_bpgl

...

One thing that's slightly interesting about Dyson Sphere's is that spherical shells of matter exert exactly no gravitational pull at any point in their interiors. That is, the integral of gravitational pull over the whole shell is identically 0 for every point inside the shell.* This goes for any force that falls off by the inverse square rule, e.g. the electrostatic force.

In short, there's no practical way to stick an atmosphere, much less a civilization composed of life forms that depend on gravity, to the inside of the sphere, even if the other engineering challenges could be overcome.

* One thing that's suddenly more interesting to me now is if the space at any point is being "pulled" equally in all directions, resulting in zero force, or if the bends in space-time cancel out. In the limit, for example, could you "tear" space-time inside a dense enough and/or heavy enough shell of matter?