Here I am standing on the sidelines, watching a good discussion. It's actually reassuring to see that it's not only Vision units that have "issues".....
My Conclusions?
It looks to me like the reasons given for not having a "hard off" are being side stepped by the users, so that the solenoid is in the "Soft off" (OC) mode, which could put the user in the position of getting in the water with the solenoid disabled.
The brown out circuits should be designed to work properly so that the rebooting and firing the solenoid underwater can be avoided. The solenoid could be firing into a loop of unknown PPO2, and if shit happens to a battery, so that you need to fly manually, you really don't want the additional layer of shit caused by losing all your deco data. Isn't this what the BSAC used to refer to as "the incident pit" aka "a cluster f***"
Some type of battery monitoring should be fitted.
All in all, a five minute job to redesign the software/firmware/hardware.
And finally, is the inverse of a "Soft Off" a "Hard On"

Just a short follow-up since you quoted me, the Hammerhead solenoid will fire even in OC mode if the loop PO2 falls below .19. This is one of the saftey protocols built-in I feel this method is an acceptable solution to "turning off" the solenoid once out of the water.