As I delve into the actual implementation I am struck with the conclusion that this is 100 times more complex than it seems on the surface. The level of interaction between the different sciences is what really is problematic. I just don't know nearly enough about planetology (like in Dune :P). This would be a huge research project.

Of course, you could just fake all the science. Most people wouldn't really know the difference anyway, but then I think it would lose a bit of its allure.

Would anyone with some knowledge in the sciences (geology, astronomy, biology, chemistry, linguistics, ecology, etc.) be interested in working with me to try and produce a kind of small design document or even an (extremely limited) working demo?

I consider myself extremely well rounded in most sciences but this goes over my head in more ways than one.

Quote:

Original post by Numsgil As I delve into the actual implementation I am struck with the conclusion that this is 100 times more complex than it seems on the surface. The level of interaction between the different sciences is what really is problematic. I just don't know nearly enough about planetology (like in Dune :P). This would be a huge research project. Of course, you could just fake all the science. Most people wouldn't really know the difference anyway, but then I think it would lose a bit of its allure. Would anyone with some knowledge in the sciences (geology, astronomy, biology, chemistry, linguistics, ecology, etc.) be interested in working with me to try and produce a kind of small design document or even an (extremely limited) working demo? I consider myself extremely well rounded in most sciences but this goes over my head in more ways than one.

I think that it doesn't really matter. Just fake it. If you want it "accurate", you'll need at least a cluster to get anywhere in a reasonable amount of time. Weather predictions are complex and not always accurate. Planetary simulations take longer and are much coarser grained than what you want. Just stare at a globe for a bit and see how the different climates are related to each other. Don't worry about the science behind it, just copy what happens at the boarders. The truth is, not a whole lot is known about what really happens or why the Earth is precisely how it is (Discussions about global warming really bring this to the forefront).

There are many areas where I think you'd jsut have to fake it, but there are other areas that are reasonably well understood that I don't think you should fake or you'd lose alot of the magic.

For instance, if you have a wind that moves to the west from the east carrying moisture off a large body of water and it encounters a large mountain range, the windward side will be lush while the downwind side will be desert.

I don't think you have to necessarily model the mechanics that cause this fact to be true, but putting this relationship in the world creates some more interesting and compelling results.

Likewise organizing the organisms in a food web pattern and adjusting the populations of different levels (by changing the luminosity of the parent star for instance) should give you appropriate relative abundancies of different organisms, which in turn would effect any human populations in the area in question.

SimEarth did alot of what I'd expect a program like this to do, and that was over a decade ago. Modern processing power should yield even better results, especially since it doesn't have to process this all in real time, and you don't need to model the whole planet to get worthwhile result. Worlds are relatively static from the time frame of a human.

In the Myst games you always explore relatively small islands. I think modelling the geo-biological interactions rather convincingly on a small scale wouldn't be terribly difficult for the computer. The main difficulty is in understanding teh relationships enough to tell the computer what to do.

A game like this would be equal parts simulation and artists' canvas. Shorting either would be a mistake I think.