My guess here would not be to have the entire interface equidistant from the user, but, rather, to have the equivalent of a wrap around monitor.
We're already used to having flat screens to interact with and simply curving the display doesn't, in my experience with a curved monitor, have any detrimental effect on your ability to view the display.
Beyond that, in a mathematical sense, a square and a cylinder are very topologically similar. Of note is that the relative size and shape of things on the surface are well preserved[1] as you don't need to significantly deform a square to turn it into a cylinder.
This is not the case for a sphere.
As a practical demonstration, take a piece of paper and draw some boxes on it before wrapping it around a reasonably sized ball. You'll see that the boxes (particularly the larger boxes) become deformed in a way that likely will make them seem to "bulge out". As far as I'm aware, the effect happens to any regular polygon wrapped around a sphere. The same thing will happen to any interface that attempts to project onto a sphere.
So, given that:
a) we're already used to flat interfaces
b) curving that interface into a semi-cylinder doesn't[2] have any detrimental effects
c) that traditional square interfaces don't map directly/cleanly onto a sphere
It's easy to conclude that a cylindrical interface is probably the simplest to make and easiest to understand interface. It requires the smallest amount of work to make and the least learning on the part of the user.
1: Obviously, things on the cylinder will be curved, but the 2 dimensional shape is still consistent.
We're already used to having flat screens to interact with and simply curving the display doesn't, in my experience with a curved monitor, have any detrimental effect on your ability to view the display.
Beyond that, in a mathematical sense, a square and a cylinder are very topologically similar. Of note is that the relative size and shape of things on the surface are well preserved[1] as you don't need to significantly deform a square to turn it into a cylinder.
This is not the case for a sphere.
As a practical demonstration, take a piece of paper and draw some boxes on it before wrapping it around a reasonably sized ball. You'll see that the boxes (particularly the larger boxes) become deformed in a way that likely will make them seem to "bulge out". As far as I'm aware, the effect happens to any regular polygon wrapped around a sphere. The same thing will happen to any interface that attempts to project onto a sphere.
So, given that:
a) we're already used to flat interfaces
b) curving that interface into a semi-cylinder doesn't[2] have any detrimental effects
c) that traditional square interfaces don't map directly/cleanly onto a sphere
It's easy to conclude that a cylindrical interface is probably the simplest to make and easiest to understand interface. It requires the smallest amount of work to make and the least learning on the part of the user.
1: Obviously, things on the cylinder will be curved, but the 2 dimensional shape is still consistent.
2: for me, at least