife.
Why did you choose Debian over other distributions?
It was the natural choice, having been a Debian developer since the late nineties. While the custom Linux OS that ASUS contracted to Xandros to be made for the machine was Debian-based under the hood, I knew the average user was going to have challenges getting Debian itself running. I wanted to ensure that my favourite OS would “just work” on these nice little machines.
Where is the Debian Eee PC project at right now, what goals are there still to reach?
We’re now well advanced in the preparation of the Debian 6.0 “Squeeze” release. We find Squeeze works well on all older Eee PC systems and many newer systems as well, though not entirely without quirks. It’s mostly the newest systems that give us trouble, as our upstreams have not yet finished support for them.
We have just one, simple goal, as stated on our home page: “full support for the ASUS Eee PC in Debian”. If you consider the broad range of systems ASUS continues to release under this name, that’s a goal we might not fully reach for a long time. We’re always playing catch-up. ASUS has also made some mis-steps by sometimes choosing hardware with no free drivers, as they did with the Intel GMA 500 “Poulsbo” graphics chipset, and they often bring new models to market without complete support from our most important upstreams: the Linux kernel itself, of course, and Xorg. That being said, none of these problems have been insurmountable. We get tremendous support from our users and team members, who tirelessly assist in sharing test results, filing and fixing bugs, and keeping our wiki updated. Thanks to their help, Debian is in good shape for supporting all models of Eee PC over the long haul. This matrix will give you a rough idea of where we are today: http://wiki.debian.org/DebianEeePC/Models .
What sort of issues does the Debian Eee PC project have to tackle most, and how successful has the team been at overcoming them?
Well, first it’s the usual set of challenges that any free software project has to tackle: guiding new users through the rough spots as they try out things and keeping their interest long enough in working through problems so that they can enjoy positive results in the end. This can sometimes be a trial for the impatient, as some of the more irritating and persistent bugs are in drivers that we are not equipped to fix ourselves. But often, we manage to find workarounds while we’re waiting on real solutions. The other big challenge is staying on top of support for such a wide variety of hardware, where our developers only have a handful of the models out there. Our users are a tremendous asset here. I think we do fairly well on both fronts, as we continue to hear good reports from satisfied users, and we’re managing to attract and retain people who are helping out with the newest models that we don’t have ourselves.
Can you explain the difference between your project and other projects for the Eee PC like Eeebuntu?
From the outset, we knew that our project could only be successful with a reasonable expenditure of effort on our part if it meshed perfectly with Debian instead of forking a splinter distribution. Desktops tailored for the small display of a netbook, or hardware-specific control panels may be nice ideas, but neither has anything to do with making a distribution. In fact, I’ve never seen anything in the Eee PC hardware that warrants making a whole new distribution. You ought to be able to just install Debian itself on your Eee, and then how it looks after that is up to you. Debian has a wide range of choices that should be able to meet your needs, and everyone’s tastes are different.
So, while we’re busy working on infrastructure, which is quite important to making sure Debian “just works” on these machines, just like any other computer, other projects focus on the differentness of netbooks, and strike out to release something new-looking, distinct from the mainstream. I don’t think this serves the needs of users well. There will always be an initial flurry of interest in such distros for how they look and feel, but after the “shiny new” wears off, people find there’s a lot of grunt-work to keep a distribution going. Interest wears off, releases start to lag behind, some of the splinter distros die off altogether, and when they do, users are left in the lurch. By contrast, we want to support you over the long term, not just impress you with slick packaging, and tapping directly into Debian itself instead of striking off on our own helps us deliver that. I think all of the feedback I have received so far indicates that we’re succeeding.
What has the response from the Debian development team been towards the Eee PC project?
The Debian project’s response comes from two main directions: other developers who have purchased or are thinking of purchasing a machine, and the kernel and Xorg teams. The former have provided us with valuable bug reports and patches, or have gone one step further and joined our team. The latter have been responsive to the Eee-specific issues we have raised, and it is largely thanks to them that we are able to continue to provide such good support for the Eee PC line of systems within Debian. So, the Debian project as a whole has been quite supportive of our work.
Has ASUS ever contacted you or the project team directly for any reason, and if so why? Have they assisted with the project at all?
This is a troubling question. I don’t want to blame ASUS entirely for the failure in communication here. They did, at one point, contact us, setting up people at ASUS as a team to communicate with us, but beyond that initial gesture, we did not hear back from them. I wonder now if we had made more of a united effort to bring them our concerns, to overcome the challenges of differences in language and free software vs. corporate culture and values, things might have turned out differently. So while some small efforts were made on both sides to talk, apparently it was not enough, and ultimately failed.
The Eee PC line initially came with Linux installed by default, but has since moved away from that to Windows pre-installs, what are your feelings on that?
While the party line on that is “we are just providing what the market is asking for”, I think it’s a shame ASUS didn’t really “get” how to work with the free software community. Had they really known while they were still developing the product, they’d have worked with a mainstream distribution from the outset and maybe that would have helped Linux to make a bigger impact in the marketplace. But it seems that their vision was for a different user experience that the mainstream distros didn’t seem to address at the time, and this drove them to a lesser known but more commercially friendly distribution vendor, Xandros. I think I understand their motives, here, even though I disagree with their approach. They thought they needed a distinctly different look and feel to sell this machine to the public, which would therefore require a high degree of customization and needed to partner with some commercial entity to deliver it. But those are only surface things that could have as easily been implemented directly as free software on top of Debian, no matter who got the contract to do the work. Then they would have tapped into the huge community support behind Debian. This would have reduced the cost of support to ASUS and helped them to remain competitive with their Linux-based offering.
That being said, the failure to provide a successful pre-installed Linux option on the Eee is not the end of the story. I hope that some day, rather than shipping with some default OS, they might offer a cheaper, OS-less edition that anyone could install their own favourite OS on. This is certainly something I would buy, and I think a lot of others would too.
What comments can you make on being a project leader in the Maritimes?
Sometimes I feel the physical isolation from my teammates, as face-to-face meetings can really help motivate and help a team to gel. But mostly the distance is a non-issue, as we stay in touch around the world daily via mailing lists and irc. I’m particularly attached to Nova Scotia, having grown up here. It’s a beautiful province with a lot to offer, the pace of life suits me well, and I can’t imagine living anywhere else. I’d love to get some of my peers to move up here or at least drop in for a visit.
What challenges and rewards are there to leading a distribution project?
People expect us to be able to solve all of their computer problems once they install Debian on their Eee. But we’ve deliberately set out not to be so much a “distribution project” as a subproject working within a distribution to make it better for owners of this hardware. So once users have installed Debian on their Eees and start to ask general Debian configuration and maintenance questions, I have to gently redirect them to general Debian support channels and make sure we’re not dragged off focus. On the flip side, our members are generous to a fault with their time. It’s tremendously rewarding to see all of that generosity pay off when users express to us their thanks for doing a good job. But beyond just the support, which consumes most of my time — just keeping tabs with users to stay on top of what’s broken and what’s still to be done on all of those models, and checking back to make sure we’re following up on delivering solutions, as they become available — it still amazes me to see how far just a little bit of encouragement on my part goes to keep things moving along. It’s to the point where most of the time, I can just let the whole thing coast, and everyone stays busy getting things done fine without me.
Are there any aspects of leading a project like this that you dislike?
I’m really quite a lazy leader. I feel most motivated when the problem-du-jour first presents itself and is fresh and interesting. The duller, but no-less-important parts of the project are ensuring you stay on top of what needs to be done, delegating jobs, recognizing areas that are being overlooked and putting them back in view so that they can be solved, and generally keeping a regular flow of information both back to your team and users, and to other developers in the broader Debian project. When it comes down to it, I don’t dislike doing any of these things, but I feel the guilt when I recognize I’m letting things slip, and that’s unpleasant. I follow a cycle of frenetic activity followed by periods when I let things slide, and then I have to kick-start myself back into action. Still, that’s just a part of me I have to live with anyway, and it has never discouraged me to the point of wanting to quit. The project is still fresh and fun after over two years of work on it, and despite my obvious flaws as a leader is still chugging along.
Is there potential in the Maritimes for more projects like Debian Eee PC to be started or contributed to?
Sure! Why not? We have a gorgeous province. We have great, friendly, smart people … We have excellent Internet infrastructure, and that’s a boon to staying in touch with the broader, international free software community. I don’t see any reason the Maritimes shouldn’t be brimming with free software projects.
ACOSS thanks Mr. Armstrong for taking the time to tell us about the Debian Eee PC project. You can learn more about Debian Eee PC at the link below. You can follow Ben Armstrong on his blog syn.theti.ca and on Twitter here.
Original photo by ragwing, Creative Commons Attribution-Noncommercial 3.0 Unported