I have just watched the National Geographic programme about this accident. It gives a fair (if incomplete) summary of the events and the investigation, and correctly (in my view) credits the crew for their judgement in handling the ditching.
Reflecting on my recollection of the investigation report, about which I posted
earlier in this thread, I have realised there is a spectacular fallacy in one of the key recommendations - namely that the FQIs for the ATR42 and ATR72 types should be redesigned with different connectors, so they cannot be mistakenly interchanged. The entire premise of the recommendation is that different connectors would ensure that the error of installing the 'wrong' FQI cannot be repeated. That is false. The FQIs rely on internal logic to provided a fuel quantity indication that is based on data provided by a number of sensors in the tanks. To prove this is false, simply ask this question -
suppose that type 42 and type 72 FQIs did have different connectors but that a 'type 72' FQI had been mistakenly programmed at manufacture with 'type 42' logic and this was fitted to the accident flight. All of the subsequent errors and the resulting accident would have occurred in exactly the same way!
A post-installation cross-check (e.g. drip-stick measurement) to validate the accuracy of the new FQI WOULD have trapped such an error (and makes the argument for different connectors redundant).
Such a fault would also have been trapped by the diligent application of the simple cross-check against refuelling records. This procedure is based on the simple equation:
Reading 1 + fuel loaded - fuel used = reading 2.
Fuel used is measured independently of fuel reading and this procedure, if applied, is virtually infallible. For it to fail in a way that allows a fuel quantity indication to be materially erroneous (and go un-noticed) requires that two independent fuel measuring systems fail not only simultaneouly but in a way that gives equal and opposite measures of error. The maths for calculating the probability of that happening are beyond me, but I would imagine it would meet the 10^-9 test.
A better recommendation would have been to emphasise the critical importance of the cross-check prcedure in assuring sufficient fuel quantity for any planned flight.