Off the Beat: Bruce Byfield's Blog

I was surprised by the passion generated by my blog entry last week about Ubuntu's decision to replace GNOME with its own Unity desktop. Apparently, contrary to the pundits and usability experts, users have strong feelings about their interfaces of choice. But, when I stop to think, I should have expected that. For many free software users, the choice of desktop is still a deeply personal matter.

Journalists like me often leap to write about what's new. That tendency can be seen as a service, but it also means that the importance of trends is often exaggerated in the rush to report first.

For instance, in the last five years, network appliances and cloud computing have been hailed by many as the next revolution in computing. According to the standard story, in a few years, computers will only be launchpads for online services. Either the browser will replace the desktop, or else the services we now access through the browser will be dispersed so that the entire orientation of the desktop will be to online services. And, hearing this new gospel, designers have been quick to introduce interfaces to meet this new phase, introducing new default applets and widgets and new desktops like Chrome OS.

However, the expectations have existed long enough that now we can safely say that they aren't happening -- at least, not in the way everyone expected.

Oh, people are using online services, all right. But they aren't using them as much as predicted, and most of us aren't using them instead of our local applications. Here and there you might find people who have jettisoned their local apps for online services, but that's not what the majority are doing. Instead, we are using online services and desktop applications together, using whichever seems most convenient at the time.

This state of affairs is explained partly by concerns about the availability and security of online applications. Yet, even if these concerns were fully addressed, I doubt that the situation would be much different, because paradigm shifts in technology are rarely total. Photography did not replace painting, although it may have encouraged the rise of surrealism and abstraction; the two co-exist. The same goes for TV and radio, or movies and DVD players. So why did anyone seriously expect online services to replace the desktop?

The only way that might have happened is if online services significantly improved users' ability to do what they wanted to do -- which is what happened when the internal combustion engine replaced the horse and cart.

Add the failure of most online apps to provide more than the simplest of functions compared to local ones, and we are left with a state of affairs that is likely to continue for some time. Far from dominating computing, online services remain a niche market, relied on by a few, used sparingly by some, and not at all by many.

Like Java, Agile Computing, and most of the last three or four dozen revolutions in computing, the move away from the desktop proves less all-encompassing than promised. For better or worse, the desktop remains, altered here and there by the new idea, but still recognizably the same.

A lifestyle, not a tool

Some users, of course, claim they are indifferent towards the desktop. In fact, in some moods, the majority of us probably feel that way. When you are thinking of contacting someone to arrange an appointment, or to write something down before you forget, what matters is the task at hand, not what wallpaper is on your desktop or what utilities are available. Free software users may be especially prone to this attitude, because, if GNOME is acting up, they can start Xfce instead. For that matter, if FIrefox won't work, they can start Chrome or Konqueror instead, so that even applications matter less than the task at hand.

Yet, for most users in most moods, the desktop still matters. Unless, like me, you are constantly experimenting with desktops (and often even then), you will find that some fit your work habits or preferences better than others. One reason why GNOME and KDE have come to dominate the free desktop is that they both try to accommodate as many different use cases as possible. But this effort brings its own problems, so that those who place a high value on efficiency and speed will avoid GNOME and KDE in favor of LXDE or Enlightenment.

Moreover, your desktop is often a statement of who you are. That is true even in the proprietary world of OS X and Microsoft, but if anything it is even truer on the free desktop. Your preference for BASH may be genuine, but it also proclaims you an expert user, a hacker or an old timer. Using a tiled window manager like Ratpoison also suggests expertise, as well as a tendency to minimalism. I have even heard it seriously suggested that GNOME is for business and KDE for artistic types. So far as that goes, using any free desktop in this proprietary world is a major statement of what matters to you.

From this perspective, it is not surprising that flame wars erupt over desktops and can flare up in an instant. A desktop and its utilities are not just tools: for many of us, they are also ways of life.

A clash of cultures

These attitudes, I suspect, are often overlooked by developers and usability experts. Users are not simply conservative (although often they are); they resist radical change in the desktop because their sense of who they are is bound up with their choice of interface.

In the case of any free desktop alternative, this identification becomes stronger because it is so well-articulated. Regardless of their interface of choice, free desktop users expect to be able to do things their way, not the way the application forces them to. They do not simply expect to customize their wallpaper, but to be able to choose whether to use the desktop or the menu or the panel to launch their applications, or even the exact menu style on the desktop. The availability of these choices does not cause anxiety-option for the experienced free desktop user -- instead, choice is an assertion of their values and the mark of a properly-designed desktop or application.

The problem is that these attitudes can quickly run into opposition from developers and usability experts. Anyone in a position to make decisions needs to watch themselves for the assumption that anything they decide is automatically correct, but in software design, the need is even greater than elsewhere. Usability testing can be wildly unscientific, with unspoken assumptions being surrounded by the illusion of objective fact simply because someone somewhere generated supporting statistics. Unless care is taken, usability expertise can become simply a disguise for top-down decision making.

This danger is especially strong when usability truisms are applied to free software. Much of what we think we know about usability comes from proprietary software, where users get fewer chances to assert their preferences or do things their own way. Apply many of the common assumptions about usability to free software, and you risk user dis-satisfaction, if not outright revolt.

At times, this is what I suspect is happening at Canonical; if it is, then the outspoken resistance to some of the changes being instituted become more understandable. Perhaps what is happening is not a difference of opinion, or a stubborn refusal to accept objective facts about usability so much as a culture clash between usability experts and free software values.

However, whether that is true, the basic fact remains: The desktop still matters, especially in the free software community. You might question whether it is the best metaphor for computer interfaces, but at this late point, changing it can be a risky business.

Comments