The cash-cow MOO meets the Zen MU

You access a shared virtual space. Your see your own hand poised above a crisp, ingot-shaped object labelled "schedule". Various objects are arranged in clusters on what seems to be an infinite piece of graph-paper, and you see other hands moving them around. One hand picks up a square ingot labelled "SEE" and drops it onto "you": your colleague Afia in Yokohama appears, asking when you can run the turbine-blade simulation,

You join a shared virtual space. You see a stone plaza, in a landscape vaguely reminiscent of Big Sur. Standing on the beige flagstones are several slightly geeky-looking cartoon figures. Three announce themselves as Ms Jones, Bob and Linda; the fourth must be you.

You plunge, nose-first, into a shared virtual space. You are greeted warmly by a translucent Cubist dragon named Clytemnestra who invites you to relax in the hot-tub. Miranda drifts into the garden -- carrying a copy of Shakespeare's Tempest. She ignores your greeting, and kneels before an empty plinth labelled mysteriously "lag gods". Clytemnestra, meanwhile, continues her conversation, about the philosophers Deleuze and Guttari, with a small constellation of yellow stars.

The first scenario comes from the Shared Alternate Reality Kit project at the Xerox Palo Alto Research Centre five years ago. The second is from the WorldsAway "fully animated virtual world" project which Compuserve and Fujitsu plan to launch later this year (OnLine, 16 May). The third exists, as yet, purely in the text-only environment of MOOs (OnLine, 24 November 1994). But, one way or another, the hallucinatory interactions of MOOs are going to go graphical very soon.

The last Next Big Thing to hit cyberspace -- the World- Wide Web -- was developed in the traditional way by academics, spread by the traditional internet standard-setting process of "That's really neat! Can we run one?", and was commercialised later. It is a sign of the way that the net is evolving that in the next big thing -- the highly visual virtual environment -- the commercial developers appear to be ahead of the game. And yes, Microsoft seems to have a project, but isn't talking about it.

Tools to build do-it-yourself and more participatory consensual hallucinations in cyberspace are, however, beginning to arrive. The first step will be the integration of MOOs into the World- Wide Web, providing a "flat" but graphical interface with which a few million people are already familiar. As Chris Hand, Senior Lecturer at De Montfort University and co-ordinating a Web-MOO project, points out, the effect will be far-reaching: "Browsing the Web is a relatively passive, one-way experience oriented towards receiving data, while the MOO allows fully interactive multi-user dialogues."

The price of interactivity in a MOO at present is an interface from the Stone Age of computing, with typed commands like "say to Clytemnestra `hello'".

Last year several groups of academics, including Chris's, hacked up trial Web-MOO interfaces to demonstrate the potential for virtual tutorials and world-wide conferences. In these you could click on a character's picture to "speak" to them. One public pilot, running since last autumn, is ChibaMOO -- though this uses a non-standard "mark-up language".

The next step will be an on-screen representation of three- dimensional spaces and characters, transmitted using Virtual Reality Mark-up Language (VRML). This, in essence, defines scenes and objects built up of flat polygons (OnLine, 27 April).

One VRML browser -- part-built by Silicon Graphics, who stand to sell a lot of hardware if takes off -- is currently in beta test on unix machines, with a Windows version due soon. At least two more are in the wings. And an "authoring tool" called HomeSpace has been announced by ParaGraph International.

The limit on all these developments reaching the public is the amount of information which can be crammed down a phone line. This is presumably one reason WorldsAway users will create characters by picking clothes from a pre-defined wardrobe and attaching them to a flat body. (Add little paper tabs and you'd have a virtual cereal packet.) The interactive Doom game is only a little more flexible.

A kind of free-form 3D virtual reality is possible. Realistic motion for a minimal Robbie-the-Robot "body" built from 20 cuboids -- enough to have rectangular eyes, nose and mouth and thus a crude semblance of expression -- would require sending about 50,000 bits of compressed VRML per second -- twice the speed of the fastest practicable modems in 1995. But for Roberta to raise an arm smoothly, while keeping the rest of her body still, could require as little as 800 bits per second.

Once the tools are available, enthusiastic experiment will determine which more artistically Cubist representations better convey emotions. Indeed, confronted with a bandwidth limitation, people will tend to use disembodied heads as avatars: a reasonable amount of facial expression will thus be obtainable. They might choose to become a hand to demonstrate a spatial point to another character. Other body parts will, inevitably, be modelled. (Letters of protest are already being drafted.)

Philosophers, already intrigued by the fluidity of identity in text MOOs which offer the user creating a new character a choice of seven genders, will have a field day.

There will be plenty more for the philosophers to get their teeth into as these things develop. The traditional forms of group interaction on the net -- Usenet news and mailing lists -- may be the last great flourishing of text as a medium. They are asynchronous -- you have up to two days before you totally lose the thread -- and they leave traces. MOO-style interaction is (in principle) as immediate as conversation, and as evanescent (unless some system operator is keeping a log...)

The existing text MOOs already provide a happy hunting ground for sociologists, though participation is culturally restricted: anyone who's visited one will understand the idea of a Lonely Hearts ad which specifies "160 words per minute minimum; gender optional". When MOO-interaction becomes primarily visual, it will be an entirely new medium.

Mind you, like most things on the net, there's a hype-fringed chasm between the potential and what works now. You can spend hours failing to get any response at all from ChibaMOO through the World Wide Web -- the server is overloaded. When you do get connected, you confront the problem of "lag": at peak times it can be ten minutes between your greeting a character and receiving their response. A document on LambdaMOO declares, not obviously ironically, that the operators haven't figured out what causes lag -- hence the archive of definitely ironic lag-god theology in its library.

The god Mammom should also have temples on the MOOs. Whether used for socialising, mind-expanding fantasy or serious business, the potential demand is enormous. So, too, will be the marketing opportunities for ever-more-powerful hardware.

And as soon as users have splashed out for the inevitable VRML- rendering accelerator card, along will come the ads for artificial intelligence hardware. MOOs already contain primitive "artificial characters" -- "Miranda" in the example above might be one. Work is beginning on software "characters" which respond to emotions. You could find yourself applying the Turing test to everyone you meet in cyberspace: can you tell who's human?

home

Written: 15 Jun 1995

An edited and doubtless thus improved version of this article appeared in the Guardian OnLine section. TRY this; you may need to register.

This version is © copyright 1996 Mike Holderness; moral rights are asserted.

back