Stacking cycles: functional transformation of circular data

  • Authors:
  • Baltasar Trancón y Widemann

  • Affiliations:
  • Department of Computer Science, Technische Universität Berlin, Germany

  • Venue:
  • IFL'02 Proceedings of the 14th international conference on Implementation of functional languages
  • Year:
  • 2002

Quantified Score

Hi-index 0.00

Visualization

Abstract

Functional programming is very powerful when applied to tree-shaped data. Real-world software problems often involve circular graph-shaped data. In this paper, we characterize a class of functions on circular data graphs extending the class of primitively corecursive functions. We propose an abstract, effective implementation technique for these functions under an eager evaluation strategy on standard stack machines. The proposed implementation ensures termination and can be tuned either for exactness or for speed. The latter variant is asymptotically as efficient as standard implementations of algebraic recursion, at the price of suboptimal homomorphic result graphs with decently bounded redundancy.