more hindsight
I could add the following to my previous post on what influenced the search strategy.
The retro drift analysis by the Brazilian Navy and USCG, pinpointing the impact so close to LKP and to the actual impact site, didn't get the attention that it deserved. Because there were other analyses, mainly by Meteo France (based on other models) that gave hugely divergent results.
And also the item that was found on 7 June on the surface of the impact spot was not seen other than an anomaly, while (as mm43 said before) it was probably a floatable piece recently escaped from the wreckage on the ocean-floor.
Moreover, because ULBs normally don't fail the search was very much focussed on that: until 10 July (the 30 days certified + 10 days of expected but not certified days of functioning) nobody was expecting not to find any pinger signals.
It was not after all the vast areas covered by sidescan sonar had not rendered any result either, that based they came closer to LKP again. On the one hand this was determined by what was left "un-sidescanned" and on the other hand on the recent the Russian study that showed how close to the initial upset point A/C that experienced LOC in cruise ended up at 0 ft.
But in one thing BEA has always been adamant: "AF447 has never been south of LKP" (except in cruise of course).
Looking fwd to the read-outs of the recorders......