Inmarsat's comments, specifically the Chief of Flight Safety's, seem a tad naïve. I wonder if they really know about flight data analysis, the investigative process and flight safety work.
Privacy, confidentiality and assurances of data security seem to have been brushed aside in favour of "exciting new technologies that can prevent crashes". I've seen/heard this before. So far, this is just so much marketing hype.
In terms of tracking & locating technology, it has existed for some time - we have been using it to follow our aircraft locally and around the world. It is real-time interrogation using web-based interfaces with user-selected 30" to 10' polling periods. I have no idea why tracking is such an issue; we've been doing it for at least six years now; we know where our aircraft are all the time. And, for a price, it can do parameter/event monitoring say, for high-acceleration events for example, so that maintenance can meet the aircraft based upon data not crew reports, etc. It isn't an entire DFDR process but it could be sufficient for initial explanations and certainly LKPs.
In terms of retro-fitting for deployable recorders and/or streaming data, I would argue against the concept, and against an industry-wide, regulatory requirement.
All recorders in recent over-water accidents have been recovered and read with perhaps one exception, the Asiana B744 freighter loss over the Korean Straits.
Acknowledging that, like all human activities, aviation works on risk probabilities, the design and engineering manufacturer's groups of the industry work on the basis of what an acceptable failure rate of mission-critical elements/components is; the certification standard is, as some here will know already, 10^-9.
We can reasonably consider that the loss of MH370 is such an event to which the same standard may be applied, and that the "normal" pattern is the historical one in which in all cases, above exception noted, the recorders have been recovered and accidents understood.
The argument for deployable recorders & datastreaming is essentially an economic one only, and that is a different arena than the case for flight safety. Such standards (for flight safety), are borne by the industry and ICAO member countries when/where accidents occur, and yes, it is expensive but also extremely rare.
I think the case for deployable recorders/data-streaming has not been demonstrated against this accepted standard. I think what we are seeing is a bandwagon response, not flight safety work.
If we are to invest significant funds and the ensuing subsequent certification/regulatory work at all in changing the way aircraft recordings are done, it would be reasonable to focus on power sources for both the data and voice recorders.
Public cries for instant sources of data do not demonstrate the flight safety case for such capability, they demonstrate a willingness to engage in media and political arugments. But with recent information regarding MR804, we may again be faced with a loss of power to the recorders and loss of data which inhibits understanding of the accident.
Ensuring that recorders have uninterruptible sources of power is a flight safety case, and as such is demonstrable.
Factors which are necessary to consider are not limited to electrical system auto-responses to load-shedding and/or damage to wiring through fire or mechanical processes etc., but also loss of electrical power through the intentional use of emergency electrical configuration drills which are intended to remove all power from normal AC & DC busses during smoke of unknown origin emergency drills.