-

In the writer's opinion the above statements are essentially correct, but some precisions may still be made. The first concerns the statement by Pijaudier-Cabot. If one considers the inelastic displacements introduced in the localized region, the inelastic strain distributions are different for the non-local and for the cohesive crack. However, this is only a matter of scale of representation. Indeed, if one represents graphically the inelastic displacements, a heaviside step function is obtained for the cohesive crack and a continuous distribution for the non-local model. However if the scale of the plot is reduced at will, the continuous distribution will approach the heaviside function to any prescribed accuracy level. Hence, if the thickness of the softening region, as described by the non-local model, is small compared to the size of the structure under consideration, the cohesive crack and the non local model will lead to essentially the same results.