Predictions about when the year of the Linux desktop might finally arrive are a long-standing joke. They are so widespread that even Jim Zemlin, executive director of the Linux Foundation, laughed about them in his keynote at this year's LinuxCon.
However, in the past year, such predictions have been increasingly replaced by more basic questions: Will the Linux desktop -- whatever its actual market share -- survive at all? And, if not, what are the operating system's survival prospects in a post-desktop world?
By "desktop," of course, I mean the traditional workstation or personal computer -- nobody's predicting a mass retreat from graphical interfaces back to the command line. Instead, the questions are a variant of those circulating throughout IT, as phones, tablets, and other mobile devices become the most common computing devices.
But whether in the general or the Linux-specific form, the question has become a common meme among computer journalists. Searches on "is the desktop dead" and "is the desktop dying" return 440,000 and 160,000 results respectively, with everyone from Jason Perlow and Stephen J. Vaughan-Nichols to Mark Shuttleworth answering from the Linux side.
With all this discussion, the questions sound urgent and revolutionary, especially when Shuttleworth quotes Paul Maritz of VMware as predicting that, three years from now, desktop computers will account for only 20% of Internet traffic.
However, a little perspective is in order. If you've been around IT for a while, then you can't help noticing that journalistic headlines and the statements of would-be visionaries always sound more dramatic than they turn out to be in practice. Often, they fail to materialize, or come true in only severely limited forms.
You might notice, too, that one technology almost never replaces another completely. Contrary to countless predictions, radio and movies were not replaced by TV, nor TV by DVD videos. Go back five years, and you find similar predictions being made about the laptop replacing work stations.
That didn't happen, and the workstation isn't likely to disappear completely in an avalanche of phones and tablets, either. The most that will probably happen is that its marketshare diminishes, perhaps maintaining its numbers while becoming a smaller percentage of total computer sales.
What we are currently seeing is a move away from all-purpose computers to specialized ones. The desktop has always been ridiculously over-powered for the needs of casual users. Now, with phone and tablets, those whose computing consists largely of texting and web browsing have tools better scaled to their needs. Because these tools have no need for large screens or fast processors, they have the advantage of being more portable than workstations.
They are also cheaper -- although often not that much cheaper -- than workstations. Where in the past, one person might use a single workstation for all their computing, that same person today may also have a laptop, a phone, a tablet, and several other devices as well, and switch between them as needed or desired. This reality has already shifted the primary case study for interface design from the desktop to the mobile device, and will probably drive other unforeseen changes as well in years to come.
That is the situation that faces the Linux desktop: not a threat of immediate extinction so much as one of shifting priorities and new competitors. And, in facing that reality, Linux has both advantages and challenges that set it apart from other longstanding operating systems.
Adapting to a World of Mobile Computing
To start with, while mobile devices will become more important in the future, within the remnants of the desktop market, Linux could have a strong chance of increasing its market share. Although desktop Linux users have become more common in the past decade, Linux almost certainly remains an operating system for high-end users -- precisely the sort for whom a workstation's advantages are likely to matter.
Moreover, the package management systems of distributions are far closer to the app stores of mobile devices than anything that Windows or OS X can boast. In a world dominated by mobile devices, Linux can already deliver the service that users expect on the desktop.
For such reasons, the year of the Linux desktop could arrive precisely when the computer desktop no longer matters much.
However, whether Linux can also gain a share of the new mobile market is another issue altogether.
On the one hand, Linux is in a much stronger position from which to enter the mobile market than it was to enter the desktop market fifteen years ago. Significantly, they are no longer starting from scratch. When desktops like KDE and GNOME began development, their first priority was to equal the functionality and aesthetics of existing proprietary interfaces. Nearing that goal took over a decade, and, meanwhile, many incorrectly believed that Linux was primarily run from the command line.
By contrast, today Linux interface design is advanced enough that, if all else were equal, it would be a strong contender in the mobile market. As the speed with which the community has produced alternatives like Ubuntu's Unity or KDE's Plasma Active demonstrates, Linux now has the expertise to develop interfaces that rival Windows or ios 5.
Just as importantly, in distributions like Ubuntu and Jolicloud, it's also embraced cloud computing more than its proprietary rivals have. In general, too, the community's collaborative development model and each project's ability to develop multiple approaches, Linux is vastly better suited to the faster pace of mobile development than any of its competitors are.
On the other hand, the problem is that all else is not equal. Apple's advantages of scale and of being first to market in mobile devices, and the permissive free licenses that allow Android to be competitive means that Linux is, once again at a disadvantage as it tries to enter a market. The fact that in Android, it will be facing its own proprietary bastard, only adds an unpleasant taste of irony.
In addition, Linux also faces the future with other handicaps. For instance, what are the chances of free licenses like the GNU General Public License (GPL) being enforced in app stores with tens of thousands of daily downloads? An unenforceable license could bring the whole idea of free software into disrepute.
True, the GNU Affero GPL is supposed to ensure the software freedom of users of cloud computing. Yet, with the Free Software Foundation expressing deep suspicions about of the whole idea of cloud computing, the Affero GPL has been promoted and used far too little to be as influential in mobile computing as the GPL has been for desktops.
Another self-inflicted problem is that, in the past couple of years, most free software designers have taken to designing all interfaces, including ones for workstations, as though they were intended for the screens of mobile devices. You only have to look at GNOME 3 or Unity to see how widespread this design decision has become.
From some perspectives, the decision makes sense. If mobile devices are going to be the most common computers, then why not give all computers users an interface with which they are familiar? The decision also means that separate code bases no longer need to be maintained for different sized screens.
The problem is, an interface that works with the small screen on a phone becomes torturous on a twenty-inch wide screen monitor.
In addition, while users might endure the restrictions imposed by a mobile device's screen, they don't necessarily enjoy them. Far from appreciating a common interface on all their computers, users might prefer an interface that takes advantage of each device's separate capabilities. Yet, so far, only KDE in Plasma Active has considered this more varied approach.
Admittedly, this approach is easier to code and maintain in KDE than in GNOME or Unity, thanks to the modular structure of the KDE 4 series. But users care only about what they see, and not the back-end. Without such a flexible design, the Linux desktop may gamble heavily on the mobile market, only to fail very messily and thoroughly, increasing existing users' discontent while doing little to appeal to newcomers.
Triumph, Failure, or Death
The shift away from desktop to mobile devices could be either the making or the breaking of Linux. Although it promises a fresh start, it also presents new external and internal challenges, all the more so because most of the community has chosen the one-size-fits-all principle of interface design.
Yet, while a failure of imagination might mean result in Linux's failure to gain market share, an even worse possibility exists. Instead of making Linux more or less popular, the efforts to anticipate the mobile market could result in simply a replay of the desktop market, with Linux remaining a niche player.
If that happens, then Linux will have failed in two major markets to become a major player. And, with two major failures to its discredit, then rumors of its death could become the plain truth rather than an exaggeration.