Fatigue corrosion crack initiation and propagation is modelled as a moving boundary value problem. The model is based on three physical processesoperating at the solid-environment interface ñ material dissolution, passive filmformation and surface straining. The dissolution triggers boundary advancement.The rate of boundary advancement depends on the passive film damage causedby the surface straining. Plane edge cracks, nucleating from surfaceirregularities, are considered. The cracks obtain realistic geometrical shapeswhere the near-tip region is an integral part of the crack surface. Elastic-perfectlyplastic materials are considered and a low-cycle fatigue load is assumed. Theproblem is solved using a FEM based program and procedures for movingboundary tracking and interior re-meshing. A crucial ingredient of the boundarytracking is the evolved surface re-meshing, where a scheme based on length andcurvature constraints is utilised. The work studies how the choice of theseconstraints influences the results for crack surface evolution. It is shown thatcharacteristic length parameters in crack nucleation and short crack growthdepend on the choice of the constraints. It is concluded that an additionalphysical process operating at the surface has to be accounted for in order todescribe the length scales observed in reality.