Abstract: | While there have been considerable efforts over the past 30 years to improve productivity in scientific computation through the creation of subroutine libraries, much of the mundane, error-prone work in developing simulation codes has remained. This situation has spurred the development of specialized efforts in both the numerical and symbolic computation domains. For instance, numerical software like PDECOL, L1SODE, and UNPACK will solve large classes of partial differential equations, ordinary differential equations, and linear equations, respectively. On the symbolic side of this issue, a few basic tools for developing simulation codes were created by Wirth in the late 1970s. We introduce more advanced uses of symbolic techniques, including two strategies that link the symbolic and numeric computing approaches in the context of simulation codes. |