The use of the moon as a passive reflector for radio interferometry experiments with baselines on the order of the radius of the orbit of the moon is discussed. Because of the extreme loss in the lunar interferometer path, only exceptionally strong point sources are candidates for this kind of interferometry. As the moon is far from an ideal reflector due to the irregular surface, the data processing necessary to optimize the fringe detectability is rather complicated. The optimum processing procedure is given and the signal-to-noise ratio for these processors is computed. Two specific applications are discussed, one being the spatial resolution of Jupiter bursts, the other the resolution of the H2O maser at 22 GHz in Orion. Both experiments are definitely possible with existing (Arecibo) or currently planned telescopes (Green Bank). 8 refs.