It seems the meteorite "theory" is picking up, even on Slashdot... And it is going even statistical:
IMHO, it is picking on the wrong way.
First, not all Earth's surface is equally probable on meteority strikes. Earth is moving on Space in a determined direction, and carries a specific positon in time. Now, space objects in the Solar System are not equally distributed all over it. Under this, as far as I can record, it seems that the Equator is one of the less probable zones for meteorite strike.
Under the weather conditions, we know were ocurring at the flight path, the flash being a "meteor stike" is even less plausible. Much more plausible is a massive bolt striking the plane. They also flash brightly and an example from a plane in Japan, seen on several sites, even shows the camera going blind.
Yes, this does not "explain" the "vertical fall" seen after the flash. Is it so? If the flash did blow up the fuel, then that would be the most probable thing to be seen.
Yes, many people claim that lightning cannot do such damage to airplanes. But no one denies, that a catasthrophic set of events could lead to something unexpected. Now such things do happen and when they happen, things go wrong in every possible way, specially, it occurs if models consider every catasthropic event as an individual parameter, but ignore their simultaneous ocurrence. In my life as an expert (nothing related to airplanes), I have seen some 4 such events. I even confess here that I was responsible for the consequences of one of them - I considered that events A and B would never occur together in any possible and plausible form. They ocurred. No matter that there were things that denied their simultaneous ocurrence, I never thought that one day I would have to deal with an hybrid environment were both basic events would be possible.