For many people, Ubuntu is Linux, and Mark Shuttleworth is Ubuntu. It might come as a surprise to learn that the prime mover behind one of the most successful and visible Linux distributions out there isn't entirely comfortable with that. I jumped at the chance to sit down with Mark for an hour on Tuesday morning while at OSCON and ask him about that, and many other, things.
InformationWeek: What do you think about this whole popular perception that Ubuntu is Linux?
Shuttleworth: It's actually an offensive idea to me. We take the view that distributions should be humble in what they claim they do. The reality is that we distribute Linux, and we're at our best when we do that particular thing efficiently -- when we most effectively celebrate the work of others. If someone decides they are the final arbiters of what Linux is as a whole, that's trouble, especially since the distributions have a responsibility because they make key decisions about what the end users receive. We do this -- for instance, which version of Firefox to use -- and it's a privilege to do that, but we have to recognize that the real work and expertise takes place upstream.
This is important for both Ubuntu and Canonical. We celebrate everyone else's work, and focus our energy to deliver their thinking to other people in the best possible condition. It's also important that diversity be maintained in the Linux space; I would be very uncomfortable if people stopped using other versions of Linux, since "vanilla" Ubuntu by itself obviously isn't for everyone. We encourage adaptations -- Xubuntu, Ubuntu Studio, and so on -- each of those meeting someone's particular needs better than the default. I would hate to be in the position MS finds themselves in, of effectively giving people no alternative, and being responsible for everyone's needs.
We're not the source of the majority of the innovation, and that's something very important to remember. I'm in Free Software because I think it's a richer source of innovation than the proprietary development model. Reduce free software down to, say three institutions [this was from my own example -- Ubuntu, Red Hat and SuSE], and that reduces the total innovation possible.
InformationWeek: I read (and commented on in my blog) an interview you gave for Datamation where you cited the Mac as being a gold standard for user interfaces, something to aspire to. That said, the Mac was designed in and is a product of a totally closed-ended environment -- how do you reconcile that with the open-ended way open source is produced?
Shuttleworth: If you think about it in "cathedral vs. bazaar" terms, Apple is, in a way, the ultimate cathedral. You can't produce free software the Apple way. To that end, take a look at the GNOME interface guidelines - it's an open process with open documentation; by doing things that way you can draw on Wikipedia-like collective wisdom that you probably wouldn't have in a closed environment.
InformationWeek: One of the advantages of a closed environment, though, is that it's easier to do things autocratically from the top down, to impose a consistency that might not be there in a more democratic setting. How might you prevent the openness of the design process from producing something that's just a horrible series of compromises?
Talking about the importance of Windows in free software is controversial. We have to continue to find people who drive new ideas, and give them enough space to explore those ideas where those features/ideas can deliver. [One example of this: a feature in Ubuntu 8.04, which allows a full installation of Ubuntu to be placed on the same partition as a Windows install without reformatting, was not developed within Canonical.]
InformationWeek: How do you plan to draw people towards Ubuntu as contributors? (As an aside, Microsoft's model seems to be to watch other companies develop things, then step in and buy them.)
Shuttleworth: Large companies tend to stifle innovation by dint of their size. Free software has the same danger -- that effectively we will end up with a couple of companies that become dominant. One company can't take everything, and then be under some moral obligation to underwrite everything else. A key part of the Ubuntu framework is to allow each component of the ecosystem thrive.
We could never do what Red Hat did [i.e., split Red Hat into RHEL and Fedora]; we feel that's like locking part of it up. Actually, I'm surprised that we haven't seen more public consequences for MySQL for bifurcating. It suggests they communicate well.
I think it's impossible to avoid change, but we've committed ourselves to defend our core values. People have faith in us that under any set of circumstances we can defend those core values. That's bound to be good for us, as well as the upstream folks.
InformationWeek: How do you plan to get continuing participation from developers?
Shuttleworth: A big theme for us is helping the free software community make the most of its assets. What we have (versus, say, Microsoft) is a lot of passion, a real overlap between personal interests and jobs. People who hack Firefox or Apache are by and large personally interested in that. As far as drawing out participation from developers - we want to create tools to develop that process. Launchpad is a good example of this.
The tools we particularly focus on are support of software processes like bug tracking. Launchpad doesn't pretend that bugs only live in one place; Red Hat's work on this bug is just as important as ours. A key goal for us is to move code around as easily as a bug report around, or resource translations -- take work done here and make it relevant there. Every time we fail at that, we introduce a level of friction -- we're tripping ourselves up. We have a huge incentive to push all of what we do back upstream.
Another thing we're proud of is the Ubuntu Code of Conduct. When someone just joins a community, you don't know if they're brilliant or not. The big thing in Ubuntu is to figure out how to encourage such people to participate, and to channel that, and not waste people's time -- but also not exclude people unnecessarily.
The key for quality free software is architecting for participation -- leading the community in a particular way, encouraging certain kinds of communication. I would like it so that with one command, any Ubuntu component could be branched and republished -- so from a participation POV, the world is flat. We have to harness our strengths to provide something better than what the proprietary software world can offer.
InformationWeek: At a panel I was attending the other day [the report for which I'm still writing up] one panelist with a background in sociology stated that she was surprised by how well collectives worked in software when they had generally failed most everywhere else.
Shuttleworth: I think the reason they've worked here is because something collectively owned is worn down by that collective. Digital assets, however, are not undermined by sharing; on the contrary, they're made all the richer.
InformationWeek: Finally, who are your personal heroes and why? [For this last question, I filmed his answer:]