(EDIT: This is incorrect, sorry. The article is poorly worded. Would you rather I delete this comment?)
(EDIT: This is incorrect, sorry. The article is poorly worded. Would you rather I delete this comment?)
IIRC, they actually had some very low level signal from the recorders two years ago that they didn't actually take notice of, and later, working on this data they finally extrapolated the probable position of the wreck after a tremendous data mining work.
That reminds me of how Gauss was able to use very limited information to locate Ceres after it had moved behind the sun.
http://www.businessweek.com/news/2010-05-06/air-france-crash...
http://www.nytimes.com/2010/05/07/world/europe/07crash.html?...
Sadly, there are no references to source material, and a quick search through the BEA at http://www.bea.aero/ (the French aviation accident investigation agency) didn't provide immediate satisfaction, alas.