Normalisation by random descent

Vincent Van Oostrom, Yoshihito Toyama

Research output: Chapter in Book/Report/Conference proceedingConference contribution

8 Citations (Scopus)

Abstract

We present abstract hyper-normalisation results for strategies. These results are then applied to term rewriting systems, both first and higher-order. For example, we show hyper-normalisation of the left-outer strategy for, what we call, left-outer pattern rewrite systems, a class comprising both Combinatory Logic and the λβ-calculus but also systems with critical pairs. Our results apply to strategies that need not be deterministic but do have Newman's random descent property: all reductions to normal form have the same length, with Huet and Lévy's external strategy being an example. Technically, we base our development on supplementing the usual notion of commutation diagram with a notion of order, expressing that the measure of its right leg does not exceed that of its left leg, where measure is an abstraction of the usual notion of length. We give an exact characterisation of such global commutation diagrams, for pairs of reductions, by means of local ones, for pairs of steps, we dub Dyck diagrams.

Original languageEnglish
Title of host publication1st International Conference on Formal Structures for Computation and Deduction, FSCD 2016
PublisherSchloss Dagstuhl- Leibniz-Zentrum fur Informatik GmbH, Dagstuhl Publishing
Volume52
ISBN (Electronic)9783959770101
DOIs
Publication statusPublished - 2016 Jun 1
Event1st International Conference on Formal Structures for Computation and Deduction, FSCD 2016 - Porto, Portugal
Duration: 2016 Jun 222016 Jun 26

Other

Other1st International Conference on Formal Structures for Computation and Deduction, FSCD 2016
Country/TerritoryPortugal
CityPorto
Period16/6/2216/6/26

Keywords

  • Commutation
  • Hyper-normalisation
  • Random descent
  • Strategy

ASJC Scopus subject areas

  • Software

Fingerprint

Dive into the research topics of 'Normalisation by random descent'. Together they form a unique fingerprint.

Cite this