Build 5 uploaded. Ullage burn is only 20 seconds, no dependence on fuel quantity though.
Major changes to the DAP. First, it is rate-based, with arrow keys to set required rates in pitch and yaw, and Insert/Delete keys (NOT numpad!) to set roll rate. Second, it _should_ work with a stack of several vessels docked (shameless self-citation, code nicked from Precession MFD). Third, you can change DAP gains after pressing Ctrl-Y.
Fourth, pitch yaw roll channels for the DAP are individually selectable (Ctrl 1 2 3 respectively to turn them off and on).
Fifth, DAP works with RCS for pitch and yaw when engines are off, and TVC for pitch and yaw when on. Roll control is always (for now at least) provided by RCS.
EDIT:
Possible next major steps:
1. Ramp up and shutdown the three NTR sisters separately, with gimballing and possibly RCS nulling out residual torque.
2. Back-of-the-envelope qualitative FMEA (failure mode and effect analysis). What happens if a control drum pair is stuck? What's the probability of that occurring? How should the user designate abort boundaries during powered flight? Shall we consider a stuck gimbal as a possible failure? What is the worst-case scenario if a turbopump blows up? Will the fragments kill two other engines? (to be honest, I was considering adding a kevlar screen between the turbopumps on adjacent engines
)
3. Electrical systems, which should be easier on this spacecraft than anywhere else (no crew, no ECLSS).