Pondering the future of virtual worlds

On Wednesday at Siggraph, I attended an interesting panel (details here) on the subject of "The Potential of End-User-Programmable Worlds."

In addition to the two organizers, the panelists included Paul Hemp of Harvard Business Review, Asi Lang of Linden Labs (the company that runs Second Life), and Vernor Vinge. Vinge is a faculty alumnus of San Diego State University, but better known as the author of "9 or 10 science-fiction novels," as he says. (I asked about that uncertainty; he said it's more about the definition of a novel than his ability to remember what he's written.)

You may have seen some of the recent news about Second Life. Last week, Linden Labs shut down the virtual casinos that had opened on the service; this week, the largest virtual bank on Second Life is in danger of collapse. These events may be related. Apparently the problem with the bank is that it was paying high interest without earning enough return on deposits (from investments in virtual land, etc.) to maintain profitability. The result, allegedly, is that the bank owes depositors much more than it has in assets. And the bank is actually owned and operated by a single individual in Brazil with no banking experience or substantial personal assets, so depositors may just be out of luck.

Lang did not get into these controversies in his presentation, but he did give a good tour of some of the more interesting applications that Second Life users have developed. There are art galleries (including an impressive recreation of the Sistine Chapel), educational activities (one academic user created a molecular modeler; enter a room, name a molecule, and the application builds a 3D model of it), and convenient tools for interacting with the real world (such as a real-time US national weather map that users can walk through). Linden itself doesn't create these applications; it just provides the environment and the development tools.

Panelist Hemp described some of his research into commercial activities on Second Life, which in his opinion aren't living up to the early hype. For example, the first real-world retailer to open an office on Second Life-- American Apparel, a clothing company-- has now closed its office on Second Life. The company isn't saying why; it has a website on the subject (here) but it offers no real explanation.

Hemp thinks these virtual offices just aren't worth the effort unless the company figures out how to attract virtual visitors. He compared the Second Life headquarters of Dell with that of Reuters. Dell's building, he says, is almost always empty. Other than a giant model of a PC that a visitor can fly through, there isn't much reason to visit. Reuters, on the other hand, offers streaming video on big virtual TV screens; visitors can sit down in a virtual couch and watch today's news.

Vinge wrote "True Names," the seminal work of the virtual-reality genre. In his talk, he described how cyberspace is "leaking into the real world" as people make increasing use of virtual reality to communicate with each other for both personal and business reasons.

He believes that this trend will continue, and I certainly agree. He points to higher-bandwidth Internet access, especially for mobile users, and smarter devices that are aware of their location and, through cameras and other sensors, their surroundings. Combining these elements means that "reality becomes its own database;" objects know what and where they are, and can communicate this information among themselves.

Eventually, he says, Web applications such as Google Maps could be augmented with this information to create something that operates like a virtual world, but represents the real world. A display device, aware of its position and orientation in the real world, could act as a window into the virtualized world. It could show information, virtual objects, and even virtual people. Imagine walking down a street in San Francisco while interacting with people touring the virtual version of the same street.

I'd also like to pass along Vinge's mention of an open-source virtual world project that appears to have some potential for private enterprise applications. The Croquet project (here), or something like it, may make it practical for businesses to create secure, practical virtual worlds for project teams, especially when the members are spread out around the world. (The science-fiction novel "Thirteen" by Richard K. Morgan, which I mentioned here, makes heavy use of this idea.)