As I understand it, the whole thing is distrssingly straightforward.
The certification requirements (worldwide, or at least as much of the certification world as I've seen, which is pretty much all of it except for the FSU) are straightforward. The rudder, fin and supporting structure must be designed for full rudder deflection at Va and 1/3 deflection at Vne. What this requirement has never included has been any sideslip component.
So we've always interpreted that requirement as rudder deflection from balanced flight. I don't think anybody has included in their assumptions that the aircraft might be already fully yawed the other way. If it is, then the force applied by the rudder, and associated stresses on the fin are much greater.
On a light aircraft, which is what most abrupt use of controls training is done on this doesn't really matter. Firstly because the aircraft has so much less yaw inertia to react the rudder force so it tends to swing with rudder input almost instantly, whereas in an airliner there's an appreciable time delay. Secondly because the calculations tend to be less accurate and also weight is slightly less critical, stress engineers use larger safety factors. So, we have been doing this sort of rudder reversal in light aircraft for at-least half a century without incident.
But in an airliner, with smaller structural safety factors and huge inertia reacting the rudder input you can get far greater stresses in the fin supporting structure than it was ever designed for. And it comes off.
I imagine that we'll be seeing changes to both pilot training / operating data, and to airliner certification requirements with regard to static strength of the vertical stabiliser before long.
G