http://en.wikipedia.org/wiki/Air_France_Flight_447

So, we're talking about more than once per decade.

There is some value in knowing early where an accident occurred, but that is rarely a problem. How much is it worth to fix it for the few cases that happen in the middle of nowhere?

"While Brazilian Navy authorities were able to remove the first major wreckage and two bodies from the sea within five days of the accident, initial investigation was hampered because the aircraft's black boxes were not recovered from the ocean floor until May 2011, nearly two years later"

If they knew where the plane went down, it wouldn't have taken 2 years to find the flight data recorders. The "pingers" on the recorders lasted for about 30 days, which should be plenty of time, if you know where to look.

If (extremely hypothetically) it turns out that this disaster could have been preventable if there had been some way for someone outside the aircraft to assume control, the tradeoff to be examined would be the benefits vs. hazards of extensive remote control capability, rather than some kind of commando raid techniques.

Can their be parameters using a known flight plan that prevent the computers from permitting actions that are contrary to the safe operation of the plane without notification of someone, if not possible prevention there of? (giving some leeway to have on board circumvention - two station bypass no adjacent/password/etc)

On the other hand, pilots need to be able to be in total control of the aircraft, including the ability to shut off any system on board. (See Swissair 111). In an emergency, you don't have time to mess about with two station bypass/etc.