Cheaper than printing it out: buy the paperback book.

To win Populous, you've got to think like a god. You cannot live many small individual lives and succeed. Nor can you manipulate every individual simultaneously and hope to remain sane. Control must be surrendered to a populous mob. Individuals of Populous land, who are no more than a few bits of code, have a certain amount of autonomy and anonymity. Their pandemonium must be harnessed collectively in an intelligent way. That's your job.

As god, you have only indirect control. You can offer incentives, play with global events, make calculated tradeoffs, and hope that you get the mix right so that your underlings follow you. Cause and effect in this game is coevolutionarily fuzzy; changing one thing always changes many things, often in the direction you wanted least. All management is done laterally.

Software stores sell other god games: Railroad Tycoon, A-Train, Utopia, Moonbase. They all enable you, the neo-god, to entice citizens to create a self-sustaining empire. In the game Power Monger you are one of four godlike kings hoping to rule supreme over a large region of a planet. The population below, which numbers in the hundreds, is not faceless. Each citizen has a name, an occupation, and a biography. As deity, your job is to urge the citizenry to explore the land, mine ore, make plows, or hammer them into swords. All you can do is adjust the society's parameters and then set the beings loose. It's hard for a god to guess what will emerge. If your folks manage to rule over the most land, you win.

In the brief annals of classic god games, the game of Civilization ranks pretty high. Here the goal is to steer your bottom-up population through the evolution of culture. You can't tell them how to build a car, but you can set them up so that they can make the "discoveries" needed to build one. If they invent a wheel, then they can make chariots. If they acquire masonry skills, then they can make arithmetic. Electricity needs metallurgy and magnetism; corporations first require banking skills.

This is a new way of steering. Pushing too hard can backfire. The denizens in Civilization might revolt at any time, and occasionally they do. All the while you are racing against other cultures being tweaked by your opponent. Lopsided contests are quite common. I once heard an avid Civilization player boast that he overran the other society with stealth bombers while they were still working on chariots.

It's only just a game, but Populous embodies the subtle shift in our interactions with all computers and machines. Artifacts no longer have to be inert homogeneous lumps. They can be liquid, adaptable, slippery webs. These collectivist machines run on myriad tiny agents interacting in ways we can't fathom, generating results we can only indirectly control. Getting a favorable end result is a challenge in coordination. It feels like herding sheep, managing an orchard, or raising kids.

In the development of computers, games come first, work later. Kids who become comfortable relating to machines as if they behave organically, later expect the same from machines at work when they are older. MIT psychologist Sherry Turkle describes the readiness of children to perceive complicated devices as organic as an affinity for a "second self" -- a projection of themselves onto their machines. Toy worlds certainly encourage that personification.

SimEarth, yet another god game, bills itself, somewhat tongue in cheek, as "the ultimate experience in planet management." An acquaintance of mine told a story of making a long car trip with three 10-12-year-old boys in the back seat, the trio equipped with a laptop computer running SimEarth. He drove while eavesdropping on the boys' conversation. He gathered that the boys had decided their goal was to evolve intelligent snakes. The kids:

"Do you think we can start the reptiles now?"

"Oh shoot. The mammals are taking over."

"We better add more sunlight."

"How can we make the snakes smarter?"

SimEarth has no narrative or fixed goals -- a nonstarter for many adults. Kids, on the other hand, fall into the game without hesitation or instruction. "We are as gods, and might as well get good at it," declared Stewart Brand in 1968, who had personal computers (a term he later coined) and other vivisystems in mind when he said it.

Stripped of all secondary motives, all addictions are one: to make a world of our own. I can't imagine anything more addictive then being a god. A hundred years from now nothing will keep us away from artificial cosmos cartridges we can purchase and pop it into a world machine to watch creatures come alive and interact on their own accord. Godhood is irresistible. The hemorrhaging expense of yet another hero will not keep us away. World-makers could charge us anything they want for a daily fix of a few hours immersed in the interactive saga of our characters' lives, and to keep our world going we will pay it. Organized crime will make billions of dollars peddling crude artificial calamities -- first class hurricanes or high priced tornadoes -- to addicts compelled to buy. Over time, god-customers will evolve fairly sturdy and endearing populations, which they will be eager to test with yet another fully rendered natural disaster. For the poor there will surely be underground exchanges of generic mutant beings and pilfered scenarios. The headlong high of substituting for Jehovah, and the genuine, overwhelming, sheer love for one's private world, will suck in any and all who near it.

Because simulated worlds behave -- in a tiny but measurable way -- similarly to worlds of living organisms, the ones that survive will grow in complexity and value. The organic ambiance of distributed, parallel world-games is not mere anthropomorphism, despite the second self projected upon them.

SimEarth was meant to model Lovelock's and Margulis's Gaia hypothesis, which it succeeded in doing to a remarkable degree. Fairly serious changes in the simulated Earth's atmosphere and geology are compensated by convoluted feedback loops in the system itself. For instance, overheating the planet increases biomass production, which reduces CO2 levels, which cools the planet.

Scientists debate whether the evidence of self-correcting cohesion seen in the Earth's global geochemistry qualify Earth as a large organism (Gaia), or merely a large vivisystem. Applying the same test to SimEarth we get a more certain answer: SimEarth, the game, is not an organism. But it is a step in the direction of the organic. By playing SimEarth and other god games we can get a feel of what it will be like to parry with autonomous vivisystems.

In SimEarth, a mind-boggling web of factors impinge on each other, making it impossible to sort out what does what. Players sometimes complain that SimEarth appears to run without regard to human control. It's as if the game has its own agenda and you are just watching.

Johnny Wilson, a gaming expert and author of a SimEarth handbook, says that the only way to derail Gaia (SimEarth) is to launch a cataclysmic alteration such as titling the axis of the Earth to horizontal. He says there is an "envelope" of limits within which the SimEarth system will always bounce back; one must bump the system beyond that envelope to crash it. As long as SimEarth runs inside the envelope, it follows its own beat; outside of it, it follows no beat. As a comparison, Wilson points out that SimCity, SimEarth's older sister, "is much more satisfying as a game, because you get more instant and clear feedback on changes, and because you feel like you are more in control."

Unlike SimEarth, SimCity is the paramount example of an underling- driven god game. This award-winning simulation of a city is so convincing that professional urban planners use it to demonstrate the dynamics of real cities, which are also driven by underlings. SimCity succeeds, I believe, because it is based on the swarm, the same foundation that all vivisystems are based on: a collective of richly linked, autonomous, localized agents working in parallel. In SimCity a working city bubbles up from a swarm of hundreds of ignorant Sims (or Simpletons) doing their simple-minded tasks.

SimCity obeys the usual tail-swallowing logic of god games. Sims won't take up residence in your city unless there are factories, but factories generate pollution which drives away residents. Roads help commuters but also raise taxes, which drive down your ratings as a mayor, which you need to survive politically. The maze of interrelated factors required to construct a sustainable SimCity can unfold along the lines of the following fairly typical account from a heavy SimCity-using friend of mine: "In one city which I built up over many Sim-years I had a 93 percent approval in the public opinion polls. Things were going great! I had a nice balance of tax-producing commerce and citizen-retaining beauty. To lessen pollution in my great metropolis I ordered a nuclear power plant built. Unfortunately I inadvertently placed it in my airport's flight path. One day a plane crashed into the generators, causing a meltdown. This set fire to the town. But since I hadn't built enough fire stations in the vicinity (way too costly), the fires spread and eventually burnt down the whole city. I'm rebuilding now, differently."

Will Wright, the author of SimCity and coauthor of SimEarth, is thirtyish, bookish, and certainly one of the most innovative programmers working today. Because Sim games are so hard to control, he likes to call them Software Toys. You diddle with them, explore, try out fantasies, and learn. You don't win, any more than you might win at gardening. Wright sees his robust simulation toys as the initial baby steps toward a full march of "adaptive technologies." These technologies are not designed, improved upon, or adjusted by the creator; rather, they -- on their own accord -- adapt, learn, and evolve. It shifts a bit of power from the user to the used.

The origins of SimCity trace Will's own path to this vision. In 1985 Will wrote what he calls "a really, and I mean really, stupid video game" entitled Raid on Bungling Bay. It was a typical shoot-'em-up starring a helicopter that bombed everything in sight.

"To create this game I had to draw all these islands that the helicopter would go bomb," recalls Will. Normally the artist/author modeled the complete fantasy in minute pixelated detail, but Will got bored. "Instead," Will says, "I wrote a separate program, a little utility, that would let me go around and build these islands real quick. I also wrote some code that could automatically put roads on the islands."

By engaging his land-making or road-making module the program would -- on its own! -- fill in land or roads in the simulated world. Will remembers, "Eventually I finished the shoot-'em-up game part, but for some reason I kept going back to the darn thing and making the building utilities more and more fancy. I wanted to automate the road function. I made it so that when you added each connecting piece of island, the road parts on them would connect up automatically to form a continuous road. Then I wanted to put down buildings automatically, so I built a little menu choice for buildings.

"I started asking myself, why am I doing this since the game is finished? The answer was that I found that I had a lot more fun building the islands than I had destroying them. Pretty soon I realized that I was fascinated by bringing a city to life. At first I just wanted to do a traffic simulation. But then I realized that traffic didn't make a lot of sense unless you had places where the people drove to...and that led layer upon layer to a whole city; SimCity."

A player building a SimCity recapitulates Will Wright's sequence in inventing it. First, he makes the lower geographical foundation of land and water which support the road traffic and telephone infrastructure which support residential homes which support the Sims which support the mayor.

To get a feel for the dynamics of a city, Wright studied a simulation of an average city done in the 1960s at MIT by Jay Forrester. Forrester summarized city life into quantitative relations rendered as mathematical equations. They were almost rules of thumb: it takes so many residents to support one firefighter; or, you need so many parking spaces for each car. Forrester published his findings as Urban Dynamics, a book which influenced many aspiring computer modelers. Forrester's own computer simulation was entirely numerical with no visual interface. He ran the simulation and got a stack of printouts on lined paper.

Will Wright put flesh onto Jay Forrester's equations, and gave them a decentralized, bottom-up existence. Cities assembled themselves (according to the laws and theories of the god Will Wright) on the computer screen. In essence, SimCity is an urban theory provided with a user interface. In the same sense, a dollhouse is a theory of the household. A novel is theory told as story. A flight simulator is an interactive theory of aviation. Simulated life is a theory of biology left to fend for itself.

A theory abstracts the complicated pattern of real things into the facsimile pattern -- a model, or a simulation. If done well, the miniature captures some integrity of the larger whole. Einstein, working at the peak of human talent, reduced the complexity of the cosmos to five symbols. His theory, or simulation, works. If done well, an abstraction becomes a creation.

There are many reasons to create. But what we create is always a world. I believe we may be unable to create anything less. We can create hurriedly, in fragments, in thumbnail sketches, and streams of consciousness, but always we are filling in an unfinished world of our own. Of course we sometimes doodle, literally and metaphorically. But we immediately see this for what it is: theory-free gibberish, and model-less nonsense. In essence, every creative act is no more or less than the reenactment of the Creation.