Just wanted to add: TIF 16bit Half-Float enables us to do a full linear workflow, with post modelation and rendering. We could manually convert to 16bit after, but than we would still have rough 8bit color precision.
So my second request is, when exporting in 16bit, does PS calculates the more precise color values for that format. Internally you shouldn't just convert from 8bit to 16bit, otherwise you would loose color precision.
Take care