Comparing type-based and proof-directed decompilation

Abstract

In the past couple of years interest in decompilation has widened from its initial concentration on reconstruction of control flow into well-founded-in-theory methods to reconstruct type information. A. Mycroft (1999) described Type-Based Decompilation and S. Katsumata and A. Ohori (2001) described Proof-Directed Decompilation. The article summarises the two approaches and identifies their commonality, strengths and weaknesses; it concludes by suggesting how they may be integrated.