Comment Re:I think you're thinking too hard and the author (Score 5, Insightful) 503
Except that the desktop cannot work using the phone/tablet model because user expectations do not suggest that metaphor when they sit at a desktop.
Even if the desktop metaphor was too complex to master, users still sit down at a desktop and think, "now where are my files?" because they intend to "do work in general" (have an array of their current projects and workflows available to them) rather than "complete a single task."
As was the case with a desk, they expect to be able to construct a cognitive overview of their "current work" at a computer—an expectation that they don't have with a phone, which is precisely experienced as an *interruption to* their "current work." KDE, Gnome, and most recently Windows 8, made the mistake of trying to get users to adopt the "interruption of work" mental map *as* the flow of work. It's never going to happen; they need to be presented with a system that enables them to be "at work." In practice, being "at work" is not about a single task, but about having open access to a series of resources about that the user can employ in order to *reason* about the relatedness and next steps across a *variety* of ongoing tasks. That's the experience of work for most workers in the industrialized world today.
If you place them in a single-task flow for "regular work" they're going to be lost, because they don't know what the task is that they ought to be working on without being able to survey the entirety of "what is going on" in their work life—say, by looking at what's collected on their desktop, what windows are currently open, how they're all positioned relative to one another, and what's visible in each window. Ala Lucy Suchman (see her classic UX work "Plans and Situated Actions"), users do not have well-specified "plans" for use (i.e. step 1, step 2, step 3, task 1, task 2, task 3) but are constantly engaged in trying to "decide what to do next" in-context, in relation to the totality of their projects, obligations, current situation, etc. Successful computing systems will provide resources to assist in deciding, on a moment-by-moment basis, "what to do next," and resources to assist in the construction of a decision-making strategy or set of habits surrounding this task.
The phone metaphor (or any single-task flow) works only once the user *has already decided* what to do next, and is useful only for carrying out *that task*. Once the task is complete, the user is back to having to decide "what to do next."
The KDE and GNOME experiments (at least early on) hid precisely the details necessary to make this decision easy, and to make the decision feel rational, rather than arbitrary. An alternate metaphor was needed, one to tell users how to "see what is going on, overall" in their computing workday. The desktop did this and offered a metaphor for how to use it (survey the visual field, which is ordered conceptually by me as a series of objects). Not only did the KDE and GNOME not offer a metaphor for how to use this "see what is going on" functionality, they didn't even offer the functionality—just a series of task flows.
This left users in the situation of having *lost* the primary mechanism by which they'd come to decide "what to do next" in work life for two decades. "Before, I looked at my desktop to figure out what to do next and what I'm working on. Now that functionality is gone—what should I do next?" It was the return of the post-it note and the Moleskine notebook sitting next to the computer, from the VisiCalc-on-green-screen days. It was a UX joke, frankly.
The problem is that human beings are culture and habit machines; making something possible in UX is not the same thing as making something usable, largely because users come with baggage of exactly this kind.