An environment for Symbolic and Numeric Computation
Abstract
We describe the environment for symbolic and numeric computations, called SYNAPS (Symbolic and Numeric APplicationS) and developed in C++. Its aim is to provide a coherent platform integrating many of the nowadays freely available software in scientific computing. The approach taken here is inspired by the recent paradigm of software developments called active library. In this paper, we explain the design choices of the kernel and their impact on the development of generic and efficient codes for the treatment of algebraic objects, such as vectors, matrices, univariate and multivariate polynomials. Implementation details illustrate the performance of the approach. We describe the environment for symbolic and numeric computations, called SYNAPS (Symbolic and Numeric APplicationS) and developed in C++. Its aim is to provide a coherent platform integrating many of the nowadays freely available software in scientific computing. The approach taken here is inspired by the recent paradigm of software developments called active library. In this paper, we explain the design choices of the kernel and their impact on the development of generic and efficient codes for the treatment of algebraic objects, such as vectors, matrices, univariate and multivariate polynomials. Implementation details illustrate the performance of the approach.