Making systems communicate and cooperate: The Central Control approach

Abstract

In this paper we present the Central Control, a software component designed to be the kernel of environments for scientific computation and which can offer a common and concurrent access to many tools needed by the scientist and the engineer. The Central Control (often abbreviated as CC in the sequel) communicates with servers that can be general purpose or specialized computer algebra systems, numerical systems, visualization programs, graphical interfaces etc. The Central Control can abstract the syntaxic and semantic differences of the systems so that, for example, an expression computed by Mathematica can be used as input to Maple. The CC is in fact an extended Scheme interpreter. This enables the dynamic configuration of a network of servers to distribute computations using the full power of the Scheme language. Architectures which support the solution of mathematical problems by linking specialized components are important to the future growth of Computer Algebra. The CC is a tool designed to make experiments in this direction easily. It has been used to build an efficient library for computational algebraic geometry. This library has been used to implement a new algorithm for primary decomposition of ideals.

In this article, we present the main features of the Central Control, illustrated with some examples of actual Scheme code. We also discuss the possible uses of the CC, its implementation, its relation with previous work and the problems and benefits of our approach.