This paper makes explicit the fact that the ''focus variation method'', widely presented recently as a new approach to direct recovery of the complex wave function in electron microscopy, in fact embraces several methods, of which those most widely described (involving paraboloids in a 3D Fourier space) are not in fact new except in their application to larger data sets, but were established 15-25 years ago. They are powerful (if little practised), but they have not been in any way transformed by recent work, and remain precisely as effective in reducing the non-linear image contributions (by a factor root N for N images), and minimising the effects of chromatic aberration, as they have always been. There is in fact no need to record a quasi-continuous focal series, nor a long series, to secure the benefits of the restoration - nor even a focal series at all.