Link problems aside, that is very impressive – all the more so since the final grinding and polishing to optical standards is by hand! He doesn't say how long it took, but that must be hours of work with the patience and care of a neurosurgeon.
One of Holtzappfel's books shows how large telescope mirrors and lenses were ground in the 19C by clamping a precision flat disc of glass down onto concentric rings of accurately-set jacks similar to toolmakers' ones, so curving the blank against its own elasticity. (New glass is very, very slightly flexible but it hardens and becomes more brittle with age, explaining why you cannot usually re-cut old windows.) The distorted blank was then ground flat, so when released it would rebound with a concave surface of the appropriate, very large radius.