Are menus starting to disappear from the Linux desktop? A survey of the alternatives suggests that, at the very least, menus seem to be evolving out of recognition in response to modern trends, particularly the effort to make workstation and laptop desktops more like mobile interfaces. Ask usability experts, and the unexamined assumption is that the classic menu needs improvement -- although whether users feel that way seems less clear.
Ten years ago, desktop menus were straightforward. They listed most of the desktop applications, with sub-menus spilling across the desktop to help organize them. The most extreme case is the famous -- or infamous -- Debian menu, which descends four or five sub-levels, but contains every application installed on the system, if only you had the patience to keep searching. The Debian menu remains popular with some users today, including me.
So, for that matter, does the classical menu in general. Users of Xfce, LXDE, KDE 3 and other lesser-used free desktops continue to use the classical menu with few or no complaints. However, increasingly, on both free and proprietary desktops, the movement is away from the classical menu, although none of the existing alternatives is likely to satisfy anyone.
The Growing Problem of the Menu
In some ways, the change was inevitable. Where hard drives once held dozens of applications, now they hold hundreds. Even on a workstation monitor, sub-menus can easily fill most of the screen. Alternatively, if you try to restrict the menu to two or three levels, which is the general trend, the height of the open menu frequently exceeds the height of the screen, even on workstations.
Nor is the situation helped by the fact that desktop designers are now putting more into the menu than ten years ago. Now, in addition to applications, menus are likely to include lists of favorite applications, locations, and files. Whether these extras serve any purpose varies with the user, but, since they are included by default, they add to the crowding on the menu for everyone. So far as I can tell, this crowding seems to have been born partly out of a wish to get away from icons or launchers on the desktop, even though both are popular with large numbers of users.
Along with these increased demands comes the assumption among usability experts that users -- especially new ones -- should not be overwhelmed by all the available choices, as they were in the past. Instead, only a selection of applications likely to interest new users would be included in the menu. The result was more elegant, less cluttered menus, but at the cost of concealing some applications so successfully that some users would probably never find them.
All these problems are magnified several times over by the smaller screens of netbooks and other mobile devices. Menus do, of course, exist on mobile devices, but often require complete changes of screen, as well as alternatives like scroll wheels to speed navigation. Instead, the mobile trend seems to be towards different forms of launchers, which can be described as a compromise between menus and desktop icons. Now, in new, mobile-inspired interfaces like Ubuntu's Unity or the upcoming GNOME 3.0, launchers rather than menus are being offered as the main navigation tool.
The Search for Alternatives
Some of the first attempts to improve the menu were merely organizational -- efforts to standardize or reduce the number of top level menus. GNOME in particular worked hard on this effort, with KDE not showing a similar concern until the fourth series of releases began. In addition, GNOME moved away from the single menu to a series of three: Applications, Places, Menu.
The trouble was, these organizational standards have never been consistent across distributions. Since menus are relatively easy to edit, distributions would arrange applications or functions as they saw fit. Although at first users switching between distros might think they know where to find their favorite applications, in practice, they usually have to do some searching.
Mercifully, no free desktop (so far as I am aware), ever attempted to emulate the so-called personalized menus in Windows that changed with your use and only showed the most frequently used menu items. The drawback to these personalized menus was the assumption in their design that most users only clicked regularly on a few items; if you wanted anything unusual, they required an extra click and a few seconds before you could seen a complete list of choices.
Instead, the free desktop has explored other alternatives in recent years. In KDE 4.0, KDE introduced Kickoff, which seems inspired by innovations introduced in Windows Vista without being entirely an imitation. Kickoff's main difference from the classical menu is to confine the display to a single window showing one level at a time. This structure makes Kickoff more suitable for smaller screens, but requires users to click back to a higher level before they can see, much less switch to another sub-menu.
To satisfy users, KDE continues to include the classical menu as an alternative to the default Kickoff. Moreover, KDE also includes a third menu choice called Lancelot among its desktop and panel widgets. Like Kickoff, Lancelot is confined to a single window on the desktop, but, unlike Kickoff, Lancelot also shows more than one menu level at once. This innovation makes navigation easier, but has the drawback of being cramped if more than three levels are visible at once; resizing the Lancelot window is possible, but requires some fiddling, even if the desktop widgets are unlocked.
Enter the Mobiles
More recently, designers have started exploring mobile-inspired desktops that abandon or minimize the classic menu altogether.
The first of these next-generation desktops to be released was KDE's Plasma Netbook. It moves the applications menu and the taskbar right on to the desktop, a solution that is legible and highly usable even on a small screen like those found on mobile devices. However, Plasma Netbook is no better at preserving screen space for applications than the classical menu.
An even more radical departure is unfolding in the upcoming GNOME 3.0. In GNOME 3.0, the menu is labeled Activities, and simply changes the desktop view to show a launcher and the available virtual workspaces. The GNOME 3.0 launcher is moved to the left side of the screen in acknowledgement of the fact that wide screen displays have more horizontal space than vertical, but otherwise does little to overcome the problems with the classical menu. Places and Devices and Recent Items menus still occupy valuable space on the launcher, and, to find an item to launch, you need to scroll through an alphabetical list of icons that -- at least in the version offered in the latest Fedora release -- has no high-level organization whatsoever.
In these limitations, the launcher is no worse than the interface of a typical mobile device, but someone accustomed to a workstation may find it a step back from the classical menu. However, possibly, it will become more usable by GNOME 3.0's general release.
Another radical departure with a launcher on the left side of the screen appears in Unity, which is scheduled to become Ubuntu's new default desktop in the next release. However, although Unity's launcher has large, easily readable icons, it suffers from too little organization. In the end, like Plasma Netbook, it falls back on large buttons on the desktop -- a solution that makes for legibility, but is no better than the classical menu when it comes to taking up space.
Different, Not Better
Of all the alternatives, Lancelot comes closest to overcoming the problems with the classical menu while not creating new ones. However, it is used by only a minority, even among KDE users. Given the mobile-inspired designs of Plasma Netbook, GNOME 3.0, and Unity, the users of the major desktops may soon find themselves next to menuless.
Nothing would be wrong with that it the change had any advantages, but, for the most part, these innovations are not better so much as simply different. The classical menu may be cumbersome on modern computers, but it has the advantage of being familiar, and of working with a minimum of fuss -- and of mouse-clicks.
All the same, I suspect that the classical menu is about to become rare on the major free desktops. Having created some of the problems by overloading the menu, usability experts seem less to be solving the problems than exchanging them for new ones.
The worse thing is, I am not convinced that users want the change. Unless I missed something, there is no consensus that the classical menu is hopeless. Nor is there any indication that users want or expect their laptop and workstation interfaces to resemble those on their phones and music players. If anything, I suspect that the average users are sophisticated enough to use the interfaces that are most suitable to the displays.
I have no sentimental attachment to the classical menu. I have little doubt that one day it will be replaced by something better. However, more and more, as I try the efforts to replace it, I am growing convinced that the usability experts are trying to fix something that is not so badly broken that it urgently needs repairs -- and that part of the problems that do exist are the result of their own tinkering.