Here’s another topic that just won’t die. I have been hearing about this almost as long as I’ve been using computers—and by “almost” I mean since I was, maybe, ten. Because I’ve been using computers since birth, judging by the actual non-digital photos of me from that time in my life. I can’t say I remember that.

Anyway, one of a handful of TV shows I used to watch at the approximate age of ten was called Beyond 2000, which had people talking in high tech sounding accents about things that we were going to have beyond the year 2000. One of the things I always wondered about the show was what they were going to call it when it actually got to be the year 2000. Beyond 3000 initially seemed like the logical choice, except the show was talking about things we’ll be using beyond the specified number-year, and who knows what we’ll have by then. So it would make a lot more sense to call it, say, Beyond 2010, and then just keep upping the year in the title as it drew close, so the predictions were at least accurate for the time frame. Of course that could become cumbersome…

Yes, I had considered this line of reasoning when I was ten. I was very concerned. But guess what? The problem has been solved. I only had to wait fifteen years to find out what they were going to do.

Now I’ve really digressed. What I was going to say was that sometime in the past, probably on the show called Beyond 2000, I was introduced to the concept of the 3D computer desktop. Over the years it has repeatedly popped up, always slightly improved, and always being heralded as something new. Needless to say, it’s not, yet I saw it presented just a few weeks ago as the Desktop of the Capital-F Future.

The idea is basically self explanatory. Rather than having flat pictures and words spread out across the two dimensional surface of the computer screen, you’ll be in a 3D environment to navigate the computer’s file structure. It could look like anything—your house, your office, or even some far away, exotic location such as your friend’s house. The possibilities are limited only by your imagination, the amount of time you’re willing to spend customizing the environment, the power of your 3D accelerator, the number of upgrades and props you can afford to purchase to populate the world, the color gamut and resolution of your monitor, and the amount of RAM you have. But other than all that stuff, the possibilities really are endless—pretty much.

Despite all these people thinking it’s the next big thing, it has stubbornly refused to be used at all. Back in the day when Wolfenstein 3D was at the very height of fashion (not to mention immersion), I myself tried a 3D desktop. It was excrutiatingly slow, since it was rendering a true 3D environment. But that’s no longer an excuse today. We have—indeed, we have had for some time now—the ability to easily create grand 3D worlds, and it would be a trivial matter to have these work with a file structure. Obviously people like hanging out in 3D worlds because we have non-games like Second Life which are completely open ended, allowing you to do such things as practice all manner of bizarre sexual fetishes. And chat.

What we haven’t seen is this same technology applied to a computer desktop and have it be even slightly successful. Why?

There are probably a number of reasons why 3D desktops will never take off, but here are the two I think are primarily responsible.

1) A 2-dimensional space limits the amount of chaos that can be created in that space. It should come as no surprise to anyone that people who have messy real-life desks also have messy virtual desks—and vice versa. (Although, I also know plenty of people who have clean real life desks yet still have messy virtual desktops. I suspect the reason for this is an unfamiliarity with the virtual environment and how it works.) I’ll be the first to admit that my desktop is not a paragon of orderliness, however, I’ve known some people that keep almost every file ever worked on on the desktop. Even using the “auto arrange” feature doesn’t help. It just serves to place the chaos in straight rows. But there’s no organization there, and it would be a time consuming pain to clean it up, so it stays messy. And that’s a 2-dimensional mess.

Extrapolate this into three dimensions. I already know what would happen there. I’ve got a real-life desk, too. If I were more industrious, I would take a photo of my real life desk just to prove the point. It would be like that one time I dissected my wallet. On my personal desk I have, among other things, a stack of VHS tapes stacked 16 high, an unopened bottle of vitamin C tablets, and a wristwatch that stopped working two years ago. And that’s in addition to important stuff, like bills.

Now imagine the same thing in 3D. No one wants a 3D computer desktop because it’s already a pain to manage their 3D real desktop. In another sense, the computer desktop is like a bulletin board in an already messy 3D space. You can 2-dimensionally mess up a bulletin board, but only to a certain extent. Using a 3D virtual desktop is more akin to having a portal to a totally different world. Now you have two offices to organize, and most people can’t handle one.

So a 2-dimensional desktop, while limited in capability, is actually a blessing in disguise. It is the most manageable number of the fewest number of dimensions in which we can usefully convey information. (Obviously, because you can’t convey much useful information in a functionally 1-dimensional space.)

2) The second reason 3D desktops won’t catch on is due to Fitt’s Law. Or, more precisely, a varient of Fitt’s Law designed for 3D virtual worlds which I have been unsuccessful in tracking down. It might not be formulated yet. It doesn’t matter; Fitt’s Law is an equation which predicts how long it will take someone to acquire a target with a pointer based on the distance between the two and the size of the target. Generally speaking, the formula shows that the smaller and further away the target is, the more time it will take the point to it.

This is why important buttons are bigger and closer to the cursor, while less important features are squished in little squares to the corner of the screen. In a 3D environment, there would probably be these sorts of messages popping up and conforming to Fitt’s Law in their usage of buttons, but in navigating file structure and icons, the law doesn’t work.

But it doesn’t matter, because even without the fancy math, you can still see that it would still take longer to acquire the target. There are two ways the target could be presented in a 3D environment. It could still be a 2D icon, just plastered on the wall or floor, or on a 3D virtual representation of a real desk surface, or it could be a 3D prop in the environment.

Both of these take more time to acquire than having the same information presented in a 2D space. In order to contain the same number of icons that are on today’s 2D desktop, you would need to find a balance between the size of the icons in the room and the size of the room itself—but here’s the problem. Regardless of how you arrange a 3D desktop, it’s still subject to the laws of perspective. This means if you’re acquiring a target on the wall, you have to “walk” to the wall, and then hit the target—which changes shape and size depending on where you “stand”. On a 2D desktop, even if you can’t see the desktop, the “walk” to get there is just a few mouse clicks, and the targets are always the same. In other words, now you have to find a balance between moving forward and backwards from the target and acquiring the target itself—and that movement also take a finite amount of time. It would be an interesting experiment to see if human brains can do the optimal amount of movement + target acquisition in one step, or if people continually overshoot the bounds, making for a less efficient selection process. This would be especially true when the maneuvering and aiming is done in the same process, as is the case today with most first person shooters.

If you’re using props instead of icons, you could then do it by having items scattered around the room. Then it’s just like I said earlier: a 3D respresentation of a real office, and you have to hunt crap down just like you do in real life. It’s taking all the things that suck about real life and importing them into the computer, using an environment that is not suited for the purpose of file navigation for any reason other than it’s how we experience the world in real life.

In the same way that the mouse cursor moves in a finite speed across the 2D desktop, so too would you have to move in a finite speed in the 3D desktop. Finding a particular program then becomes a matter of navigating through rooms and selecting the target. The 2D desktop, on the other hand, is like a series of pages stacked on top of each other which you shuffle around. The computer can shuffle pages much faster than you could if you actually had a stack of pages in real life. This is a better way of organizing information, then, because it uses the speed of the computer to pull any particular page you need with the simplicity of organizing information in two dimensions, which is more natural for us, anyway.

Basically what it comes down to is that when people are working or doing some non-recreational task that requires utility, they want the simplest and fastest environment possible in which to work. A computer desktop doesn’t need to be any particular way, but your average, everyday user is going to use the one that doesn’t require any effort more than is absolutely necessary. A 3D desktop is not this environment, a 2D desktop is. That’s what people will continue to use.

Actually that last paragraph raises a good point. If anything, computer interaction will grow to the point where it will require less effort than it already does. Currently, to use a computer I only need to move my fingers and a little bit of one arm. This takes a small bit of effort. Once we get to a point where you can put some sort of claw-like device on your cranium and control the computer by thinking about it—and it’s more accurate than the system we have now, which, in all fairness, is pretty dang accurate—that will be the control scheme of choice. At that point, I suspect the 2D desktop will be significantly simplified, if only because space that is devoted to common commands today, such as “save”, will be unnecessary.

What we also won’t see is that sort of nonsense that was in Minority Report, with Tom Cruise waving his arms around like some sort of cyberpunk mage, trying to control the windows on his 2D desktop. That take a lot more work than we have now; nobody would ever put up with it. Did they un-invent the mouse in the future? I don’t get it.

In any case, hype only carries an idea so far. The bigger question to ask is “will people put up with it”? If it takes more work to accomplish the same thing than what we’re already using, forget it. People aren’t that dumb.

-Ted