Unloading would save the AF447 but it is far from being only possible solution. Other ones: not pulling in the first place, not disregarding the stall warning, returning to assigned altitude instead of merely reducing the RoC from humongous to merely excessive, etc...
Not pulling in the first place would have prevented the the whole event, as would have functioning pitot tubes. It is self explanatory.
Returning to the assigned altitude, there starts the problem, if you refer it to the part after initial pullup. This returniung to the altitude was imho exactly what the PF tried, smooth and easy and no coffee cups on the floor and no passenger complaint. But he f**d up badly in the pullup and a normal return to the assigned altitude like in day to day flying with easy initiation and all doing nicely worked not out as planned. The speed decayed unobserved due to failed pitots and unnoticed due to lack of common sense concerning aerodynamics and energy management near the service ceiling of the aircraft.
The necessary recovery action was time critical.
In this situation returning to the assigned altitude has to observe those specifics and the primary focus has to be angle of attack reduction and speed conservation. The normal parameters for changing altitude or regaining altitude like a special pitch change, a special change of vertical speed, a known SS input might not work in the time available.
I referenced the amount of change to the normal load factor value, as that one is documented in the FDR. I´m well aware that
afaik neither AOA nor normal loadfactor are readily displayed in the cockpit.
Unloading is the most effective on highly aerobatic, very powerful and draggy designs, such as Su-26 or Su-27.
Unloading can be performed in any aircraft and is done during any approach to stall recovery. On which aircraft it is most effective is of no relevance to the discussion here.
My post has nothing to do with a special maneuver, what i´m saying is that an early recovery atempt before SW2 was made, documented by the normal load factor in the FDR, computed by HN to be +.87 in the average, but that it was not agressive enough.