Is the Desktop Doomed?
It’s been eight years since the introduction of the first multi-touch device. Today, smartphones are ubiquitous, and have come to dominate the personal computing landscape. But the windowed desktop environment, controlled by a mouse or trackpad, still persists — and remains the go-to interaction model for those interested in getting work done. Why is the desktop metaphor still the best interaction model for productivity? How can lessons derived from years of experimentation with mobile platforms influence and improve the desktop experience? And what will happen to the desktop in the coming decades? To explore these questions, I think it’s helpful to first dial things back a bit, and examine the best practices of design in the real world. These tangible design constraints and solutions can provide insights into the creation of more intuitive digital interfaces, which has ultimately been the goal of both mobile and desktop platforms.
Reading Don Norman’s “The Design of Everyday Things” changed the way that I think about my interactions with objects in the real world. Door handles, cups, lamps, etc: most of the objects around us have been very consciously designed to be highly understandable, and to reduce confusion during our interactions with them. There’s a bit of a paradox here, in that the best-designed objects call very little attention to themselves. Instead, they are both highly functional and highly intuitive. Well-designed objects could be said to exhibit “as little design as possible.” In some cases, that goal — to promote clarity — is met, but in an imperfect way. I think the example of doors, which Norman explored extensively, is a helpful one.
We’ve all had the embarrassing experience of not being able to figure out how a door works. (Do I push or pull… where?) When a lot of people experience difficulty opening a particular door, it probably reveals a lack of proper design, not a lack of intelligence on the part of any given individual. A common solution to reducing confusion involves putting a sign on the door. But relying on a sign that says “push” is inelegant when the design of the door itself can promote an understanding of the appropriate action. A “push” sign, in the semiotic sense, is a means of communicating information, comprised of two related parts: the letters themselves are the signifier, and their interpreted meaning — “I’m supposed to push, not pull” — is the signified. But a more elegant design solution views the door itself as a sign. The correct interaction with the door, the signified interaction, can be written into the door’s physical appearance. A long horizontal bar can instruct us to push as much as the printed sign can, but without the extra work of reading and interpreting a written word.
Norman’s suggestion was really interesting to me: there could be an objective scale to evaluate the quality of design, and it was based mostly around simplicity and clarity. Good design could be said to denote or signify “purely,” meaning that an understanding of the object and all aspects of interaction with that object are contained and communicated within the design of the object itself. The same real-world design principles apply to our digital environments: in UI/UX, this premise is referred to as “affordance.”
Desktop User Interfaces
We’ve been interacting with digital systems for a long time now. Early computers had user interfaces based exclusively around text, and they ended up being too intimidating for the majority of people to use. As technology progressed, the development of bitmapped screens that could display a more intuitive visual interface became the true origin of mainstream personal computing. New interaction models that involved pointers, mice, and trackballs helped people navigate these GUIs. But just as people would become confused by the design of certain doors, these new interfaces were also unclear and caused frustration — most people have had the experience of trying to explain how to use a computer to their older colleagues or family members. It’s never easy.
Awareness of good design principles became a necessity in order to convey appropriate interactions with the virtual objects within these desktop environments. Metaphors like windows, software buttons, and menus were created. Even the metaphor of the “desktop” itself had to be invented, more or less from scratch. It was a big leap forward. Like in the real world, there have been many approaches taken to reduce confusion in virtual environments, some with more success than others. But on a fundamental level, the desktop environments we use to this day still rely heavily on signs that must be learned and interpreted.
In a lot of cases, our understanding of interactions within a virtual environment is the product of learning. All modern desktop environments share a similar set of common abilities and constraints. Windows can either fill the entire screen, be minimized, or closed entirely. While navigation and window management are among the most basic tasks on a computer, they are probably not intuitive for most new users because they are still built upon abstractions that must be learned. For example, the Mac OSX “stoplight” buttons are interactive cues that are simple, but still require interpretation: red means close, green means minimize, etc. Microsoft Windows has a similar set of buttons: “X” means close, “_” means minimize.
So my guess is that, due to their complexity, modern desktops aren’t inherently intuitive, and still require learning / interpretation — much like a sign on a door that tells you what to whether to push or pull. We’ve seen that doors can be designed to signify purely, simplifying user interaction. This leads to a key question: is there a way to improve digital interfaces so that we don’t have to rely heavily on written signs?
Touch Screen Interfaces
As it turns out, it is possible to create a more intuitive interaction model: touch screens on mobile devices. The first iteration of iOS on the original iPhone pioneered this model in 2007. It involved the development of a few new paradigms, all of which are now commonplace. Scrolling and swiping mirrored natural motion. Multi-touch allowed “pinch to zoom” and more intuitive rotation of pictures and maps.
Interestingly, iOS and most modern touch screen interfaces did away with the desktop metaphor, focusing instead on grids of applications (apps sound even less intimidating) that you could touch. This is much more intuitive for those who aren’t already fluent in digital media: want to check your mail? Just touch the little icon that looks like an envelope. In the past, this action was obscured behind several additional layers of interaction that involved a web browser, bookmarks, URLs, and other confusing elements. At the new level of abstraction enabled by touch screens, the process is much simpler.
The touch model works well for consuming information, but has yet to influence desktop environments to a large degree (Windows 10 seems to have reversed some of the more polarizing touch-oriented design trends present in Windows 8, for example). It’s also interesting to note that, by and large, desktops are still the preferred platform for those looking to get a lot of work done. The windowed desktop environment must be superior for this use case. But why?
When we work on the computer, what is it that we’re doing — and why is the desktop metaphor still better than mobile for these purposes? I think the answer is document manipulation. People prefer to do both administrative and creative work on desktops primarily because they are opening, sorting, moving, editing, and creating documents — like papers, proposals, spreadsheets, presentations, and designs.
While mobile, touch-based systems are very good for consuming content, desktop systems are better for creating content, for a number of reasons:
- Accessible File System — a hierarchical filing system is a must, in order to organize a lot of documents. iOS famously obscured the file system from users, opting instead for a homescreen that presents a grid of apps running in fullscreen.
- Hierarchy / Context — the concept of context also applies to screen real estate. It’s helpful to be able to open multiple documents and applications at once within windows that can be moved, resized, and compared directly to one another.
- Larger Work Area — similarly, context is aided by a bigger screen with more real estate.
- Working Posture — most workstations are meant to be used at a desk, either seated or standing. The screens are not only larger, but also placed vertically in front of the user. This encourages the use of a keyboard and mouse/trackpad that lay flat — better than touchscreens for extended use. (Touching a vertical screen will lead to exhaustion).
That is not to say that ideas inherited from mobile UI design can’t be applied to our desktops. In fact, a thoughtful blend of both interaction modes can — and probably should — be implemented to better approximate the ability of well-designed real world systems to “signify purely.”
One aspect of design that propelled mobile UI in the early years was skeuomorphism. I think it was a good trend, because it allowed people to understand interfaces very easily. Sure, there may be examples of when skeuomorphism went too far. But overall, I think that skeuomorphic elements are a net positive: people intuitively understand what a digital element is supposed to do. I think that the trend towards flat design is overzealous. Obscuring appropriate interactions behind excessive whitespace or esoteric gestures is actually a step backward.
Here are a few ideas, inherited from mobile, that I think should inform the design of Desktop UIs over the coming years:
- Apps — on your phone, when you want to look at your mail, you touch the mail icon. With many applications moving to the web, the delineation between browser tab, desktop file, web app, and desktop app will become more blurry. A seamless interface should connect these previously disconnected worlds. Chrome OS is an example of an early pioneer in this area.
- Skeuomorphism — the trend toward flat design will reach a plateau, and the pendulum will swing back a bit. Shadows and textures will be helpful in promoting a sense of depth and reality within digital interfaces. See material design.
- Tacticle Input Methods — not exclusively touch, per se. Incremental improvements to pointing devices will solve the “gorilla arm” problem while allowing desktop users to use more intuitive gestures like “force touch.” I think Apple has the right idea by embedding touch gestures into computer mice. I just don’t think the magic mouse is a strong implementation yet.
- Portability — your work should be able to come with you. One of the great (or maybe frightening) things about our phones is that they are always on and always nearby. They always preserve their state, and your open apps, browser tabs, etc. In the future, a desktop class work environment should be able to come with you too. It will be personalized, just like your phone. In fact, in the future, our phone might become a singular device that can pair seamlessly with larger screens in your vicinity — at the office, at home, etc — to offer the best user experience for each situation.
“The First 100 Years”
In the near future, virtual reality (VR) will present huge opportunities to developers. They’ll be able to experiment with new interaction models and new input devices. I can definitely envision a future where many people will choose to work predominantly in highly realistic simulated environments — with major implications for communication, collaboration, and rapid creative ideation. But even within those environments I believe we will still mostly be using “screens” with desktop UIs to get our work done, instead of (for instance) manipulating virtual representations of documents with our hands. Why? Because even now, people prefer receiving email to snail mail. Services like Evernote encourage people to “go paperless.” I personally have many more folders and files on my hard drive than in my filing cabinet at home. And I can access and organize (and obsessively reorganize) them faster when they’re on a hard drive. In VR, there will be definitely be simulated representations of paper documents — but I expect that there will also be simulated representations of screens (no longer bound by the laws of physics, they might be floating, or have dimensionality) with desktop UIs that are optimized for creative work and fast, easy document manipulation.
I think that the quality of design is probably stronger now than at any point in the past; there’s a lot of talent and attention paid to making beautiful, intuitive interfaces. As the recent frenzy over mobile subsides and evolves, I’m looking forward to seeing how desktop and mobile UI continue to develop in tandem and in relation to new emerging interaction models — like wearable devices and VR.
Ultimately, the goals of good design have remained unchanged for as long as they have been studied. Simplicity and ease are critical. Intuitiveness and “joy,” too. There will never be a perfect interaction model, but that hasn’t stopped us from trying. So we’ll all keep striving, hopefully putting the stitched leather — and also the flat tiles — behind us.