In the decade following the publication of William Gibson's wildly popular science fiction novel Neuromancer, we've seen an entire world of technical innovations - first found in the pages of his text - come to fruit. The technical community - touched by Gibson's vision of a "consensual hallucination" of computers, named cyberspace or the "Matrix" - began to build the pieces needed to make his vision real. Gibson's cyberspace was a network of all of the computers in the world, together communicating with each other and billions of individuals, who would use the Matrix for education, for business, and entertainment.
In 1984, the Internet consisted of a few thousand computers; most of these were located at universities, military installations or defense contractors. Beginning as a "fallout shelter" for the defense data networks, the Internet grew slowly, mostly because it was very difficult to use. Even expert users needed to refer to manuals to understand the cryptic command sets, and no one could remember where important information was located - there was no map to Internet resources. This situation created a great opportunity for a generation of systems administrators ("sysops") who could maintain the big Internet systems, and even find things for you. More than just administrators, they were community librarians and maintained the community's memory.
Five years later, the size of the Internet - and the amount of data within it - was beginning to overwhelm everyone, including the systems administrators.The Internet needed context; it was important to comprehend how different documents related to each other, especially in the sciences, where new work always draws upon the prior work of others. This problem, especially acute in physics, faced Tim Berners-Lee, a young software engineer at the European Center for Nuclear Physics, CERN, in Geneva. To counter it, he created a set of protocols which became the core of the World Wide Web. With these protocols it became possible to show the connection between documents, and to connect these documents together. This worked well for the physicists at CERN, who could now create documents that would illustrate the correlation between last month's papers, this week's experiments, and next year's proposals.
While the Web solved a fundamental problem with the Internet, it needed an interface which could mix different document types - text and images, for example - into a single document. Marc Andreesen, an undergraduate at the University of Illinois, created a Web browser which could do this, and gave it the name NCSA Mosaic. Marc went on to found Netscape Communications Corporation.
There are two ages of the Internet - before Mosaic, and after. The combination of Tim Berners-Lee's Web protocols, which provided connectivity, and Marc Andreesen's browser, which provided a great interface, proved explosive. In twenty-four months, the Web has gone from being unknown to absolutely ubiquitous. Yet, as good as it is, it's still somewhat confusing. The Universal Resource Locator (URL) - which tells you where things are in the Web - is pretty cryptic. The URL for my home page is http://hyperreal.com/~mpesce. I could tell that to my mother (who has Web access through AOL), but she'd never remember it, and it's unlikely that anyone could, for very long. It's not human-centered information, and, because it doesn't make much sense to us, it's easy to become confused when using it.
In October of 1993, I downloaded NCSA Mosaic and the Web server onto my SUN workstation and began to explore. I knew, as small as the Web was, that it would some day span the whole Internet.
I also could see that it wasn't enough. Linked pages are great, and that would be enough if the world were entirely made of books - but it isn't.
We'd need a sensual interface to the Web - one which could represent the real world as well as the world of publications - for it to be truly useful to large numbers of people. A sensual interface would need to work in three dimensions, just like we do; then I could tell you to get to my home page by saying, "go down Market Street, make a right on Noe Street, down two blocks and make a left on Henry Street." That would make a lot more sense to people, and would open the Internet to people who'd never dream they'd be able to use it.
That December, Tony Parisi moved to San Francisco, where I live and work. We both came from Boston - where we had some friends in common - and he gave me a call after he arrived. On New Year's Day 1994, amid the packing boxes in his new apartment, I sat and talked with him and his wife Marina Berlin about this crazy idea I had to create a "virtual reality" interface to the Internet. Tony's a programmer (among his other talents) and I interested him in helping me develop this interface. His talents in computer languages perfectly complemented mine in networking. We needed some help with the graphics, though - 3D graphics are very sophisticated and difficult to program. I had the good fortune to be introduced to Kate Seekings and Servan Keondjian, founders of a tiny company called RenderMorphics. They'd created a programming library called Reality Lab which made it possible for any programmer with a basic understanding of 3D graphics to write applications with sophisticated 3D capabilities. (RenderMorphics was acquired by Microsoft in March and Reality Lab is now a part of Microsoft Windows.)
On the 14th of February 1994, our experiment was up and running; Tony and I had written the first 3D Web browser. We could go to a Web site in 3D, or click on a link in a Web document and enter a 3D world. Just as important, we could go from a 3D world back into the Web; this meant we could construct 3D models of the various resources in the Web. I sent mail to Tim Berners-Lee, telling him what Tony and I had done; in turn, he invited us to present our work at the upcoming First International Conference on the World Wide Web. There, Dave Raggett - one of the fathers of HTML - named our creation a "Virtual Reality Markup Language" (we changed 'markup' to 'modeling' later on) - and Brian Behlendorf, the sysop at WIRED magazine, volunteered his organization's computers and disk space to set up a mailing list to discuss VRML issues. We knew that our work was preliminary, and we wanted to establish a community forum where we could design a real VRML together, drawing from the enormous talent available to us on the Internet. Within a week, over a thousand people signed up for the list, and today roughly 3000 people worldwide contribute to the discussion about what VRML should be, and how it should be implemented.
Later that year, Silicon Graphics took the file format of their Open Inventor scene description language and put it into the public domain; that file format, commercially useful and debugged, became the basis for VRML 1.0. In October, at the second Web conference, in Chicago, we presented our results to packed room of 100 Web researchers and developers. A real buzz was beginning - over the winter many companies began to develop VRML browsers, including SGI, Intervista Software (Tony's company), Template Graphics Software, and Paper Software. All of this was aided by SGI's contribution of a VRML parser - which converts the text of VRML files into objects the computer can understand - into the public domain, as a gift to the entire VRML community.
On the 3rd of April 1995, the worldwide announcements of VRML began, and so did the hype. A month later, a two-page article in NEWSWEEK heralded the advent of a technology just barely out of the womb. With just one VRML browser in release (SGI's WebSpace) and practically no tools, it seemed as though VRML might drown in its own hype. But the spring and summer saw a slew of releases from other companies, each eager to contribute their best efforts to the design of cyberspace. The Interactive Media Festival produced a VRML model of their Los Angeles-based event; research chemists started to exchange the results of their work as 3D molecular models; and an architect in San Francisco produced a VRML realization of South Park, the center of our Multimedia Gulch.
We're at the beginning of an enormous revolution; computing and the Internet are just about to dive into a three-dimensional universe which will change both forever. Just as the two-dimensional graphical user interface swept in and took over in the 1980's, we'll see the real world interfaces of cyberspace become commonplace in the years before the millennium. Right now, cyberspace is mostly empty; we're in an awkward period between the creation of the VRML specification and the release of tools which can really do VRML justice. When that happens, thousands of graphic artists, architects, sculptors and urban planners will create their own interfaces to the Web, which we'll use to communicate with them, and with each other.
Mark Pesce - co-architect of the VRML standard, travels the world evangelizing the benefits of 3D space on the World Wide Web. Mark is a tireless student of the history, future, and impact of 3-D world-building on international community and commerce.