Well after a lot of reading I've finally found out a few more interesting things.
"CABIN VERTICAL SPEED" is triggered if the cabin air pressure changes with an excess rate of over 1800 fpm (either +/-) over 5 seconds - so a 150 ft cabin alt pressure change over 5 seconds will trigger this alert.
We can't be sure what the cabin alt pressure was set to, but due to the fact that this was a longer sector (documents state that under 2,5 hours it is typically set to 8000 ft, over 2,5 hours it's typically set to 7350 ft) it might have been 7350 ft.
One way to trigger that event is busting pressure difference limits (+8.85 psi, -0.26 psi iirc) due to a descent below an altitude where there pressure difference would have been in excess (in absolute terms) of -0.26 psi (I did some calculations and it would have been slightly below 5000 ft MSL) or severe air pressure drop in the environment. If the air pressure difference went above -0.26 psi they went from FL350 to 5000 ft in under 5 minutes, so their descent rate would have been in excess of 6000 fpm
. Another alternative would be a decompression, but the fact that none of the recovered passenger oxygen masks were deployed (found stowed) speaks rather against this.
Not completely sure how much pressure is lost with normal cruise outflow valve positions and the (completely normal) imperfect seals - but could it be possible that engines spooling down simply caused the rise of cabin alt pressure (so it went from 7350 ft to above 7500 ft within 5 seconds) - if that were true then that last message could also confirm the loss of engine power. Highly speculative (and very unlikely to say the least), but trying to think outside of the box.
I will also have to second guess my own statements that it could have been a ditching gone bad ... 0 flaps speaks against this (reread the BEA documents).
After rereading the BEA documents one specific passage caught my eye. The NAV TCAS fault logged via ACARS at 0210Z could have been cause either by an electric fault or the standard altitude being rejected by the TCAS system for not passing the credibility test (ie: what was read from ADR1 or ADR2) - while not knowing the exact procedure of the credibility test I would imagine that AoA and previous altitude would be used to predict the new value and allow for some tolerance. If they did get caught up in a massive updraft I could imagine the predicted value being too different from the read altitude. The other alternative is incorrect altitude readings received from the ADRs - which would quite likely mean that besides their speed indication being unreliable (apart from the GPS deduced values in the GPS monitor) also their altitude indication was unreliable. With that in mind - they were effectively blind and relying on the standby attitude indicator, the EPR/N1 indication to try to maintain altitude and speed, and the GPS speed + GPS alt values (which would be rather "laggy").
So after exploring all of this I only see 2 options of the reason for the disaster:
1) loss of control due to disorientation (hitting the water near wing level is somewhat odd for this speculation, but possible if they entered a sequence of stalls that they did manage to recover from - until the last one)
2) loss of engine power and subsequent ditching (0 flaps speaks very much against this, also all of the debris starting point location (path regression) also favor the first option)
D.