Virtual worlds today are walled gardens, their 3-D landscapes divided in much the same way that AOL, Compuserve, and others divvied up the 2-D Internet in the 1990s. Now a group of more than 20 companies, including IBM, Linden Lab, Multiverse, and Forterra Systems, has begun talking about how to link together today’s virtual worlds to form a 3-D Internet by establishing a set of common standards. Although there are still many questions about what form that 3-D Internet might take, a major goal will be to make it possible for users to move from one world to another as easily as people today navigate between websites.

“We see this as the next logical step in the growth of the Internet as a whole,” says Michael Rowe, manager of 3-D Internet and intraverse (an in-house virtual world) for IBM. Rowe says that today’s walled-garden situation hinders the development of the Internet by making it exceedingly difficult to offer virtual goods and services in more than one location. For example, he says, someone who wants to create a virtual storefront in several virtual worlds now has to design and build the store separately for each world, using radically different processes. Standards for a 3-D Internet, on the other hand, might allow the shopkeeper to provide links that people could use to reach the store from any virtual world, in much the same way that people can now use a domain name, such as Amazon.com, to point to popular shopping sites.

To make the virtual-world experience as rich as possible, Rowe says, developers need to find ways to consistently render objects–including avatars, which people use to represent themselves in virtual worlds–across several worlds. For example, he says, someone who has built a car in MTV’s Virtual Pimp My Ride might want to take it into Linden Lab’s Second Life to show it off. In order for the car to remain the same in both worlds, its underlying programming must make sense to both systems. Adopting 3-D Internet standards, Rowe says, will allow developers to solve these types of problems and open their borders to one another.

But the issue goes deeper than virtual cars and shopping malls. Jaron Lanier, interdisciplinary scholar in residence at the Center for Entrepreneurship and Technology at the University of California, Berkeley, and a virtual-reality pioneer, says that the search for a 3-D Internet is important for humanity. “Human cognition was designed to function in 3-D, and our computation eventually has to have a 3-D interface to maximize the matchup with the human brain as it evolved,” he says. People will need to find a way to combine a concrete, 3-D spatial understanding with the connective power of the 2-D Internet, Lanier says.

So far, a few steps have been taken toward this goal. Jon Watte, CTO of Forterra Systems, has proposed a possible standard designed at his company, called paged terrain format, which he hopes can keep virtual worlds literally on common ground. The format, if adopted, would allow developers to create landscapes so that multiple systems agree on the features of any given terrain. Paged terrain format is particularly designed to allow developers to use high-quality graphics, a feature that Watte says is missing from existing terrain standards, which were developed by the military. Forterra plans to make the specification for the format freely available.

Other companies, including Multiverse and Linden Lab, have designed platforms that independent developers can use to create virtual worlds more easily than building them from scratch, resulting in networks of virtual worlds that are interoperable, since they share the same foundational code. Although these networks of virtual worlds are a small-scale version of what a 3-D Internet might eventually look like, further standards need to be developed to interconnect worlds built on different platforms.

Corey Bridges, cofounder and executive producer of Multiverse, says that his company has created code that developers can insert at their discretion, which would allow users to bring their avatars across world lines using the Multiverse platform. Bridges says that a key problem is making a platform flexible enough that it doesn’t force developers to create the same world over and over again, while maintaining the consistency needed to keep the worlds connected.

Cory Ondrejka, CTO of Linden Lab, says that the first step toward bringing disparate virtual worlds together is to search for existing candidates for standards. He plans to look at systems such as OpenID, which people could use to store information about their identities and transfer it between worlds at their discretion.

Lanier, who notes that he has many professional connections to people involved with virtual worlds, says that while he very much wants the 3-D Internet to succeed, he is skeptical about whether it will be possible for developers to agree on a set of standards. “There’s a virtual land rush of people who want to come in and grab the standard,” he says, noting that the history of IBM and Microsoft provides some indication of the money that can be made by establishing a standard. But Lanier thinks a successful standard for the 3-D Internet is unlikely to develop the same way that HTML did–that is, as an abstract definition that people then adopted. He thinks it is more likely that a well-designed package will become a standard, similar to the way that Adobe Flash is becoming standard for rich Internet applications.