My non-expert take on this, to get the ball rolling, is as follows:
Normally as a battery charges up, its internal voltage increases. If the charging voltage is set to the correct value this means that the charging current falls to near-zero when its fully charged. Conversely the charging current will be high when the battery is in a low state of charge. In other words the charging current is a function of the difference between the internal voltage and the charging voltage.
When the battery is charging, a chemical reaction occurs which stores the energy being put into the battery by the charging current - the battery only heats up a bit due to resistive losses within it. Once the battery is fully charged, any remaining charging current just goes to making heat in the battery (and decomposing - aka "boiling" the electrolyte.)
However a feature of NiCads is that the internal voltage also changes with temperature. Unfortunately it decreases with increasing temperature so that a hot battery will take more current than a cold one for a given state of charge and charging voltage.
This means that a thermal runaway state can occur when, for some reason, the charging current is too high with the battery fully charged - the excess current causes temperature rise which decreases the internal voltage of battery, which increases the charging current which decreases the internal voltage etc - you have thermal runaway.
I think this would be likely to occur either because the charging voltage is incorrect, or because some fault in the battery, such as a shorted cell, has reduced the fully-charged internal voltage.
My take is that as soon as you remove the charging current the process stops. The only exception to that would be if you have left it so long that the battery has overheated to the point of distortion and possible short circuit on the cells.
But it will take a long time for the heat to escape and the temperature to return to normal, so you can expect the light to remain on for a long time (in my one experience it was 45 minutes). The other hazard is that the decomposing electrolyte will produce explosive gases (oxygen and hydrogen like a lead acid battery I believe?) so if they are not properly vented, or if a spark occurs it could go boom anyway, even though you have removed the charging current.
When I did a groundschool course at another operator, we were told that slowly discharging the battery helped to reverse the process and actually helped to cool the battery faster than isolating it completely, and this was the reason why only the battery positive contactor was switched off (there being some services that were powered direct from the battery positive terminal). However I am not sure that I believe this and it might have been purely so that the direct battery services remained powered. (Super Puma family)
Well that's my offering, who is going to start the argument?
EC