One thing that I find interesting - shocking really - is that from the pictures of the intact battery, there appears to be no attempt to provide a thermal barrier between the individual cells. This would seem to virtually guarantee that if ones goes, they all go - kind of like storing eight Roman candles right next to each other. If there was some sort of thermal barrier that could limit heat spread while the shorted cell uses up its energy and burns itself out, perhaps the other cells would remain intact thus greatly lessening the total damage. In the old days, a metal/asbestos laminate often worked quite nicely, but I'm sure more modern materials are available.
Another interesting observation is that there appears to be no direct cell temperature monitoring (to be fair, it may not be visible from the top). All I can see is what appear to be thermistors attached to each of the jumper busses. The thermal lag time from this arrangement would make it essentially useless for detecting cell thermal abnormalities (but at leas the totally non-flammable copper buss will never overheat!). Even my lowly Makita Li-ion drill has a thermistor directly attached to the cell casings.