Discussion of results. Figures 11 and 12 show the behaviour of two different meshes which undergo optimization. In one case optimization is not successful, as can be seen in Figure 11c. This is not the fault of the method but rather the fault of the design itself which places mesh polylines relative to asymptotic directions such that case (ii) of Observation 2 applies. It depends on the nature of the mesh if optimization manages to move vertices such that smoothness can be achieved or not. In the case of Figure 12 this works because there is not much to do: Vertices