In 1971, not only the theoretical and by definition already existing ‘ideal memristor’ concept was introduced, but a real memristor device was suggested on grounds of the already known real inductors. The latter is a scientifically significant hypothesis grounded in fundamental symmetries of basic physics, here electro-magnetism. 2008 claimed the discovery of the “missing memristor.” Controversy arose: The devices were not new, and the hypothesized device needs magnetism but has no material memory, while the available devices constitute analogue memory that would work in a world without magnetism. Nevertheless, even the originator of the prediction accepted the discovery. Defenders of the 2008 claim emphasize that the devices are not merely ‘memristive systems,’ which is however a distinction defined in 1976, not 1971. We clarify widely confused concepts and maintain that the originally hypothesized real memristor device is missing and likely impossible. The argument is illustrated also by finding an ideal mechanical memristor element and purely mechanical memristive systems, and hypothesizing a real mechanical memristor device that requires inert mass just like the 1971 implied device requires magnetic induction.