The data flow programming language CAJOLE - an informal introduction

  • Authors:
  • C. L. Hankin;H. W. Glaser

  • Affiliations:
  • Westfield College, London;Westfield College, London

  • Venue:
  • ACM SIGPLAN Notices
  • Year:
  • 1981

Quantified Score

Hi-index 0.00

Visualization

Abstract

Data Flow represents a radical departure from the conventional approach to computer organisation. Rather than a centralised control unit that controls the sequencing of operations, the scheduling of instructions for execution is controlled by the flow of data between operations. In this way the control is distributed and several operations may be executed concurrently. This new approach has major implications for computer architecture and several "data flow" machines have been proposed in the literature.In this paper we present a low level graphical programming language that can be used to represent data flow programs. We also introduce a high level textual language, CAJOLE, that reflects the data flow philosophy. The notations used are simple and elegant in contrast to "fat and flabby" Von Neumann languages.With the advent of fast and cheap microprocessors, we are likely to see a vast increase in the number of tightly-coupled microprocessor systems (a system where several processors cooperate in executing a single task). In this sort of system it is essential to distribute as much as possible the functions that have traditionally been performed by the central control unit. Data flow provides an elegant conceptual basis for this.