The research issues to be addressed include: study of mesh design methodologies; recognition of geometric features of workpiece and tooling; decomposition of two-dimensional domains into regions bounded by four curves; suitability of the hierarchical refinement model of design for mesh design tasks; research of geometric data structures for finite element systems; and automated smoothing of evolving boundaries. As the outcome of this investigation, a novel technology for automated two-dimensional metal forming modeling should result. This technology will contribute to the transformation of non-linear Lagrangian finite element analysis from a research analysis method to an engineering design tool. Furthermore, the availability of automated finite element analysis could enable the integration of process design and analysis systems. Design could be readily evaluated through modeling, and modeling results fed back to the design system for redesign. Finally, the study and understanding of the way in which humans perceive and understand shape has important implications in the automation of those design activities that, like mechanical design, are dominated by geometric reasoning.