There seems to be the assumption by some earlier posters that improved vibration monitoring techniques might have identified an impending failure and prevented this accident.
I submit that while vibration monitoring is a good technique for analysing rotating components, it is possible for static components such as a gearbox casing to fail without any vibrational warning.
Dye penetration tests and X-rays prove only that a component is servicable at the time the test was carried out. A slight misalignment or incorrectly torqued bolts during reassembly or maintenance work at a later date could cause stresses that initiate a crack that then propogates without any sign of problems until the component suddenly fails in overload.
This process of fatigue cracking was first identified during testing after crashes of early Comet aircraft. Repeated stress allows seemingly insignificant flaws to grow over a period of time until sudden failure. A common factor in all the Comet crashes was that they all occured without warning. A redesign fixed the problem, but public faith in the aircraft was lost.
Testing and monitoring programs are a lot more stringent now but it seems that two nearly identical rotor separations have happened, although the exact root cause in each case is still to be determined.
Better vibration analysis and HUMS systems are probably not going to be the answer if the parts that fail are not actually moving, just providing a support structure.