Last summer, Mark Shuttleworth, the founder of the Ubuntu distribution and of Canonical, Ubuntu’s commercial sponsor, challenged the free and open source software (FOSS) to deliver a desktop experience superior to Apple’s.
Now in the process of introducing the first steps in meeting this challenge, recently Shuttleworth took the time to talk to Datamation about the origins of his interest in usability, the difficulties in implementing it, and the organization and road maps he’s creating to focus upon it.
To a casual observer, Shuttleworth’s focus on usability might seem less than a year old. However, Shuttleworth himself describes it as the latest stage in a lifelong interest.
“Thinking very carefully about what you are trying to deliver is essential in a successful process,” he says. “Personally, I’ve been fascinated by product design for a long time. I’m drawn to companies and processes that deliver great product. If you look back at the work I’ve been involved in elsewhere and in Ubuntu, there’s a consistent pattern of trying to make things simpler, clearer, and more useful.”
Shuttleworth explains that usability was the main reason for Ubuntu standardizing on the GNOME desktop. Specifically, Shuttleworth was drawn to GNOME because of its Human Interface Guidelines, which, while not officially adopted for many years, were unofficially a major influence on GNOME developers.
Shuttleworth describes the guidelines as “a rallying call for specificity and clarity in the software user experience. When we were picking a platform for Ubuntu, we wanted to pick a platform that already believed in the idea of making something easy to use — and GNOME met that challenge.”
With GNOME as a foundation, he adds, in making Ubuntu, “we invested a tremendous amount in how you put the pieces of free software together to form something that feels designed. I think that’s why Ubuntu has a reputation of being something very rich that you can install and that works out of the box, where the pieces fit together in an elegant way.”
Increasingly, however, Shuttleworth suggests that “we’ve reached the limits of that. So the question is: Where else can we apply this energy? And, at that point, I started to want to drive Canonical in the direction of being a community leader in design. So I articulated the challenge to the free software community to show that we would aim to show that you can deliver great design inside something that is effectively crowd-sourced.
“Now, I don’t imagine that bringing these two forces together will be easy, because there are elements of both that are inimical to one another. But I think that if we can unify those forces, then we can transform the free software experience from something of the ugly duckling of the software world to the powerful swan that it really is.”
The Challenges of Usability
Shuttleworth acknowledges that the goal of combining radical changes in the name of usability with FOSS community development is a controversial one. “The open source community is always fractious. [The proposal] certainly generated a lot of opinions, right? Sort of loudly-expressed opinions,” he adds as an aside.
A large part of the controversy stems from the fact that, until the last few years, FOSS has been written by developers for developers, with little concern for the user experience of those less expert than themselves.
However, Shuttleworth suggests that now, “More and more open source projects are interested in the idea of making something very easy to use, the challenge of making something beautiful. They’re starting to realize that it’s not a meaningless exercise. There are some elements in the approach that we are going to be taking that are already controversial and will be more controversial, but there is also a lot of support for the idea that bringing user experience to the front is essential if we want to grow the base of Linux users beyond those who are self-selecting for what’s different and what’s more powerful.”
Another major concern is how to implement changes on the scale Shuttleworth envisions: in a single giant leap forward, as KDE did with its 4.x series of releases, or incrementally, as the GNOME desktop project prefers.
On the one hand, Shuttleworth says that “The KDE team ought to be commended for the quality of the default 4.2.2, which is the latest version, and for tackling the challenge of making a transition. A major transition is a special thing — it’s very tough to pull off, right?” On the other hand, Shuttleworth also maintains that GNOME’s record of doing point releases every six months is also “an extraordinary achievement.”
Ironically, he observes, the two major desktops are now switching places, with KDE edging towards incremental releases after focusing on its major transition, and GNOME debating whether to abandon incremental releases in favor of a major leap forward in order to realize GNOME 3.0.
What Shuttleworth concludes from these contrasting experiences is that both methods have to be balanced over the long term.
“A really predictable release cycle is very energizing, as GNOME has learned,” he says. “Conversely, what KDE showed is that you can make substantial jumps forward if you have a well-run program and you’re bold in your vision. So putting those two together is really exciting. Within the open source desktop, we have the experience we need to make big transitions, and we also have the knowledge to have predictable, energetic regular release cycles.”
The question, of course, is which approach to use when — knowing that, no matter what choice you make, some parts of the FOSS community will criticize you and your timing.
Asked whether the problem is as much one of diplomacy as of software engineering, Shuttleworth replies, “There is an element of that. And the challenge is that we very consciously know that we’re not going to be able to give everyone what they want. I don’t know how that’s going to work out. I know that that’s going to be socially very challenging. But I also strongly believe that, in the absence of a will to make tough decisions, all of this will lead to failure.”
The Organization and the Road Map
Because Ubuntu supports all three major free desktops, Shuttleworth suggests that “Canonical is in a good position to help this process. Because we embrace the whole free desktop, we are able to drive an idea forward across the whole desktop.”
Since Shuttleworth issued his usability challenge, Canonical has been hiring employees to answer it, dividing them into three main groups: The Design group, which studies usability and makes recommendations based upon their work, as Matthew Thomas did in evaluating the Pidgin and Empathy messaging tools; the Desktop group, that writes the code to express the usability ideas that are approved; and the Platform team, that integrates new code into the desktop.
All these groups work closely with upstream projects, the producers of the software affected by changes. Together, these teams consist of about twenty-five people — a “significant focus,” considering that these comprise about ten percent of Canonical employees, “but only part of the broad thrust of what goes on in Canonical and Ubuntu.”
As you would expect in FOSS, community members play an additional role in meeting the challenge. Accordingly, Canonical and Ubuntu are active in developer sprints and conferences held by upstream projects. In addition, “we have a variety of public conversations in GNOME, KDE, and Ubuntu,” Shuttleworth says, referring, no doubt, to the regular rounds of discussions in project forums.
“I also expect that we will offer to open source projects the ability to come to London to work directly with the Design and User Experience teams,” he adds. “We recognize that, ultimately, the free software desktop experience is made up of the individual experiences of each of the applications that make up that desktop. So we want to engage with each of those projects and invite them to work with [our] usability professionals.”
What might be less expected, though, is that part of the usability challenge is being met privately. “There are other process that being developed internally in Canonical by projects that we have underway with commercial partners, and those I can’t speak about at this stage. Those will obviously only be available when the products themselves are available. But, in each case, we engaged partners on the understanding that, as soon as the products are available, we can move that code into the public area under a license like the GPL [GNU General Public License], and we can integrate that code and that experience within Ubuntu, and make it available to projects like GNOME and KDE directly.”
All these efforts will be coordinated by an umbrella organization within Canonical known as Ayatana, a name that Shuttleworth defines as “the Zen term for your sphere of consciousness.”
However, Ayatana is still in the progress of being created. Meanwhile, the first efforts in usability have focused on notification messages, and are due to make their first appearance in the upcoming 9.4 release of Ubuntu.
This choice has received some criticism on GNOME and Ubuntu mailing lists — not least from people who wonder where the Canonical Design team suddenly appeared from — but Shuttleworth defends the decision as a necessary testing of the waters.
“There is the thing that, if you have a grand vision, where do you start?” he asks. “And notifications were very carefully chosen as a starting point. We wanted something which affects multiple applications, so we wanted something systemic. We also wanted something on which good progress had been made, and where there was some consensus on a need for a common approach.”
Yet another element in the decision to start with notifications was that “we also wanted to work on something that would be slightly controversial. We wanted to make some tough choices, like removing actions from notifications, and we knew that that would trigger discussion and debate about the design process. You know, ‘How can we have a design process that takes something away?’ That’s something that’s very difficult in the open source community.
“We are used to having every version giving us stuff — more options, more features. So it was important to pick something where part of the design process would be a whittling away. So, for all of those reasons, notifications were chosen. [They are] small enough that we believed that we could do it in a single cycle, but meaningful enough and visible enough that it would draw attention to the work that we’re doing.”
What Shuttleworth and his supporters will tackle next remains to be determined, although Shuttleworth suggests that it will be something that involves “existing design and user experience processes” — perhaps becoming involved in the development of the GNOME Shell, a revisioning of the desktop that, according to current thinking, will be a major part of GNOME 3.0.
However, although every milestone may not have been declared, Shuttleworth is firm on the end goal.
“In eighteen months,” he says, “I would like us to have proven that the combination of a strong design-led process and open source can produce something really stunning and remarkable. I don’t expect it to be complete by then, but, by Ubuntu 10.10, I expect each of the major threads that we now have in play, either publicly or privately, to be integrated into Ubuntu, and to start to be setting the pace on those elements for desktop user experience.”
In other words, Shuttleworth’s goal is to have Ubuntu leading desktop innovation by October 2010. “I don’t think it’s interesting to do this unless we can set the pace,” he says, speaking with a quiet determination.
Ubuntu vs. Apple
Shuttleworth explains that he chose Apple as the competitor to beat because “they bring that philosophy [of usability] to a complete operating environment and the set of products to go with them. I think we need that holistic view if we are to achieve the same thing in the free software environment.”
However, Shuttleworth now adds that other models of individual elegance in usability exist. “Google Maps, for example, is a wonderful example of how Web 2.0 has driven innovation in user experience and made it tangible to open source developers. It’s just a wonderful piece of web work.”
Mention of Google Maps also raises the point that, today, desktop design must include web connectivity. “We must blur the line between the web and the desktop,” Shuttleworth declares, “so both elements are connected.”
But, always, Shuttleworth keeps returning to the problem of how to implement major changes without provoking hostile responses.
“The most important thing that we want to figure out is how to have participation without conflict. It is very clear that, in order to challenge Apple, we’re going to have to make a lot of changes. Nobody would make the case that the free software environment, whether on Ubuntu or any other distribution, is a world-beating experience from a design and user perspective. It’s world-beating for other reasons, right? But it certainly doesn’t win from a design and user perspective.
“If we’re going to put ourselves at the forefront, we’re going to have change a lot. That change is going to be controversial and difficult, and it will not serve our purposes at all if that becomes an excuse for vicious argument. The folks with passion need to get invested in it, either as part of a process like the GNOME 3 discussion, or as part of the Ayatana effort that Canonical is leading, or just by diving into their favorite application and being passionate about user experience.
“We will disagree with each other occasionally. But if we allow those disagreements to become destructive, then we ensure that we will fail. So, to me that’s the greatest challenge: how we are going to marry the community processes that often lead to visceral disagreement with the need to design something that is consistent and holistically designed.”