After reading today that the Paris Court of Appeal has found Air France and Airbus guilty of involuntary manslaughter for AF447, I've been re-reading through information about the accident.

My layman's understanding is that a key tenet of Airbus Fly-By-Wire design is that flight computers will prevent the pilot from taking the aircraft outside of its safe envelope - including preventing stalls through the use of high-AoA protections (Alpha Prot / Alpha Floor / Alpha Max). I understand that the A330 switched into Alternate Law after the pitot tubes froze and lost reliable airspeed information, and Alternate Law kicks out high-AoA protection, essentially allowing the plane to fly like a "normal" jet that CAN stall.

Where I'm still fuzzy (please correct me if there is any misunderstanding):

1. Does Alternate Law removal of stall protection boil down to the computers essentially losing faith in their airspeed inputs - i.e. without trustworthy ADR information the protections simply cannot operate safely so the engineering decision is made to return the envelope to the pilots? Or is it more complicated than that?
2. AoA comes from separate vanes, not the pitots. Was there ever post-AF447 discussion about retaining some degraded version of AoA-based stall protection even if you have unreliable airspeed? Or are there sound engineering reasons that that would be a bad idea?
3. The stall warning reportedly activated \~75 times. It also stopped functioning at very low airspeeds because the system considered the airspeed data invalid. That (as I understand it) is why pulling back momentarily silenced the warning, but may have also strengthened an incorrect mental model in the cockpit. Was that known about at the time, and did it change in later software versions?

I realise this accident has been scrutinised for years by people far more qualified than I. I'm just trying to ascertain exactly where the Airbus envelope protection stops and where the crew is truly on their own.