Thanks for all theses information!
I will be interested to know what kind of "hardware" used to block theses issues, diodes or just a capacity to control the discharge?
I haven't quite worked out what is causing the remaining resizing issues. One problem is that the write signal to the RAM chip is 'active low' which means that if the RAM chip sees a low voltage then it will write data into it's storage. This is an issue as the write line is driven from externally and when the card is removed the external signals go low. This was causing the odd write when the pack was removed or inserted as a write signal would be seen while the signals from the organiser sorted themselves out. To help with this problem I added a transistor that is powered from the battery and holds the signal high even when the pack is removed.
In a similar way, the RAM chip has a chip enable that has to be high to enable the IC. I wire this to the 5V supply so when the pack is removed the 5V supply is removed and the RAM chip disables itself. That wasn't enough to prevent some writes, which is why I had to add the transistor circuit I described earlier.