Waveform (output least squares) inversion of seismic reflection data can reconstruct remarkably detailed models of subsurface structure, and take into account essentially any physics of seismic wave propagation that can be modeled. However the waveform inversion objective has many spurious local minima, hence convergence of descent methods (mandatory because of problem size) to useful Earth models requires accurate initial estimates of long-scale velocity structure. Migration velocity analysis, on the other hand, is based on the Born approximation but is capable of correcting substantially erroneous in itial velocities. Appropriate choice of objective (differential semblance) turns migrat ion velocity analysis into an optimization problem, for which Newton-like methods exhib it little tendency to stagnate at nonglobal minima. The extended modeling concept links these two apparently unrelated approaches to estimation of Earth structure: from this point of view, migration velocity analysis is a solution method for the linearized (single scattering, Born) waveform inversion problem. Extended modeling also provides a basis for a nonlinear generalization of migration velocity analysis. Preliminary numerical evidence suggests that this new approach to nonlinear waveform inversion may combine the global convergence of velocity analysis with the physical fidelity of least squares.