Experience gained in a world wide effort to exploit restoration methods in order to improve the scientific return from the degraded HST optics calms down fears that HST might have lost all of its spatial resolving power. However, the problems posed by the characteristics of the point spread function (PSF) and of the detectors are such that “black-box” image restoration in an automatic procedure during pipe-line calibration is not applicable. Results from restoration experiments show the directions for future research: On the practical side ways are required to properly estimate the complicated PSF, to objectively describe errors, to treat undersampling and non-linear detectors, and to gain factors in throughput for large amounts of data; innovative example are methods to combine data obtained with differing PSFs (ground – space, space – space) and sampling.