[chrispenycate]

The only problem with having the entire ship rotate that I can see is in turning corners, from the gyroscopic effect, and I'm sure that could be compensated for (no, not in "grab the steering wheel and pull us out of the line of that planetoid" space opera, but in "in eighteen months and two days our primary thrust axis needs to be 2.07° closer to Polaris").

I would hate to design an air-tight, frictionless bearing/seal which would last – oh, decades at least with the sort of mission lengths a craft like this could expect, and really difficult to bring to a halt for repairs. (Frictionless suggests magnetic bearings; I wonder if you could shape the field so it also held a ferrofluid layer in place solidly enough to resist not merely atmospheric pressure, but the extreme peaks of molecular displacement, and make a ferrofluid that resisted vacuum long enough?) Otherwise you could run a central shaft down the entire ship (structural and your main material conduits), thickened up where the outer skin is missing, maintaining rigidity. The bars leading from this central pillar to the rotating section might be a bit of an inconvenience to anyone trying to get along the ship at speed, but, if you build big enough, your carrousel only needs to turn at the same speed as a rotating restaurant; say, once every two hours. And that means your "holes that line up" biting your ankles off problem is a lot less than you've considered, too.

The main argument against propulsion gee forces is that you are expelling reaction mass continuously, rather than freewheel most of the trip (oh, and energy, too, but a decent energy source is essential to the project from the start) and, for more than about a hundredth of a g, an appreciable fraction of the vessel's total mass. Not easy to replace, then you spin the ship round mid voyage and go straight on to deccelerating just as hard…