Derivable partial locking for algebraic data types

Abstract

Parallelism and concurrency are one of the most actively researched fields in Computer Science. Writing concurrent programs is challenging because of the need for synchronization and solving possible race conditions and deadlocks while avoiding unnecessary waiting and overhead. The integrity of the program data can be archived by providing locks for its data structures or using concurrent data structures. Partial locking allows threads to lock exactly those parts of the global data they need to read or update. This article presents a method that helps the implementation of threadsafe programs with Algebraic Data Types [1], By transforming the data model of the application to thread-safe data structures with a built-in, configurable locking mechanism including partial locking. With this support, the programmer can focus on the business logic of his application when writing the program. As part of this article, we prove that the shared version of the calculation will produce the same result as the original one.