Holography, migration, and tomography can be derived from the Porter-Bojarski integral. The different algorithms simply deal with data differing in angular and/or frequency diversity and whether inhomogeneous waves are completely ignored or assumed to be weak. If we apply these methods to image the subsurface, we should cover the biggest portion of the k-space possible: We should combine transmission mode and reflection mode data. Angles and frequencies should be as diverse as possible to allow superposition of many halfcircles. If we remember that the data and a priori knowledge is imperfect, a non-linear inversion should be applied on the highly diverse data. Mora (1989) summarized these ideas as he wrote ``inversion = migration + tomography.''