Synchronization as a Special Case of Access Control

Synchronization ensures exclusive shared-variable access at runtime, and static access control mechanisms give similar guarantees at compilation time. Usually we treat these language concepts as separate. In this work we propose to integrate synchronization into access control in a Java-like language: Shared-variable access depends on the availability of tokens (as a form of access control), and the compiler generates code for locking to gain the needed tokens (synchronization). We get more freedom in expressing synchronization at appropriate points in a program and weaker influence of concurrency on the program structure.