Of course, but the point was, that a drift relative to actual time does not cause problems as long as all clocks in the system drift at the same rate while the failure was caused by different clocks in the system drifting relative to each other and therefore with different rates relative to actual time. Therefore I treated actual time as a special clock.