Skip to Main content Skip to Navigation

Type-based static analysis of structural properties in programming languages

Abstract : We introduce a general static analysis framework for inferring `structural' or `usage' properties for functional programs. the term `structural', borrowed from proof theory, is used here to suggest a strong connection with linear logic, for which the structural rules of weakening and contraction play an important role. We present the problem of static analysis in this context as that of finding a translation from a PCF-like source language into the intermediate language decorated with structural annotations. we show how the set of all possible translations can be characterised as a set of inequations over a suitable algebraic annotation ordered set. in particular, we are interested in the least solution of this set of inequations, corresponding to the most accurate, or optimal, translation. Most of this thesis dissertation is concerned with the detailed presentation of a case study, linearity analysis, aimed at determining when values are used exactly once. We begin by describing a very simplistic version of linearity analysis and then introduce two further extensions which embody notions of annotation subtyping and polymorphism. The latter turns out to be a key feature in practice, as it allows the analysis to retain its expressive power across separately compiled modules.
Document type :
Complete list of metadata

Cited literature [69 references]  Display  Hide  Download
Contributor : Francisco Alberti Connect in order to contact the contributor
Submitted on : Monday, October 3, 2005 - 3:39:30 PM
Last modification on : Saturday, March 28, 2020 - 2:12:41 AM
Long-term archiving on: : Friday, September 14, 2012 - 2:45:08 PM


  • HAL Id : tel-00010369, version 1



Francisco Alberti. Type-based static analysis of structural properties in programming languages. Software Engineering [cs.SE]. Université Paris-Diderot - Paris VII, 2005. English. ⟨tel-00010369⟩



Record views


Files downloads