Over in Second Life, a user has created an emergent artificial life simulation that grows plants, rains, and so on.
This sort of thing is, to my mind, the future of dynamic world environments; in order to get away from static worlds, we have to have plausible reasons why things change. Natural resources and natural shifts in them offer plausible reasons for AIs to behave differently over time. Add in users affecting abundance or scarcity, and you get systems with changing dynamics. If it doesn’t spin out of control, that is. But you can curb that with balancing mechanisms.
Going to a simulation level also allows players to interact with the world and actually affect it.
Ironically, based on what I’ve seen and done, I’d guess that it’s far cheaper to make a world this way than the traditional way.
Given that Kotaku just reblogged this, we may get invaded. Therefore, I should mention that there are graphical virtual world projects that have done similar stuff on the world level.
SL’s clouds are a 3d cellular automata system already, else this couldn’t have been done. There was also a nice fish flocking algorithm demonstrated within SL a while back.
Was it Dark & Light that had ice forming over fjords in the winter? I remember seeing that at a demo a few years ago at GDC (they just released, but I haven’t checked them out yet…).
And of course, UO’s original ecology system was based on artificial life principles, where objects had abstract properties and therefore could theoretically interact with each other on the basis of things like “I, a dragon, need food, but the temperature killed the plants off, so the deer moved to hunt other plants, so I will have to range farther too, and the first meat I will find will be people.”
I have commented on this before, so it’s not like the whole dyanmic world via simulation is new. On MUD-Dev those of us who favored it were termed “simulationists.”