the whole situation just reeks of un-revealed things.
according to Malaysian reports, the autopilot was reprogrammed 12 (?) minutes prior to the last voice contact - the famous "Goodnight" thing.
per Malaysian radars the plane turned west, enroute to an established "waypoint" - gobblespeak for a point in an air traffic corridor route.
then proceeded to a second waypoint and looped around that; before turning west again and flying down the west side of the island chain into "nowhere" aka southern Indian Ocean.
note that this 'down the west side' route does not agree with the arc of the engine system ping data in the early hours of the flight; only much later do the two paths converge.
the plane could not do that except under human control - which human...? that's a question.
catastrophic fire disabled the electrical system(s) - okay, how'd the autopilot continue working?
it's a fly by wire system; one would certain hope the designers "protected" the electrical supply to flight control surfaces from any kind of disruption. batteries? how long do those batteries last?
they have a magnetic compass, flying off into never never land for lack of a sense of direction.... not too likely. tossing aside all 'standard procedures' where the pilot/co-pilot would have immediately notified somebody on the ground "Malaysia, we have a problem."
"soft landing, plane intact, sunk intact = no wreckage / debris" - so many problems there. in a soft landing the surviving passengers would have evacuated aka the 'Miracle on the Hudson' and something would have been found, by now.
did the passengers not notice they were supposed to land hours ago? hence the theory the passengers and crew were "disabled" - smoke, toxic fumes, depressurization, pick a theory - major holes in all of them. problem with that is a disabled cockpit crew could not hop from waypoint to waypoint then later set a course into never-never land.
disabled passengers and crew + soft landing - not going work - the autopilot system seeks to keep the aircraft on a course, at the specified altitude and speed. autopilots are probably (because I'm not really an expert in autopilots) not programmed to revert to "when all else fails, do a soft landing" - more likely when the fuel was exhausted the autopilot would blindly attempt to control course, altitude and airspeed - eventually pulling the nose up to the point the aircraft stalled and hit the water in a violent spin. problem: no wreckage / debris.
a 'slow leak depressurization'? nope. cabin pressure is maintained by compressed air from the engines - which continued to work. 'fresh' air is continuously supplied under pressure and 'stale' air vented out the back of the plane. small "holes" cannot cause a complete / debilitating depressurization, and the Hollywood version of a bullet blowing out a window and everything in the cabin gets sucked out is just that - Hollywood, not reality. bullet hole(s) would put a small diameter hole in the window / fuselage and not much more would come of it other than less stale air being exhausted from the cabin.
"the pings" - hopefully this will lead searchers to a specific location. there is a problem: sound travels through water in mysterious and sometimes unpredictable ways / paths. the biggie in the issue is "thermoclines" - layers of water at different temperatures which "refract aka bend / transmit" sound somewhat unpredictably. it's going to take multiple "got it" events to pin point things; and then there's the question of how long the devices may continue to ping....
given the wind and currents in that area of ocean - and the elapsed time, surface wreckage could be hundreds of miles from where the aircraft pingers settled on the bottom. it will be really a bad scene to find dead survivors in life rafts after a soft landing - dead because nobody knew where to start looking.
at the moment, there's precious little other than intentional - and skilled - action which explains the totality of the circumstances.
why? that is a valid question.