This paper reviews a major disk-drive product development effort in IBM during the 1950s, one that was originally based on perpendicular recording. The goal of the project, identified as the Advanced Disk File (ADF), was to develop the successor to the first disk drive, the Random Access Method of Accounting and Control (RAMAC) 350. The ADF eventually became the IBM 1301, the first disk drive to use a flying head per surface. Whereas the RAMAC used longitudinal recording, the ADF project chose perpendicular recording as the mainstream recording technology for magnetic disk data storage and continued on this path for a period of five years. A crisis arose in the later stages of prototype testing due to unacceptable failure rates; one decision made was to change the recording method from perpendicular back to longitudinal. The reasons for the original selection of perpendicular and the subsequent return to longitudinal recording are described and address technology, product, and business issues. General observations on factors influencing the choices made and the final outcome are offered as well as the long-term impact of these events on the disk-drive industry.