Towards linking correctness conditions for concurrent objects and contextual trace refinement

TL;DR

This paper explores how correctness conditions for concurrent objects relate to the correctness of client programs that use them, linking these conditions with trace refinement to ensure overall correctness.

Contribution

It introduces a framework connecting correctness conditions of concurrent objects with client program correctness through trace refinement.

Findings

Links correctness conditions to trace refinement

Addresses correctness in client contexts

Provides insights into ensuring overall program correctness

Abstract

Correctness conditions for concurrent objects describe how atomicity of an abstract sequential object may be decomposed. Many different concurrent objects and proof methods for them have been developed. However, arguments about correctness are conducted with respect to an object in isolation. This is in contrast to real-world practice, where concurrent objects are often implemented as part of a programming language library (e.g., java.util.concurrent) and are instantiated within a client program. A natural question to ask, then is: How does a correctness condition for a concurrent object ensure correctness of a client program that uses the concurrent object? This paper presents the main issues that surround this question and provides some answers by linking different correctness conditions with a form of trace refinement.