23 June 2008

In the image of man

At the time of writing, Sporepedia lists more than 907,000 new species. Will this creativity and playfulness broaden imaginations to the pre-existing natural world or channel them away from it?

above: a contender for title of the world's ugliest dog.

1 comment:

Anonymous said...

From the New York Times, 2 September 2008

September 2, 2008
Gaming Evolves

By CARL ZIMMER
NEW HAVEN — By day, Thomas Near studies the evolution of fish, wading through streams in Kentucky and Mississippi in search of new species. By night, Dr. Near, an assistant professor at Yale, is a heavy-duty gamer, steering tanks or playing football on his computer. This afternoon his two lives have come together.

On his laptop swims a strange fishlike creature, with a jaw that snaps sideways and skin the color of green sea glass. As Dr. Near taps the keyboard, it wiggles and twists its way through a busy virtual ocean. It tries to eat other creatures and turns its quills toward predators that would make it a meal.

The chairman of Dr. Near’s department, Richard Prum, watches him play and worries about his reckless lunges.

“You’re just attacking them?” he asks as Dr. Near tries to eat a fat purple worm that looks too dangerous to bother.

“If you kill them, you unlock their parts,” Dr. Near explains. But then the purple worm sticks its syringelike mouth into Dr. Near’s beast and begins to drain its innards. “Uh-oh, I’m about to die,” he says. The screen fades to black.

The next time, Dr. Near’s luck changes. He gains enough points to move to the next level of the game. His creature grows a brain. “Oh man, it’s like I graduated college,” he says. Dr. Near can now alter his creature. He stretches the body to give it a neck. He adds a pair of kangaroolike legs.

His creature — or, rather, a swarm of his creatures — charge out of the ocean and onto land. Dr. Near pushes back the laptop as his creatures find a place to make their nest and lay eggs. “So that’s pretty cool,” he says with a grin not often seen on a professor.

Dr. Near and Dr. Prum have spent a few evenings testing out Spore, one of the most eagerly anticipated video games in the history of the industry. After years of rumors, the game goes on sale Friday. Spore’s designer, Will Wright, is best known for creating a game called the Sims in 2000. That game, which let players run the lives of a virtual family, has sold 100 million copies. It is the best-selling video game franchise of all time — an impressive achievement in an $18-billion-a-year industry that is now bigger than Hollywood.

Spore, produced by Electronic Arts, promises much more than the day-to-day adventures of simulated people. It starts with single-cell microbes and follows them through their evolution into intelligent multicellular creatures that can build civilizations, colonize the galaxy and populate new planets.

Unlike the typical shoot-them-till-they’re-all-dead video game, Spore was strongly influenced by science, and in particular by evolutionary biology. Mr. Wright will appear in a documentary next Tuesday on the National Geographic Channel, sharing his new game with leading evolutionary biologists and talking with them about the evolution of complex life.

Evolutionary biologists like Dr. Near and Dr. Prum, who have had a chance to try the game, like it a great deal. But they also have some serious reservations. The step-by-step process by which Spore’s creatures change does not have much to do with real evolution. “The mechanism is severely messed up,” Dr. Prum said.

Nevertheless, Dr. Prum admires the way Spore touches on some of the big questions that evolutionary biologists ask. What is the origin of complexity? How contingent is evolution on flukes and quirks? “If it compels people to ask these questions, that would be great,” he said.

Evolution may seem impossible to capture in a computer. It is a hugely complicated process by which millions of individuals change over millions of years, as thousands of genes mutate and are spread by natural selection and other forces. Yet scientists have managed to distill some of the most important features of evolution into the language of mathematics.

In the early 1900s, mathematicians figured out how to represent a population of organisms in simple equations. They used those equations to show how natural selection can spread some genes from one generation to the next. Their work transformed the study of evolution into a modern, rigorous science.

Today, mathematicians use far more sophisticated equations to analyze evolution. And some of their most important insights have come from treating evolution like a giant game. Organisms can evolve different strategies to survive, in the same way game players can choose different strategies to win the most points in a game. Using a branch of mathematics called game theory, scientists can figure out if natural selection will favor a strategy over all others, or if it brings them into a stable balance. Game-theory models have shed light on the evolution of things like human cooperation and the deadly relationship of parasites and their hosts.

Today’s computers make it vastly easier for scientists to build these models. They have also allowed researchers to study evolution by building digital organisms. Scientists at Michigan State University and the California Institute of Technology, for example, have developed software called Avida that allows tiny computer programs to behave like real organisms. They make copies of themselves and mutate (randomly changing lines of programming code).

As the programs process more information in more powerful ways, the mutations are favored by a digital version of natural selection. The Avida team has published a string of papers in leading scientific journals on their experiments, testing ideas about complexity, mass extinctions and even the evolutionary benefits of sex.

Computers have also made it possible for scientists to build simple simulations to help people understand the principles of evolution. This year, for instance, Ralph Haygood, a postdoctoral researcher at Duke University, built a Facebook application called Evarium that lets users watch flowerlike creatures drift around a box, attracting one another with their colors. They mate and shuffle traits in their offspring, which then go through the same cycle. Players can control how quickly traits mutate and how strongly the organisms are attracted to some traits and not others. Or they can just watch the creatures change each time they open their Facebook page.

Mr. Wright came to the challenge of an evolution game with a long track record of simplifying complex systems without losing the feel of reality. He first came to fame in 1989 with SimCity, a game that allowed players to build and oversee a city. He simplified the workings of cities so that the slow personal computer of the late 1980s could simulate them. But he included enough feedback loops between elements of cities — like tax rates, incomes and traffic jams — to give SimCity the unpredictable complexity of real cities.

Mr. Wright followed the success of SimCity with a string of open-ended games, like SimAnt (a simulated ant colony) and SimMars (a simulated Red Planet players could make habitable). Around the time he released the Sims, he began to contemplate an all-encompassing game. At first, he called it SimEverything.

The game, which he eventually renamed Spore, would give players an experience of life and the universe across billions of years, from microscopic creatures to interstellar civilizations. “There were deep motivations in the early phase from the work of a lot of evolutionary biologists, like Richard Dawkins and Edward Wilson,” Mr. Wright said in a telephone interview.

Mr. Wright wanted Spore to communicate some of the grand patterns of evolution. But he did not want players to spend a million years waiting for something interesting to happen. He also did not want the game to look like an abstract cloud of drifting spots.

“I spent a fair amount of time going around to talk to scientists here and there,” Mr. Wright said. “You have to explore a huge amount to figure what 20 percent will be cool and fun for a game.”

One thing Mr. Wright and his colleagues decided Spore should reflect was evolution’s ability to produce life’s staggering diversity. “We wanted to convey the sense that evolution can bring up a surprising diversity of weird, interesting, strange things,” he said.

The game begins with a meteorite crashing into a planet, sowing its oceans with life and organic matter. Players control a simple creature that gobbles up bits of debris. They can choose to eat other creatures or eat vegetation or both. As the creature eats and grows, it gains DNA points, which the player can use to add parts like tails for swimming or spikes for defense. Once the creature has gotten big and complex enough, it is ready for the transition to land.

On land, the creatures can grow legs, wings and other new parts. And it is at this point that some of Spore’s features really shine. Mr. Wright’s team has written software that can rapidly transform creatures in an infinite number of ways, as players add parts and alter their size, shape and position.

This summer, as part of the buildup before the release of Spore, Electronic Arts offered software for building new creatures on its Web site. So far, people have built more than three million creatures. Electronic Arts uses that growing zoo to populate each player’s planets with life.

Neil Shubin, a paleontologist at the University of Chicago, was enchanted when Mr. Wright came to show off Spore to him. Dr. Shubin’s own research has helped reveal how real evolution recycles and modifies pre-existing biology to produce different body plans. In 2006 Dr. Shubin and his colleagues reported the discovery of a 370-million-year-old fossil called Tiktaalik that illuminates our ancestors’ transition from sea to land. It offers clues to how our hands and feet evolved from swimming fins.

Dr. Shubin found that Spore gave players a feel for how evolution uses the same basic tool kit to produce different body plans. “Playing the game,” he said, “you can’t help but feel amazed how, from a few simple rules and instructions, you can get a complex functioning world with bodies, behaviors and whole ecosystems.”

Spore also mimicked evolution in another way that pleased Dr. Shubin. “Will asked me, ‘Why did creatures evolve to walk on land?’ ” he recalled. “I mentioned that the freshwater ecosystems of the Late Devonian were pretty predator-intensive. He smirked.”

Mr. Wright built a Tiktaalik with Dr. Shubin’s help. “We let him swim around in a Spore Devonian world. And every time our little silicon Tiktaalik went in the deep water, a huge creature ate him in one bite. Tiktaalik crawled on land and thrived,” Dr. Shubin said.

Spore embodies another major theme of evolutionary biology: evolution is not a simple kill-or-be-killed affair. If a Spore player ends up with a carnivorous creature, it will certainly do its fair share of killing. But it will not make it very far unless it makes alliances. In Spore, creatures bond by dancing, wiggling and singing. Taking the time to bond allows players to move in packs and herds, which do a better job of fighting off predators and attacking prey.

“You always wonder why life tends to become more complex over time,” Mr. Wright said. “If you look at this balance between cooperation and competition, at almost every level it explains it neatly. You have agents competing at some level. The agents might be cells. At some point the cells can group together and work collectively and outcompete the other ones that are not cooperating. Then competition jumps to the next level. At every level you have to have the right balance between co-op and comp. That balance is driving the organizational complexity.”

Even as scientists praise Spore, they voice concerns about how the game does not match evolution. In the real world, new traits evolve as mutations arise and spread gradually through entire populations. Winning Spore’s DNA points does not work even as a remote metaphor.

“I do hope that it doesn’t confuse people as to what evolution is all about,” said Charles Ofria, a computer scientist at Michigan State University and a creator of Avida.

Spore may also mislead players with the way it is set up as a one-dimensional march of progress from single-cell life to intelligence. Evolution is more like a tree than a line, with species branching in millions of directions. Sometimes species become more complex, and sometimes they become less so. And sometimes they do not change at all. “There’s no progressive arrow that dominates nature,” Dr. Prum said.

These caveats notwithstanding, Dr. Near hopes that Spore prompts people to think about the evolutionary process. “This may be totally off about how evolution works, but I’d much rather be dealing with a student who says, ‘O.K., I have no problem with evolution; I think about it the same way I think about gravity.’ If it does that, it’ll be great.”

Mr. Wright said he had been hearing similar reactions from other scientists. “I find that scientists are incredibly open and excited that we can portray this stuff in games, even if it’s not perfectly accurate,” he said. “It’s manure to seed future scientists.”

Dr. Shubin said: “The differences between Spore and nature do not bother me. I see Spore for what it is: a game. And it is a game in the best sense of the word. It is not identical to nature, but it is a world that evolves, that changes and where the players are part of those processes.”