I think part of the problem with tail sitter deckplans is that its simply easier for someone to make a 2 or 3 deck belly sitter design that a 7 or 8 deck tail sitter of the same size, and the customers, in general prefer the big, lavish belly sitter decks that are easy to read and play on to smaller, less visually impressive tail sitter decks, that add extra bookkeeping for keeping track of who is on what level in a fight.
on topic, I think the engineering requirements of making everything rotate are not
insurmountable, but that its needlessly complex compared to other solutions (for example, having the ships software cut the M drive acceleration if the grav plates fail).
Also, designing the rooms so that they can rotate 90 degrees puts major limits on what you can place where, in terms of deconflicting the placing of furniture so that it can transition smoothly, and without power (since we are talking about a situation where major system has just failed, its not unreasonable to assume that several other systems might be offline as well, so no guarantee of room power to shift things for you, it would all need to be "gravity" driven), ship designers may just think its too much hassle for a limited payoff, and take other options. For example, always placing the beds in a stateroom against the aft wall, so that if power fails while the occupant is asleep, he only "falls" a foot or two, and designing lockers to hold closed while suspended face down. obviously anything not strapped in place is going to become a projectile, but thats a matter of discipline and housekeeping by the crew, so all designers can do is provide enough storage space in convenient locations that most items are secured while still being easy to access, and recommend certain standard operating procedures (like having the pilot be strapped into his seat at all times, so a gravity shift doesn't throw him away form the controls).
another factor is "total loss of gravity" isn't the only possible failure mode for the ships gravity. a software glitchy or faulty sensor might cause the system to, for example, shift perceived gravity 45* to the right, or onto the ceiling, or straight "forward" towards the direction of travel. all these would need to be factored into the design as well.
edit:
here and
here are some videos talking about the real life work done to train military vehicles crews to deal with vehicle roll-overs, which give you some idea of what a gravity shift or failure might be like. They were filmed at Camp Bastion, Afghanistan, and I've used that very machine for RODET training. they stick some foam shapes into the back of it to represent loose items (backpacks, unsecured rifles, etc), and they go
everywhere, with pretty much everyone in the back getting hit at least once in the face by something.
