[iamthad]

Although this approach is obvious, there are complicating factors. The Coriolis force makes high-RPM centrifugal gravity unpleasant and perhaps dangerous to humans. So, to achieve 1g, structures would need to be very large (224 meters). Nonetheless, this is probably the best approach.

http://en.wikipedia.org/wiki/Artificial_gravity#Rotation

[danielweber]

Two modules separated by a tether requires much less mass. It makes Coriolis effects hard to notice.

I readily admit they haven't done any of this. The lack of engineering work into making spinning structures is criminal.

[drakeandrews]

In structures with a small radius, the "gravity" differential between your feet and your head is a bigger issue than the Coriolis effect. Large structures using the tether approach produce sufficiently large forces on the tether that it becomes a significant engineering problem. Tether structures also suffer the issue that they are much harder to stop and start than a ring structure. This is important because you do not want the structure spinning whilst you make course adjustments. There are modifications to the design that compensate for this, but they add significant complexity to the design of the ballast.

The ESA have put a lot of thought into this, as have Rosmoscos. I assume NASA have put a fair amount of thought into it as well, but I've nothing to back that up.