New 3-D Graphics Are Revolutionary--but Are They Necessary?
The difference between personal computers and the more powerful engineering and scientific computers known as workstations has narrowed steadily over the last 10 years. Thanks to economies of scale, PCs have improved faster than their specialized cousins. But one key difference has remained: Workstations leave PCs in the dust when it comes to three-dimensional graphics.
Now two emerging technologies--virtual reality markup language and 3-D accelerator chips--are poised to make even this distinction disappear. The result may be the first truly qualitative change in personal computing since the Apple Macintosh.
But whereas companies ranging from Netscape Communications Corp. to Intel Corp. agree that the arrival of affordable, interactive 3-D graphics is revolutionary and will be a critical technology in 21st century computing, nobody has given much thought to exactly why. Instead, it is largely an article of faith that 3-D is inherently better than 2-D and that software will become easier to use once interfaces look more like the real world.
If you think about it, though, this makes as much sense as saying that word-processing programs should sound a bell at the end of the line.
Slavish imitation provides the benefit of familiarity, but it can also limit our imaginations about how to best use a new technology. Understanding the trade-offs--when to use 3-D and when not to--will be critical to the next generation of software.
For evidence, just take a look at some of the primitive and clunky 3-D interfaces on the Internet today. Trying to navigate through these spaces too often feels what it must like to walk through a fun house on acid.
Admittedly, this is partly due to the early stage of the technology. Many of these interfaces use VRML 1.0, a specialized language for describing 3-D objects and scenes. (Think of it as a 3-D version of HTML, the language used by the Internet’s World Wide Web to describe the texts and images that appear on a Web page.) VRML is important because it allows networking and 3-D to work together, but this first version suffers from serious limitations. The most important is that objects are completely static; there is no way to build a door that opens as you walk toward it, for example. Fortunately, a new version of the standard that better supports interactivity is due out later this year.
The other technical constraint on 3-D interfaces today is computational speed. Every time you move in a 3-D virtual world, the appearance of each 3-D object must be recomputed and redrawn. So the more complex and lifelike the world, the slower the animation.
This is where 3-D accelerator chips come in. These are special microprocessors designed solely with 3-D in mind; plug one into your PC and it will handle all the recomputing and redrawing at lightning speed while your computer’s main processor does more interesting things.
Nearly a dozen semiconductor companies, including NEC Corp. and nVidia, are racing to market with such chips, priced at less than $200.
However, these technical limitations are not enough to explain why attempts at 3-D interfaces usually fail miserably.
The real problem is that 3-D is too often used for its own sake. Interface designers can become too enamored of the technology, just as movie directors get carried away with special effects and forget about telling a story. The few 3-D interfaces out today that do work use 3-D as more than just eye candy.
The two best examples I know of are Doom, the popular shoot-’em-up game from id Software, and OnLive, a VRML environment where people all over the Internet can meet and chat (https://www.onlive.com). In both cases, the use of 3-D feels absolutely necessary.
In the case of OnLive, it’s because the notions of distance and space are crucial to preventing cacophony and allowing private conversations. As you move away from another person on the system, for example, his or her voice becomes quieter. In the case of Doom, it’s because 3-D allows you to feel fully immersed in the game in a way that 2-D never could. You really feel as if you are there, and the success of the game derives from that.
According to F. Randall Farmer, one of the founders of Cupertino, Calif., start-up Electric Communities and an expert on building virtual worlds, this points to the simple question that designers should ask themselves when considering using 3-D: “Do you really have to be there?”
In the case of a game like Doom, the answer is clearly yes. But in the case of, say, an online bank, the answer is probably no. Even though we are accustomed to physical banks, even though it would be easy to use VRML to create an online bank that looks like a real one, there is no reason you have to be there.
“3-D is an attribute, like the color blue,” Farmer once wrote. “Any time you read or hear about how great 3-D is, substitute the word blue for 3-D.”
His point was that technologists too often promote 3-D as the answer to everything. But if we can remember that lesson while still exploring the potential that VRML and 3-D chips are about to unleash, computing will become a far richer experience.
Steve G. Steinberg (steve@wired.com) is an editor at Wired magazine.
