Over the past year, Ubuntu has become one of the centers for usability design on the Linux desktop. You might criticize this effort because it takes place in the distribution, rather than as contributions to the GNOME desktop, but at least it is happening. Moreover, this effort is being discussed far beyond the outer reaches of the Ubuntu community.
Part of the reason for this discussion is because Ubuntu's popularity automatically makes it influential.
Yet an even more important reason for the interest is that Ubuntu founder Mark Shuttleworth has made usability his personal obsession. Not only has he withdrawn from managing the business affairs of Canonical, Ubuntu's commercial arm, in order to devote his full attention to it, but -- what really matters -- he blogs about usability decisions as they are announced.
In these blog entries, Shuttleworth gives observers a rare chance to see the rationales behind usability decisions, and to measure the rationales against the results.
A case in point is the Unity desktop, announced two weeks ago as the basis of the next Ubuntu Netbook Edition. In the future, Unity will also be the basis of Ubuntu Light, which the release describes as "intended for the dual-boot 'instant-web' market" -- that is, web-centered machines that have both Windows and Ubuntu installed.
Shuttleworth's blog mentions four main design assumptions or goals: a "finger-friendly" layout (one suitable for touch screens on mobile devices), a faster boot, an effort to use horizontal more than vertical space because of the present prevalence of widescreens, and a limited number of local applications.
The Ubuntu Unity desktop
By downloading Unity to an existing copy of Lucid Lynx, the latest Ubuntu version, and comparing it to Shuttleworth's comments on Unity, anyone who is interested can compare the reality to the rationales that created it, and evaluate the results on their own terms, as well as measure them against reality.
The opportunity is very nearly unique -- and completely irresistible. By my evaluation, the current version of Unity misses three of Shuttleworth's goals, and meets only one of them.
The first missed goal is that of making Unity ready for touch screens. The icons on the panel on the left seem large enough for selection with a finger (or, more likely, a thumb), but are possibly too close together to avoid mistakes, although Shuttleworth does promise that "well expand that left-hand launcher panel so that it is touch-friendly."
The top panel is even worse, having been ported unchanged from standard Ubuntu into Unity. Clearly, more work remains to be done with this goal.
The second miss is a fast boot time, a goal that is apparently a pre-occupation in Ubuntu development circles. During development of Lucid, one of the goals was to reduce the boot time on test workstations to ten seconds.
Personally, though, I have yet to hear of any installation of Lucid that came within six seconds of that goal. Although that is still impressive, if that gap between the goal and the reality is typical, it suggests why Shuttleworth begins his blog on Unity by describing the new desktop as "the fruit of that R&D."
In other words, the goal has been transferred to Unity as something still unobtained, although the news release raises the stakes by declaring the goal as "under ten seconds."
Shuttleworth may also imply that Ubuntu cut some corners in the effort to boot quickly, since he adds that one of the lessons learned from observing users is that "Its not about how fast you appear to boot. Its about how fast you actually deliver a working web browser and Internet connection. Its about how fast you have a running system that is responsive to the needs of the user."
At any rate, the current release of Unity has yet to reach that goal on my test machines. In fact, if I set the installation to log in automatically to Unity, I can use the desktop in more or less the same time as when booting GNOME.
The third miss is the change in the orientation of the screen. As Shuttleworth observes in his blog, "Notebooks in general are moving to wide screen formats. So vertical space is more precious than horizontal space." In keeping with this dictum, Unity eliminates the bottom panel that Ubuntu has always added to GNOME in favor of one on the left -- a simple and obvious move.
Yet, by the same logic, why keep the top panel? While Shuttleworth does talk in passing about a dashboard "which presents files and applications as an overlay," he does not say whether it is intended as a replacement for the top panel.
My guess is that it is not, since elsewhere in his blog he talks about making the top panel "smarter" by adding a global menu and having window titles appear in it. In this respect, as with "finger friendliness" and boot speed, the present version of Unity shows only tentative movements away from earlier versions of Ubuntu.
According to Shuttleworth's blog, work on Unity began with the assumption that the target installations computers have "no need of heavy local file management." Instead, users would be largely interested in online services, which they could use to keep their Ubuntu and Windows installations on the machine in sync with each other.
Nor would the target machines be "environments where people would naturally expect to use a wide range of applications." The exceptions would be "media playback, messaging, games, and the ability to connect to local devices like printers and cameras and pluggable media."
This logic is already visible in Unity. The left panel is reduced to eight icons. From top to bottom, they are: Ubuntu Software Center, Help, Firefox, Applications, Rhythmbox, Empathy, Evolution, and File Manager.
In this version, Applications is not a menu, but a file manager view of /usr/share/applications, a piece of inelegance that emphasizes better than anything else could the assumptions that Shuttleworth and his team make about how Unity will be used.
The layout could be criticized for its lack of priorities -- for example, placing the Ubuntu Help Center on the top seems to contradict the contention that users will not want many applications. Yet the order of application launchers is not stressed by Shuttleworth as a goal, so Shuttleworth and his team can still be said "to preserve that sense of having a few favorite applications that are instantly accessible."
The reality check
Shuttleworth's blog shows that new features are already being considered for Unity. In particular, although Shuttleworth states that, typically, "instant-on environments are locked down," plans do seem afoot to put customization options on the desktop, rather than burying them several clicks beneath the Applications launcher.
Still, for now, comparing Unity to Shuttleworth's comments, what becomes obvious is that we are not seeing anything like a finished delivery of development concepts. Instead, we are seeing the playground on which those principles are being actively worked out.
Given this state, the next question is: Are Shuttleworth's assumptions valid ones?
Although Shuttleworth talks of studying "a couple of hundred different desktop configurations from the current Ubuntu," possibly not.
When Ubuntu already has one of the fastest boots of any distribution, are users really that concerned about reducing the boot time by another few seconds? In the past, the main concerns about boot time have been in comparison to proprietary operating systems, and the latest Ubuntu's boot already compares favorably to Windows' or OS X's. Perhaps Ubuntu has already done enough work on this goal for now.
Other questions also arise. Do most users further divide the relatively small hard drives of netbooks by installing multiple desktops? Are web-centered computers really going to major players, or is that a hope for the increased use of Ubuntu One's online storage?
So far, web-centered distributions like Chromium and Jolicloud seem more novelties than solutions to which significant numbers of users are migrating.
Most important of all, are people really interested in only light, web-based computing on netbooks? I have watched coders, writers, and even graphic designers work on netbooks in airports, so I have my doubts. Shuttleworth's team (and many others) seem to be designing for the first generation of netbooks when the second generation, with more RAM and larger hard drives are already here.
Unity still has a long development ahead. It may change in all sorts of ways, including in its assumptions. I can only hope so.
Because, you see, as fascinating as watching how assumptions play out, there is a serious problem with them. Even if you follow your own logic faithfully in your design, you still may not get to anywhere you want to go -- not if your assumptions are flawed from the start