Over the weekend, I produced a (mostly finished) game called Abandon for the Ludum Dare competition. I've learned a lot from this experience, and will be incorporating scripts, artwork, and concepts from my project into Against the Wall. For instance, large-scale buildings based on real-world architecture.
During the competition I realized that I wasn't experimenting enough with my designs. I was being too practical, thinking too logically when it came to the physical scale and structure of buildings. How would a dome or arch work structurally if it is sticking out of The Wall horizontally? Wouldn't a tower be useless if there is always higher ground? Where would people get the material for all of these buildings?
These questions are a bit similar to the one that I keep getting about how the trees stick out horizontally rather than face the Sun. One answer: Every tree uses the same model at various rotations, in addition to the Sun moving in a circle around the sky. I would have to create a large number of tree model variations AND animate them in order for the trees to face the Sun constantly. Another answer is that they are not really trees.
The gist of what I'm saying is, If I bind myself to a conventional understanding of physics, I may have fewer people questioning the game world but I also limit what I can do there. Remember the size of the facility in Portal 2? That game was set on Earth, and the physics and logistics of the whole complex make no practical sense whatsoever! The game was so cool that people suspended their disbelief and just enjoyed the ride. It would be presumptive of me to suggest that Against the Wall is as cool as Portal 2. However, the underlying premise of my game is as absurd as quantum tunneling devices and sentient robots.
I'll concentrate on making a strange, gigantic world that is fun for people to wander around, above all other considerations.