Will the Desktop Die?
With the huge popularity of mobile phones and tablets, apocalyptic voices announcing the demise of the desktop are louder and louder. We keep hearing about how, in a few years, we’ll all spend our time on our phones and tablets and let our PCs gather dust, or, even worse, find their way to the garbage dump (See, for instance, http://news.cnet.com/8301-1001_3-57530402-92/pc-market-much-worse-than-expected-says-analyst/.)
Do we need to worry? Should all of us who study or design websites and applications for PCs start designing for mobile exclusively instead? Or perhaps, should we look over the fence and borrow from the mobile designer’s arsenal of tools to replace old fashioned PC design?
Without a doubt, many of the tasks that some of us do on a PC will soon migrate to a tablet or mobile device. It’s been said that tablets are consumption devices—and, indeed, most people use them for reading news, email, magazines, or books, and for watching videos. Maybe in the near future we will do these tasks mostly on tablets and cease most of our media consumption activities on the PC.
There are, however, tasks that are better suited to a larger screen. In our tablet and phone testing, many users say that they’d rather complete a purchase on a PC than on either of those devices. While there is some mobile commerce, it’s really hard to do complex tasks on a small screen. It doesn’t mean that people won’t do it if forced, but it’s unlikely that it’s going to be their method of choice.
There are several limitations on mobile and tablets today. One is the poor input capabilities. Yes, people do text and email with soft keyboards, but writing novels and creating complex presentations are tasks that are still better done on the desktop. Some of these input limitations can be overcome with an external keyboard. The new Microsoft Surface tablet, announced to ship at the end of October, will have a cover that doubles as external keyboard, and many other tablets will dock to a keyboard.
A more serious limitation has to do with the small screen. Even on the larger tablet screen, it’s hard to have multiple windows visible at the same time, and there’s only so much content that can fit in one window at one time. For tasks involving comparisons and multiple sources of information that need to be put together, mobile screens or even a 10-inch tablet screen are going to be too small. If you want to create a chart based on data from several other documents, moving back and forth and selecting the necessary information will be more painful on a tablet. If you need to understand complex operating instructions while looking at device pictures, the small screen will also be a limiting factor. There simply is not enough space to put that much information on a small screen, and whatever useful information does not fit in there will need to find a place somewhere else—either on some external memory source (printouts, anyone?) or perhaps on a different screen.
Every once in a while we are forced to do more complex activities on our mobile screens. Say that we are at a museum (and thus away from the desk), and are faced with the choice between waiting in line for 30 minutes for an exhibition or buying tickets on the museum website using our mobile phone. Even if the museum site was not optimized for mobile use, many of us would put up with the suboptimal experience of the full site on a mobile device because the benefit is high enough. But that doesn’t mean that from now on we’ll always choose suboptimal experiences just because we can. Suboptimal experiences incur a significant interaction cost. We’ll choose them when the benefits we get from them surpass that interaction cost.
Mobile and PCs are good for different tasks and they can coexist. There’s no war between the two: They each have a place in our tech ecosystem.
The Chrome to Content Ratio, or Why Designing for Mobile Doesn’t Work on the Desktop
But how about the idea of designing for the desktop as if we were designing for a big mobile device? On mobile we always recommend an economy of features and content. We advocate stripping down to the bare essentials, cutting the extraneous out, and optimizing for quick access to information, short attention spans, and efficiency. Wouldn’t that approach benefit the desktop user as well?
Well, yes and no. Yes, because if there’s a lot of fluff and irrelevant information on your website, you do need to eliminate it and put users’ needs first.
But designing for mobile is more than that. Let’s take an example—that of chrome versus content. (The term “chrome” refers to the visual design elements that give users information about or commands to operate on the screen’s content, as opposed to being part of that content.) We often say that, in mobile design, content must be prioritized as much as possible: New information is crucial, and most of the screen should be devoted to it. Because the mobile screen (phone or tablet) is much smaller than that of average PC screen, busy chrome steals significant space away from the relevant content. Thus, a typical PC design with lots of menus and navigation bars does not work well on smartphones or tablets.
The first iPad designs went overboard and tried to hide the chrome almost completely, forcing users to rely on gestures to interact with the apps. People were often lost, and couldn’t find their way in a sea of gestures. (The first magazine apps on the iPad—notably, Popular Science—are examples of gesture overuse.) Luckily, designers realized their mistake quite soon and backed up, relying on fewer gestures and more buttons. Sometimes the controls needed to be exposed by an extra action (for instance, tapping on the screen), but that was better than simply forcing the user to use gestures exclusively. In other situations, designers used tips and progressive disclosure to teach their users to use the gestures.
How about hiding the chrome on the desktop? Wouldn’t that also benefit the users by allowing them to plunge into the content without being bothered by the ugly interface?
The answer is no. On desktops, hiding the interface is gratuitous, from a cost-benefit perspective.
First, the benefit: The desktop screen is big—even with navigation bars, menus and other widgets, content will not be penalized. The ratio of content to chrome is high enough on a big screen, even when the interface is visible and some screen space is given to buttons. (While some desktop software does have too many GUI widgets, well-designed software does benefit from exposing a fair amount of functionality when the space is available.)
Second, the interaction cost: Whenever the interface is not exposed, users must to figure out how to interact with the app. If the app hides a button, users have to take an extra step to uncover that button. Before pressing it, they have to think (either guess or remember) about where that button may be and how to carry out the action.
Many times, in a consumption environment the content drives the interaction. Users are in a browsing mode and don’t need to act upon the content that much. Yes, they may occasionally need to share an article or find the technology section, but most of the time the interaction is bottom-up: from the device to the user.
However, in a production environment, the directionality of the interaction changes. Users are more driven by their own goals: they need to modify or create the content they have in mind, not just the content that is easily accessed through the interface. Productivity happens through interface controls, and having to perform extra work to uncover them for every single action puts a burden on the user.
An Example: Windows 8
The temptation of chromeless design has become so big that some desktop designs have started to copy it. Proof is Windows 8, the new version of the Windows operating system, scheduled to become available later this month. (The interface used to be called Metro until recently, and its principles were applied before in Windows Phone 7 phones.) Windows 8 is an OS created to look the same across devices: mobile, tablets, or regular PCs. Because of that it had to be designed with mobile in mind, and as we said before, many windows, menus and navigation bars don’t work that well on mobile.
We recently finished a round of usability testing of a preview version of Windows 8 on a regular PC, equipped with full keyboard and mouse. Our study participants were users of older versions of Windows (from Windows XP to Windows 7); one of them was an expert user who was fairly versed in using interface shortcuts.
Windows 8 made it hard for all of them. Granted, it was their first encounter, and maybe after persistent use, things will become easier. However, none of them left the 90-minute session feeling that they had learned how to use it and that it was easy.
The main difficulty was discovering the controls. In Windows 8 the start menu has been replaced with a start page. Various controls in the interface can be accessed by clicking the corners of the screen or by using the right mouse button.
Figure 1 (click to enlarge): Start screen for Windows 8. To access controls, users have to click the corners of the screen, or use the right mouse button.
For instance, to expose the “charms” bar that includes the search function and the settings, users need to click the bottom-right corner of the screen.
Figure 2 (click to enlarge): “Charms” bar for Windows 8, accessed by clicking the lower right-hand-side corner.
These actions are not particularly intuitive: there are no affordances regarding clicking the corners. The left bottom corner was still discovered more quickly than the right bottom corner (in spite of the little icon with a dash shown on the bottom right in Figure 2).
In the Weather application (that comes preinstalled with Windows 8), we asked our study participants to find the weather forecast for San Francisco. They had a hard time figuring out that they had to press the right-hand-side button of the mouse in that app in order to expose the chrome and change the city of interest to San Francisco. In fact, some users never pressed that button and were stuck in an app with which there was no way to interact.
Figure 3b (click to enlarge)
In the Weather app (Figure 3a), to change location you need to expose the chrome by clicking the right-hand-side mouse button (Figure 3b). It would be much more intuitive for users to click a visible object such as the name of the current location (here, Sunnyvale), as indeed our test users attempted to do. Even better would be if that object were made to look clickable.
Beside the lack of chrome, another feature that Windows 8 has borrowed from tablets is the single window display. An app usually takes up the whole space of the screen, which makes sense on a tablet, but can be less productive on the PC. When we asked our users to do tasks such as copy and paste information from several websites into an email or into a file, it took them a long time to get from one app to another. They first had to copy the information from the browser, then went back to the start screen (by clicking the lower left corner) and selected the app, and then went into that app to paste whatever they had entered. For every single action, there were too many steps that needed to be taken.
One of our users was an expert user who had even tried Windows 8 on his PC before. To switch tasks, he ended up relying almost exclusively on the keyboard shortcuts that he was familiar with from previous versions of Windows. But even with this expert-user skill, the process of moving back and forth between apps was more costly than it needed to be, because he did not have the two apps on the same screen side by side. (Yes, in Windows 8 you can have two apps visible on the screen at the same time—one taking approximately a quarter of the screen and the other the rest of it—but none of our users were able to make it work.)
Microsoft probably recognized that this interface will not be optimal for productivity, because they decided to have their old interface stick around in the form of a Desktop app. (And, of course, they did have to ensure compatibility to previous Windows versions.) Once the Desktop app is selected, users are presented with a more traditional view of Windows similar to what they’ve used in previous versions. In that Desktop app they still can run Microsoft Office and have several windows with menus and traditional interfaces on the same screen at the same time.
Why are we complaining? Aren’t we getting the best of both worlds with a traditional desktop interface and a tablet interface on the same device? Well, the Metro-style tablet interface is the default, and users cannot simply live in the Desktop interface—they must interact with the Metro. It seems that an interface best suited for consumption is an uninspired choice for a device best suited for productivity.
Don’t put all your eggs in the tablet basket just yet. Tablets and mobile devices may be the best thing since the invention of the Web; they enable us to be connected at all times and in many ways make our lives easier by simplifying many of our everyday tasks. But PCs will still be around for complex tasks and can coexist peacefully with mobile. And in complex tasks, minimizing interaction cost is of the essence—and is much harder to do on a mobile device or with a mobile-based design.