Programming macro tree transducers

  • Authors:
  • Patrick Bahr;Laurence E. Day

  • Affiliations:
  • University of Copenhagen, Copenhagen, Denmark;University of Nottingham, Nottingham, United Kingdom

  • Venue:
  • Proceedings of the 9th ACM SIGPLAN workshop on Generic programming
  • Year:
  • 2013

Quantified Score

Hi-index 0.00

Visualization

Abstract

A tree transducer is a set of mutually recursive functions transforming an input tree into an output tree. Macro tree transducers extend this recursion scheme by allowing each function to be defined in terms of an arbitrary number of accumulation parameters. In this paper, we show how macro tree transducers can be concisely represented in Haskell, and demonstrate the benefits of utilising such an approach with a number of examples. In particular, tree transducers afford a modular programming style as they can be easily composed and manipulated. Our Haskell representation generalises the original definition of (macro) tree transducers, abolishing a restriction on finite state spaces. However, as we demonstrate, this generalisation does not affect compositionality.