Projective ML

  • Authors:
  • Didier Rémy

  • Affiliations:
  • -

  • Venue:
  • LFP '92 Proceedings of the 1992 ACM conference on LISP and functional programming
  • Year:
  • 1992

Quantified Score

Hi-index 0.00

Visualization

Abstract

We propose a projective lambda calculus as the basis for operations on records. Projections operate on elevations, that is, records with defaults. This calculus extends lambda calculus while keeping its essential properties. We build projective ML from this calculus by adding the ML Let typing rule to the simply typed projective calculus. We show that projective ML possesses the subject reduction property, which means that well-typed programs can be reduced safely. Elevations are practical data structures that can be compiled efficiently. Moreover, standard records are difinable in terms of projections.