Take a step back...
I think it is time to take a step back to examine the need for real-time, full-time data transmission (including CVR data).
In recent accident history, (last 30 years, over which a tremendous reduction in fatal accidents has been the trend, due to CRM techniques, GPWS/EGPWS, TCAS & automation), there has been only one loss-at-sea in which the recorders were not found and which so far remains a mystery and that is MH370 and we don't know if that "loss" was in some way not accidental.
Through great efforts, the flight data recorders of every other transport accident at sea and on land have been found and the causes discerned, and lessons learned. That is a statistically significant success rate. The locating of and the successful retrieval of the recorders from the AF447 is the benchmark against which all subsquent retrievals must be measured. It can be done.
Add to this fact the statistical reality that most aviation accidents occur within an hour of the departure or destination airfield, over land, or over water which is more often than not accessible; QZ8501 is such an example; there are numerous others.
Counterexamples to the above are enroute accidents, (mainly CFIT, LOC, mid-air collision), in which the accident site is in remote territory that is either ice/snow, mountainous or heavy jungle. Yet, to my recollection, there are no accidents, MH370 perhaps excepted, in which the recorders have not been located and retrieved.
The calls for change are understandable but in my opinion are misplaced, sometimes even on the part of those who really do know what they are talking about in terms of the technical solutions to such a proposal.
The data-loss and unsolved-accident rate does not support such solutions.
Not knowing where one's fleet is at any one moment is, however, unacceptable and the many solutions are already at hand and relatively speaking, inexpensive. "SkyTrac" is one...there are many others.