I have to agree with those that question the ergonomics of the rudder control system. I think the reference to the 767 system was as an example of what the 'logical' or expected feel of the input/response should be during progressive rudder travel limitations.
Reduced pedal travel as the rudder deflection is limited is one thing, but the varying pedal forces during limitation appears to have no logical relation.
Beyond this, there is the lag between input and deflection, which can be clearly seen in the NTSB graphs. There are moments when the pedals are pushed one way while the rudder is still deflected the other way before it reacts. While this may be insignificant during typically smooth control inputs, this momentary lag will have significant effects on the pilot's feel for how the aircraft is responding to commands, even inspiring the pilot to make quicker and larger inputs, especially during relatively rapid reversals.
I'm not too sure how this lag compares to other large commercial aircraft, but if compared to a direct link system (cables/rods with no augmentation), it would appear to me to be a factor. I certainly would not be happy for any of my R/C planes to have such a input/response lag.
Another point is that the DFDR data collection periods for accelerations and other pertinent parameters are too far apart to have picked up the all the 'details' of the alleged wake turbulence, making it impossible to back-drive a simulater and experience exactly what Sten Molin felt the aircraft was doing.
Making any finite conclusions on whether he was right or wrong to make such input commands based on such incomplete evidence is just not right IMO.
BTW, it wasn't a 'fatigue failure'. The forces exceeded the ultimate design load of the structure. The big issue here is that the FARs only require the structure to withstand forces from a full (inc limited) rudder deflection, centering the rudder during the subsequent yaw deflection, and centering the rudder during yaw deflection + overswing. It does NOT certify tail structures for a rudder reversal during yaw deflection or yaw deflection + overswing. One has to ask 'why not?', since the chances of a rudder correction being made during a sudden gust, or specifically just when a sudden gust stops, seem quite high to me. Add in a bit of control system lag, and the circumstances of an aircraft in sideslip with an opposing or reversed rudder deflection are even higher.