For years, the classical desktop has been the main interface for interacting with computers. Consisting of a menu, a panel, and an area to display widgets and open windows, its main virtue was originally its easy access to applications and files. It remains popular today, featured in at least five of the seven major Linux desktop environments. Increasingly, though, it is becoming inefficient — a trend that is not helped helped by experimental designs that decrease access to resources rather than increasing it.
When the classical Linux desktop emerged years ago, it was a marked improvement over the command line for the casual user. Icons on the desktop and menu items ensured that executables were always one or two clicks away, and that users spent more time on productivity than in interacting with the desktop.
However, over the years, this basic efficiency has been slowly drained away. The steady increase in hard drive size has made finding items in the menu increasingly laborious.
Shoring Up the Old Design
Realizing that all-inclusive were becoming impractical — never mind the notorious Debian menus — distributions responded by installing with only major applications in the menu, hiding other programs from those unaware of them. Sub-menus helped, but required more clicks to launch an application. Favorites menus and dashboards also provided some relief, although, with more than a dozen items, both start adding to the clutter that they are trying to overcome.
Similarly, only so many icons can be placed on a desktop. The rise of wide screen monitors added space, and so did adding folders, drawers, spinner-racks, and multiple monitors, but one monitor full of icons is still difficult to search, especially if the icon text is abbreviated — let alone two.
Recently, more extreme experiments have been tried. The classical menu whose levels swarmed across the desktop was replaced either by menus confined to a single window, or else to separate screens and overlays like those used in mobile devices. However, both these solutions mean that icons no longer compensate for the crowding of menus, and increase the difficulty of finding the application you want to launch.
More innovative solutions were well-intentioned, but missed the point. GNOME developers, for instance, perceived the increased clutter everywhere on the interface, but, instead of making the clutter useful, removed it altogether, placing more pressure on the already overloaded menu.
Similarly, Ubuntu’s Unity promoted the Favorites menu to the main menu in the form of its launcher, and buried the former menu several clicks below the top of the interface.
However, such innovations have been unable to gain widespread acceptance, despite steady improvements over the last five years. Instead, users have overwhelmingly chosen the classical desktop in the form of Cinnamon, LXDE, Mate, Xfce, and KDE, which has also experimented, but had the sense to offer a classical desktop as a default.
New Approaches and New Metaphors
The classical desktop might be said to be a solution to a command line problem: finding the executable. However, finding the executable among other files is not the problem today. Instead, the problem has become finding the exact executable when so many are on the hard drive.
One solution, which Unity encourages, is to focus on personal documents, displaying their icons, and clicking them to open the applications they run in. This approach has the advantage of launching applications with fewer clicks, but the disadvantage that, in the case of games and utilities, users still need to launch the application.
Besides, today users do not just want to have an application or file on hand. Often, they also want multiple applications, or web addresses as well, and the tools they want on hand differ with the task they are doing.
So far, though, the only desktop that has addressed these changing needs is KDE Plasma. Although it offers a classical desktop, Plasma also allows users to load different icon sets. Instead of trying to fill the desktop with general purpose icons, or else all the icons you could conceivably need, you can load smaller icons sets as needed. If you choose, you can filter out the display of the icon sets that you are not currently using, and place applications in different tabs of the same window.
Another solution in Plasma is to configure virtual workspaces and Activities separately, placing each collection of applications, files, and URLS on a different desktop based on tasks.
What both these Plasma solutions do is to restore the original conditions for which the classical desktop is designed. Instead of dealing with dozens of icons, users only deal with seven or eight. Just as importantly, all those icons can be a single click away.
KDE Plasma does require a huge amount of customization to be useful. Many users will have few objections to that, especially as they realize how quickly they can get to work. However, what it does not provide is a solution that works immediately after installation, and lures users away from application-based thinking.
What such a solution would look like is anybody’s guess. Probably, it would require a new metaphor, or a new answer to a new problem. Its main difficulty would be in coaxing users to think about computing in a different way.
Meanwhile, the classical Linux desktop is becoming full of shunts and bypasses, and operating at a fraction of the efficiency of its glory days. However, its performance has degraded so slowly that the change has gone unnoticed. Users cling to the classical desktop, not because it is the ideal solution, but because it is so familiar that for many any alternative is inconceivable.