Only which times does the potential not change with separation? You push similar charges together, doing work (like the battery) and you let them separate and do work against a load.
There's no point in talking about an ideal voltmeter because they don't exist. There is always some energy involved in making a measurement.
You can choose whatever convention you like for Logic circuits. You would also be loony to try to apply Mr K's exceedingly good laws to anything so complex as a microprocessor but you could look at any part of it and they could be useful. We can't apply them Willy Nilly at frequencies at which electromagnetic radiation becomes significant but that's nothing exceptional for a scientific 'Law'. They all run out eventually.