I scanned this thread with interest as I was about to start my own thread on a very similar topic. I have several Celestron SCTs, that I've had varying success doing imaging with due to mirror flop. My latest is an 11" Edge HD CPC, which while advertised as an ideal imaging system, clearly illustrates (on multiple fronts) that Celestron has not given much thought to fully automated imaging setups, such as the one I'm trying to bring online in my newly finished observatory. While the Edge HD does have the mirror clutches that should allow you to lock the mirror into position once focused, that doesn't do me any good for my remote control focuser, since I can't leave the locks clamped down (although I think I'm probably going to try, given I shouldn't need a large amount of motion for tweaking the focus). At any rate, I'd have to agree that there's not really much excuse for the sloppy design of the SCT focusers these days. I understand how the SCT started with a simple slip-tube design, but high quality linear ball bearings aren't exactly expensive these days. There's no reason that the mirror flop issue couldn't be addressed especially for something like the larger Edge HD and ACF scopes. The other possibility I've thought of, which would potentially be retrofit-able to my Edge would be to use a fine chain drive to tie three lead screws together rather than just using the single push/pull point of the current focuser. Essentially you'd replace the two clutches with identical drive screws and tie all three focuser screws together. That would force the mirror to track true and should eliminate the flop. This would be pretty clunky as an after-market mod, but if built into the scope design, the connecting linkages (e.g. chain) could be inside the housing. You could also use a worm gear drive shaft mechanism to tie the drives together.
I can think of a few other ways to accomplish a central adjustment scheme, including using a threaded tube for driving the mirror. That would give a center drive point and require tighter tolerances to begin with. The requires some keying to ensure the mirror doesn't rotate, so we're back to some sort of linear bearing anyway. Another similar concept would be a twist plate behind the mirror that would move a set of lever arms or even scissor lift type mechanisms around the perimeter. The more pivot points, however, the more chance of something being loose. The neat thing about this would be that it would only require a single push point to drive the plate, although it would have to be from the side rather than back. None of these are perfect, but just trying to think outside the box (or tube in this case)!