Canonical is pushing Ubuntu in so many different directions. On the desktop, it has introduced Unity; on the server, it's pursuing state-of-the-art ARM and cloud platforms; and it's even trying to get Ubuntu on to mobile phones and televisions.
Trying to keep track of how all this is going, how it all fits together and what's coming next is a full-time job… which is why we spoke to Jane Silber, Canonical CEO, whose job it is to keep track of everything.
Linux Format: Let's start with Ubuntu on the desktop. The big development is Unity, and a lot of its development has been driven by user testing. Could you talk about the motivation behind that and how you've gone about finding the people to do it? It's certainly not something a lot of free software projects have experience doing.
Jane Silber: No, you're right and I think it's one of the areas where we've really innovated and pushed the envelope in terms of that approach in the free software world. We didn't invent user experience testing in any way, but we are believers in a user-centred design ethos.
From exploring initial concepts with paper prototypes to actually getting people in and having them use the software - both during development and afterwards. Even after something's been released, we continue to test it and do milestone checks by repeating some of the same testing.
LXF: Ubuntu is a free software project done in the open, but with the testing how much do you try to do in the open?
JS: We absolutely release results of user testing, and will continue to do that. A lot of it comes up first on our design blog, design.canonical.com, so if you're interested go there first.
But we're quite public in those. There are some things we keep private in the early stages of development for competitive or customer reasons, but in general we'll do as much of that design work in the open as possible, and release the user testing results as well.
One of the other things that we've done is helping other projects learn how to do that kind of thing. At UDS, the Ubuntu Developer Summit, we'll run surgeries where people can come in and say, "this is my favourite open source project, this is what I work on, but I'm not really a designer. How can I make it better?" And these design surgeries will both give design ideas and do a little bit of user testing on their own projects.
LXF: The other big thing on the desktop for Ubuntu is the Software Centre, which is one of the things that really makes you stand out against other distributions. How is getting third-party application developers on board going, and how successful have paid-for apps been?
JS: It's going very well. We don't have 100,000 applications like some other platforms do, but it's growing at a healthy pace and we're seeing it really being a draw for the app developers.
For example, some of the big names - such as EA releasing games on Ubuntu, and just the other week we did a co-marketing, co-launch event with Humble Bundle. I don't know the stats right now for that, but in the first 72 hours after release, we saw 10,000 downloads from the Software Centre.
But that's our biggest first-72-hour launch of a new title. And new titles are coming in there all the time. We still have work to do. It's too early to declare it a success and go home. There's work around providing a more coherent SDK to allow people to develop for the platform.
LXF: That's a challenge on Linux in particular, with goalposts often moving.
JS: Yeah, and it's the embarrassment of riches. You have all these toolkits you can use, as opposed to the narrow selection that other platforms give. It's great, but sometimes that creates a barrier to entry where people just don't know where to start. And that's what we've done some work around, not limiting choice but providing a narrow on-ramp that app developers can follow. Things like Quickly, and http://developer.ubuntu.com.
LXF: Looking towards the server, I know recently that you increased your support life cycle to five years…
JS: Our LTS releases used to be three years on the desktop and five on the server, and we made two changes. One was bringing the desktop up to five years, and that was in response to enterprise demand.
And the second change we made was to make a stronger commitment around making those LTS versions available on new hardware as it comes out, so that people can get hardware refreshes and still maintain a stable software platform across the enterprise.
So we'll make the 12.10 kernel work with 12.04 and the 13.04 kernel work with 12.04. So if you need that kernel for hardware support, it will be available.
LXF: Red Hat recently announced it was increasing its server support cycle from, I think it was, seven years to 10 years; does Ubuntu feel any pressure to match that?
JS: We're not seeing that right now. I think the reason is the different use cases in terms of people who use Red Hat and people who use Ubuntu. And, interestingly, we're seeing pressure almost in the opposite direction.
One of the things that's happening in the server world is that everything cloud-related is so fast moving, it's not realistic to think you're going to do something now and want the same tools and software in 10 years. What we see is people wanting the stability of the base OS, but wanting new hardware support for one, and newer software for cloud-related activities.
So, they want the new OpenStack, for example, on a 12.04 LTS base, so that's another thing we've committed to do with 12.04. In six months' time, you'll be able to get the newest OpenStack. 12.04 shipped with OpenStack Essex; but when Folsom comes out, the next version, people are going to want that not just on Ubuntu 12.10, but also on 12.04, which is for stable production.
LXF: This week you're at Computex, and one of your big announcements is that you're demoing Ubuntu on an ARM platform. Are there any real-world deployments of that yet? Where has the demand come from?
JS: The motivation for it comes from scalable, power-efficient, low-cost energy drivers. There's real customer interest in it too, but it's very early and there's not… hardware doesn't exist in production environments yet. There are no case studies yet, it's that early.
People are excited about the promise of it, and hardware is starting to show up. Calxeda has demonstrated some hardware at UDS a few weeks ago, MiTAC is the company in Taiwan that we've demonstrated an ARM server with in Taipei, and HP has announced its project Moonshot, which will be its ARM server. So there's real hardware, and it's being used in largely test and development areas, where people are exploring the workloads, are exploring how to optimise for it, but it is very real. But it is very early days.
LXF: Could you give us an idea of how many people are working on Ubuntu on ARM?
JS: It's hard to identify a specific number because there are people across the company, it's an integrated piece of the company.
Some people on the Ubuntu server team, Robby Williamson, work to ensure that Ubuntu works both on Intel and ARM hardware… we don't have a dedicated ARM team, we as much as possible treat it as another architecture that we have to support. So it's part of our QA team, automated builds, automated testing, etc.
LXF: I know you said you had no case studies on ARM, but are there any Ubuntu case studies, on the server or desktop, where you're really proud of what's happening?
JS: Oh, there are loads! Some of the more recent ones have been on the cloud. We've just released a case study with Mercadolibre, which is a South American cloud provider, which built a cloud with 1,000 Ubuntu nodes - so a good-sized Open Stack, Ubuntu-based cloud.
On the desktop side, we've recently done a very large deployment in South Africa in an education setting, which is in 1,600 schools - 50,000 Ubuntu desktops, which combine to reach two million learners. It's a shared resource. Basically, it's providing connectivity and a cloud-managed desktop across the Gauteng province.
We've done it in conjunction with a South African partner, and I think it's the largest open source deployment of its kind in the southern hemisphere. So we continue to see good, interesting growth, both on the server side, which is largely cloud, big data, web server scaling operations; and on the desktop it's large enterprises, government and private.
LXF: Moving on to Ubuntu 'elsewhere', on TVs and mobile phones… I guess one of the big questions is, everyone's sure you can get the software together, but the challenge most open source projects have is getting the hardware partners and the software distributed to customers. Could you tell us about your efforts to get partners on board?
JS: It is a significant effort for Canonical. Luckily, it's something that we've been doing for years now. A large part of what we do as a business, that we can provide to help Ubuntu, and that the community can't do, is improve those commercial partnerships and relationships.
We've been working in that industry, particularly in Taiwan and China, for many years and have very good relationships across the board with all of the major players. So we are working hard to bring both Ubuntu TV and Ubuntu for Android to market.
Our go-to-market plan is absolutely through a hardware vendor. We're not going to start making and selling hardware ourselves. We are working to line up those engagements to bring us to market, and because it's their product and not ours, I can't give accurate timelines on that.
LXF: This is a direction Canonical's pushed for in the past, in partnership with Dell on Desktops, and on Netbooks with Ubuntu Netbook Remix, and your efforts didn't pan out quite as you'd hoped. Do you feel you've learned from that and developed relationships that are helping you now?
JS: Yes, I think we have. We've learned a number of lessons. I think the industry has changed, the dynamics have changed. I think netbooks were a good idea that was overtaken, hampered by the hardware itself, and struggled to find its appropriate place in the market.
Was it a low-cost? Was it a highly mobile laptop? And aiming at those different markets would drive you to different hardware specs and different software loads on there.
I think one of the things we learned from the Netbook experience was the value of defining the software experience, and it was around then that we started defining Unity as a product and taking more control of the user experience… not control of it, but defining our vision for the user experience for the software experience on top of that, and being more opinionated. And I think that's led to a more compelling user experience and suite of Ubuntu products.
LXF: On the topic of being opinionated, with Unity and the community as a whole - there was a split. Do you think that what happened with Gnome 3, Gnome Shell and Unity was partly a result of your desire to be more opinionated about the interface and to have more control? That's not to sound negative, but you had a vision that you wanted to implement…
JS: I think that's absolutely part of it. It's unfortunate how things played out - things turned into a split, and we simply seemed to be unable to come to an arrangement to keep the communities together. I think it's easier for a community to move forward with a driving vision, and in general that's how Ubuntu works.
It's an effective community, and it is a relatively cohesive, goal-orientated community, but when there are different goals it becomes difficult to hold things together, and we found at an individual level there were different goals with what was happening with Gnome 3 and what we felt Ubuntu needed.
LXF: And do you hope in the next few years you might see other distributions picking up Ubuntu's work on the desktop and helping with the burden of that?
JS: I don't know if contributions will flow back in terms of code - we've always measured contributions beyond just code - documentation, mentoring new users, etc - and so if those other projects in time influence us in terms of design and best practice, that's an equally valid contribution.
LXF: Do you run the company as a whole on open source software and Ubuntu, or are there systems where you can't?
JS: We run the whole company on Ubuntu. This is a side annecdote, but we've just moved office. In the old office, at the entry there was a hand-scanner run by a Windows machine, and we hated it. We had to have a Windows licence! And there's no hand-scanner here, so now we have Windows licences for testing, but not as a piece of our infrastructure.
And our software infrastructure, it's largely open source. We use OpenERP for our financial system. We do use some proprietary cloud-based services; we use Salesforce.com, Google Apps for Calendar Sync, etc. We're not anti-proprietary software. But in terms of the software we run ourselves, it's open source. We have an internal OpenStack cloud that anybody can get access to and spin up any workloads they want.
LXF: Wow, so if people want to build some software, they can just use it?
JS: Yeah, and it's generated a lot of innovation in the company, and given people that agility to move quickly. Somebody will have an idea about how to do some automated testing, they can try it out and if it's successful, they can get it deployed through our normal procedures.
LXF: On Canonical's finances, how is your drive to profitability going?
JS: We're very comfortable with where we are. Our revenue is growing at a healthy pace. We don't talk about our numbers publically, being a privately-held company.
But we continue to grow, and are seeing interest in Ubuntu grow in the enterprise, particularly in emerging workloads and cloud environments, at the desktop level, and in the device space driven by the upswing in mobile devices and convergence of platforms.
That line between desktops down to phones is getting blurred, and Ubuntu as a platform can move elegantly up and down that.