Looks like more data is coming to light.
http://www.nytimes.com/2013/01/30/bu...anted=all&_r=0
The various regulations on the design of aviation batteries, posted elsewhere, seem clear to me and what I would expect, so the problem seems to be something fundamental in the battery type and behaviour over the full temperature operating range; or the behaviour of the battery, charger, and various loads as a combined system over the full temperature operating range.
Earlier there was a mention of Testability which in simple terms is the designed means to test the full parameters of a system and simulate the operational environment in which the system will live.
I'd be interested to know the range of operating temperatures within the EE bay and APU battery location from 37,000 ft ASL to sitting on the ground, the external ambient temperatures, and anything that could influence those temperatures in the EE bay/APU battery locations.
I would then want to know when loads are placed on the battery and the timings of that in the knowledge of the changing temperatures from ground to flight levels and vice versa. The batteries that haven't failed yet might have some of that data.
Presumably there is a test rig for this system that does some of this.
I'm sure that professional organisations working on this problem are equally interested.