Windows 8 and the disappearance of the Start button

20 February 2012 Though Microsoft may yet have some trick up its sleeve, there’s a growing body of evidence from leaked screenshots that Windows 8’s taskbar will omit one mainstay of the Windows user interface: the Start button. To get to the…

20 February 2012

Though Microsoft may yet have some trick up its sleeve, there’s a growing body of evidence from leaked screenshots that Windows 8’s taskbar will omit one mainstay of the Windows user interface: the Start button. To get to the Start menu’s replacement (the “Start screen”) from the desktop, you can press the Windows key on your keyboard, hit the hardware Windows key if your tablet has one, swipe your touch-screen from the side, or, if you have a mouse, move it to the bottom-left corner or right-hand edge. What you won’t be able to do is actually click or tap a permanently-visible Start button on the taskbar.

The rumor may or may not turn out to be true, but if it is, we shouldn’t be surprised. That’s because you’ve entered the Metro zone, where the rules of human interaction have changed—and it’s a change that will be felt not just by tablet users, but by traditional desktop denizens, too. Microsoft will need to tread carefully.

The Metro factor

From the outset, Microsoft designed Metro to be different from the Windows of old. Metro applications, downloaded and purchased through the Windows store, will eschew the Windows, Icons, Menus, and Pointers we’re used to in favor of a new text-heavy, low-chrome, windowless Metro style.

The Metro aesthetic is essentially chromeless. The trappings of the traditional window—borders, title bars, permanently visible scrollbars, toolbars—are gone or at least substantially scaled back. Instead, we get much more emphasis on the use of, for example, juxtaposition and layout to convey information. Metro doesn’t do away with chrome entirely (it still has buttons, etc.), but the chrome is much reduced compared to desktop Windows.

With the new aesthetic comes a new attitude towards user interaction. Traditionally, Windows has only used a small repertoire of learned interactions: click, double click, right click, drag-select, drag-and-drop. Almost everything else is visually cued on-screen. With Metro, a whole bunch of extra learned interactions are necessary, and the number of visual cues is greatly reduced.

Windows 8’s gesture vocabulary is for the most part simple, with swipes and taps being the essential operations. The most important uses for these actions are swipe from the left and right edges of the screen (to task switch and bring up the “charms,” respectively), and from the top or bottom (to bring up toolbars). Applications can be closed by dragging them off the bottom of the screen. Then there are some smaller gestures, like nudging items on the Start screen to select them for editing. But many of these interactions won’t be prompted; they must be learned.

The swipe gestures to switch between applications, bring up the charms, or close tasks—none of these have any on-screen prompting. Nor do elements of the touch vernacular like “pinch zoom” (depending on form factor, at least; pinch zoom is peculiar on a phone that’s mostly held and used one-handed, much more obvious on a tablet that’s held and used two-handed).

It’s possible that Windows 8 will have some kind of a tutorial, as Windows 3.1 once did, but this is unlikely. Such tutorials are not exactly fashionable, and the competition doesn’t have to teach users how to use the platform (in spite of some non-obvious commands, such as “double tap home button” and “long hold”).

Explicit versus implicit

It seems almost laughable today, but Windows 3.1 included a tutorial to explain how to use the mouse. Windows 95 didn’t have the mouse tutorial. By 1995, mice were common enough to take mousing skills for granted, but it did have to teach users all the same. To encourage users to click the Start button, a bouncing message, “Click here to begin,” would slide along the taskbar if the user didn’t do anything after starting the operating system. Even the “Start” name itself was a bid to lure users into clicking; early prototypes had an unlabeled button.

Windows 95’s “Click here to begin.”

The bouncing prompt didn’t survive past Windows 95, and even the Start label was finally dropped in Windows Vista; the Windows logo on the left-hand side of the taskbar was enough of a cue that people knew to click it. Windows 7 has all manner of hidden, unprompted user interface elements. Some are new, such as the jump lists that appear when you right click a taskbar button, and others old and long-standing, such as the alt-tab application switcher.

The trend to strip away explicit visual cues and rely more heavily on a common set of learned interactions is not a new one, but Metro is more extreme in this regard than any prior Windows version (though not substantially more extreme than other tablet platforms). It’s an industry-wide trend.

The upside to this is slicker, lighter user interfaces. On some level, it’s silly to waste screen space on user interface gadgets that prompt us to do things we know how to do anyway. For desktop machines with big screens, perhaps an X in the top right corner of every window is a small price to pay. But on a tablet (or a smartphone) where the constraints are much tighter (and where accidental presses are much more likely), a gestural method of closing applications saves precious pixels.

Learned versus intuitive

For actions that we do day in, day out, these prompts and on-screen cues are more than just a waste of space. They’re straight up pointless. I don’t need an on-screen cue to know where to resize a window, for example, because I know the border is where I resize the window. Windows 7’s fat window borders, showing me where to grab hold of them to resize, are wasting space and telling me nothing I don’t already know. A Mac OS X 10.7-style non-existent border would work just as well.

These basic interactions are all learned. It might seem “unintuitive” that Mac OS X 10.7 lets you grab a window border that “isn’t there,” but in practice it doesn’t matter: we learn the interface and move on.

The learned interface no longer needs to be cluttered with affordances: visual cues that there’s some interactive, functional item on the screen. There is no need for borders to “grab hold of,” or of raised, pseudo-3D buttons to “depress.”

There are many who’d argue that we learn every interface (except, perhaps, the very first one we use), and for all the emphasis put on “intuitive” interfaces, the use of learned interactions is downright expected. An operating system released in 2012 that tried to teach people how to double click or use a mouse would be laughed at.

There’s nothing wrong with learned UI as such. A learned interface no longer has to explain every aspect of its operation to users, and can be a lot more streamlined and efficient as a result. But affordances do provide a great degree of discoverability. Users know which bits of the interface to experiment with because they are familiar with the on-screen prompts that the interface provides.

And not every interface can be learned. Tasks that are rare and infrequent are poor candidates for learning. Perhaps the most extreme demonstration of this is the command line compared to a wizard interface. Wizards are designed to be unlearned interfaces; they should be used for processes that are only performed infrequently, and so have to explain terminology and guide unfamiliar users through the process. The command line provides no guidance or explanation, and it’s only worth learning the esoterica of individual commands if you’re going to use them regularly. But ultimately, once the learning hurdle has been overcome, the command line can be far more productive and flexible than the wizard interface.

So what does this mean for Windows 8?

Moving forward, in different directions

Ditching the Start button in Windows 8 is consistent with these broader trends. The Start button is something that people will use many times a day; it’s an interface that will inevitably be learned. Hiding the button also provides a kind of uniformity with the Metro environment, where the major UI elements are invisible until swiped into view. So while the change will certainly be a shock, it’s broadly in keeping with the direction Microsoft has been heading in for a number of years, and it fits well with the overall Windows 8 concept.

Freeing up an extra slot on the taskbar is also welcome.

The taskbar without a Start button looks much as you’d expect, really. It’s a taskbar, without a Start button.

Extending elements of the Metro aesthetic to the traditional desktop will also help wed the two together. Perhaps Microsoft’s biggest challenge with Windows 8 is convincing users that the Metro world and the traditional desktop can coexist happily side-by-side. The operating system needs to be a coherent whole, not an awkward chimera of two paradigms grafted crudely together. Making things like the styling and the use of the mouse and gestures as consistent as possible will aid that goal.

But one part of Windows 8 makes a big move in the opposite direction; it’s gaining more chrome, not less, and it’s turning away from learned interactions. That’s Explorer, newly equipped with a ribbon.

The new Explorer, with ribbon

Explorer is a desktop application through and through. That’s OK; Explorer does some important things, and its information density can be very high. Making it a true Metro application without compromising its capabilities is probably not trivial. But when contrasted with Metro, we notice something more significant: the ribbon Explorer is totally unlike anything in the Metro world.

Windows 7’s Explorer is not devoid of chrome, but it is reasonably low-key. The closest thing it has to a toolbar, the command bar, is primarily text, as is the address bar. It isn’t a Metro application by any stretch, and it still contains panels and subdivisions and graphical buttons, but it’s relatively lightweight, and it would be easy to envisage a version with the paneling toned down and the different sections given more room to breathe.

The ribbon and Metro also demonstrate different approaches to user interaction. The ribbon was designed to allow the complex Office applications to expose more of their functionality to overwhelmed users. Fundamentally, the ribbon is a user interface designed to facilitate and reward exploration, with learning and memorization taking a back seat. The ribbon was designed for Office, a consequence of Microsoft’s realization that Office’s full power was essentially unlearnable by many casual users. The contrast between this and the implicit, lightweight Metro style is striking.

Getting users on-board

There is, of course, a major problem with learned interfaces: they have to be learned. Handling this sensitively and effectively will be a challenge, one compounded by Windows 8’s wide mix of users. Every group, whether using touch or the traditional keyboard and mouse, will have something to learn in the new interface.

The increasingly substantial smartphone and tablet-owning demographic will all be familiar with the basic touch vocabulary of taps, swipes, pinch zoom, and more, but the specific capabilities associated with screen edges—including the fact that different edges do different things—will have to be learned somehow.

Traditional mouse and keyboard users will also need education. Microsoft might be trusting that Windows users will automatically move the mouse to where the Start button used to be, though this seems a little unreliable as a way of getting people to discover an essential part of the interface. Switching apps on the desktop will be familiar enough, but switching between Metro applications will similarly require education.

For new users with time to explore, and especially those users who buy tablets, there will be plenty to discover and unearth in Windows 8’s interface. Learning the ins and outs of how it works will likely be an enjoyable part of the Windows 8 experience. But for those wanting to get work done, especially those using Windows 8 on traditional machines, Microsoft will have to tread carefully, lest it make the learning experience an unpleasant one.