Off the Beat: Bruce Byfield's Blog

Journalism is the first draft of history. That is why I was glad to see Christine Hall's article on how the reactions to GNOME 3 created more choices on the free desktop; too often, the community forgets itw own history. However, I would like to expand on her narrative by offering my own variation.

To start with, I would suggest that the current diversity on the desktop was not due entirely to reactions to GNOME 3. When GNOME 3 came out in 2011, it was the third effort in as many years to provide an innovative desktop.

By that point, the free desktop had caught up with proprietary ones after years of struggle. The desktop in general was struggling to deal with vastly more data than when it first became widespread in the mid-Nineties, and, on the free desktop, user design seemed the next frontier. Om 2008,When Mark Shuttleworth of Ubuntu challenged free software to produce a desktop that rivalled Apple's, the process was already well along.

The first attempt at the next generation desktop was KDE 4.0. However, through the under-estimation of end users' wish to be involved in development, and distributions'release of what was intended as a developer version, KDE 4, with its Activities and hot spots, received a hostile reaction, mitigated only as the next few releases added the features that users had been missing.

As that drama was unfolding, Ubuntu was releasing its Unity desktop. The process was marred by Shuttleworth vetoing community suggestions and a mediocre result, causing a second user revolt.
By the time GNOME 3 was released, whatever taste users had for innovation was mostly gone. The situation was not improved, either, by the reluctance of GNOME developers to compromise their vision of a graphical interface by allowing more customization.

By the time GNOME developers were willing to compromise by adding extensions, the damage had been done, and GNOME had gone from one of the top two desktops to one of half a dozen. Users were generally satisfied with GNOME-based apps, which remain the most commonly used on the free desktops, but about two-thirds had abandoned the GNOME shell in the search for something closer to the experience of GNOME 2, whose simple but unglamorous interface was now being remembered through nostalgia-tinged hindsight as the ideal desktop.

Suddenly, Xfce, which had seemed doomed to be the third most popular desktop forever, found a new popularity, registering in second place on some user polls. Linux Mint forked GNOME 2 in MATE, while developing Cinnamon. LXDE and Enlightenment revived. More recently, new desktops like budgie, budgie-remix,  and Elementary OSś Pantheon emerged.

GNOME 3 may have been one innovation too many, the one that released the full fury of users who were feeling neglect. However,from where I sit, it was part of a process that was already happening before the first plans for it had even been discussed.

The price that we pay
Like Hall, I agree that diversity is worth having -- even if it does make discussing free desktops more difficult than in the days when only GNOME and KDE needed to be mentioned. However, the diversity came at a price.

Whatever can be said about the user revolts, their era was a time when developers were rethinking the desktop. At least in theory, the attempts at innovation could have led to the development of new metaphors for graphical interfaces, more suitable to modern computers than the desktop that was imagined two or three decades ago.

Instead, desktop developers became nervous. Even KDE, which I consider the most innovative of the three revolt-spurring desktops, retreated from new features and focused on incremental changes, mostly to do with customization. By doing so, KDE managed to retain much of its market share, but in the process it became safer and much less interesting. Like other desktop environments, the lesson KDE took from the experience was to avoid major changes, and to tinker instead.

You can hardly blame them. Many users were talking as though GNOME 2 was some archetypally perfect desktop, and many of the alternatives being explored were minor variations on GNOME 2 (MATE, Cinnamon, Xfce, LXDE). In fact, considering the number of desktop environments using GNOME applications, the diversity is more apparent than real. Rather than seven or eight desktops, each with its own approaches to handling information, what the free desktop actually has is largely variations on the classical desktop.

Everything considered, the message is clear: make micro-improvements, and do nothing to surprise users too much. If a desktop does have innovations that have somehow managed to surive, as a desktop developer you might want to keep them, but don't draw attention to them.

Meanwhile, major problems remain unaddressed. For example, the traditional cascading menu is an inefficient way to start applications when the average system has dozens -- despite the all-inclusive Debian menu.  Yet all of the alternatives require as many clicks on a workstation as they do on a phone. At the same time, in the name of reducing clutter on the desktop and panels, several desktop environments allow only the most basic of applets.

From this perspective, while I appreciate diversity in the abstract, I am less impressed by the free desktop's current diversity than I might be. When it comes at the cost of less innovation and less usefulness to today's system, I start believing that diversity can come at too high a cost.