Thanks to Hazelnuts recently passing a link to the engauge digitizing software, I have finally been able to perform roll behavior analysis on the initial movements of AF447. Admitted that this is the 11th hour and 59th minute before BEA releases their report, but the evidence is very strong that the crew encountered a roll PIO that took their full attention away from the proper control of their aircraft and concentrated their attention on the roll problem to the detriment of the big picture.
I was able to digitize the roll angle to a high degree of accuracy and then calculate the roll rate on a short interval basis. The roll rate data was then smoothed by a moving average technique. The result is as follows:
From this, it can be seen that the peak roll rates exceeded 10 degrees per second. This roll rate in an oscillation would not be ignored by a pilot and he would be unable to "look through" it to the larger picture.
We had a substantial discussion at the time I put forth the possibility of a roll PIO and rather than re-post that discussion, I will merely provide a link to that portion of thread number 7.
http://www.pprune.org/tech-log/46839...no-7-a-35.html
There are ~5 different potential causes for PIO. BEA has not yet provided spoiler and aileron position data, so analysis of the causal aspect is not yet possible.