Sounds like a cool topic to me Dragonblade.
I prefer a more traditional vision of the role of character level in a world. I've found it to work quite well to assume a "ratio" approach (for example):
8:10 population are commoners
1:10 population are classed levels
1:10 population are classed npc's
1:4 of classed npc's are militia
1:2 of militia are Warrior 1
1:10 of militia are Warrior 5
1:20 of militia are Warrior 10
1:100 of militia are Warrior 15
1:1000 of militia are Warrior 20
1:10,000 of militia are Epic Fighters
This gives a good feel of "believeability" while still allowing for plenty of "major players" in the world. If you assume a typical continent with 2mm to 200mm people, the numbers above get pretty big (ie, 1:250,000 pop = epic fighters, therefore picture that the US would have roughly 8,000 epic fighters). That's plenty to deal with an "epic" party, while still very-much presenting a world where "most are commoners" and "dealing with epic/high-level npc's is pretty rare...until you get there yourself.
Changes I'd love to see:
Better balance for the spells and save dc's in epic.
A good epic spell system.
Better balance for other feat/magic item/class combinations.
More uniqueness in the classes and class or prestige class options.
My characters seemed to gravitate toward "racial alterations" and "templates" as epic alternatives (that balance well).
A terrific system for managing a network or organization.
A simple, tactically challenging method for holding large-scale battles.
A method for managing a kingdom (from one castle to an entrie sprawling continent), with a scalable system for determining actions.
A more simple method for good tactical combat at high levels. Combat shouldn't have to take 2 hours.