For the new SPS machine constructed in the early seventies, he proposed that the generation of the reference voltages for the bending and quadrupole supplies should be based on measurements of the field along the cycle, and gave an outline of the correction algorithms. His proposal resulted in the first ever computer-controlled closed-loop system for a geographically distributed system, as the 7 km circumference SPS was; this was a no simple feat for the early 1970s. Measurements of the main magnet currents were introduced only later, when the SPS had to run as a storage ring for the SPS p–pbar collider.