Global abstraction-safe marshalling with hash types

Marshalling: Abstraction, Rebinding, and Version Control

We discuss the design of programming languages for distributed computation, focussing on support for type-safe marshalling of arbitrary language values. In particular: (1) unmarshalling can involve rebinding to local resources; (2) values of abstract types can be communicated, and a globally-coherent notion of type equality ensures that unmarshalling respects abstraction; and (3) interoperation...

Safe Runtime Down asts With Ownership Types

Safe Runtime Downcasts With Ownership Types

The possibility of aliasing between objects constitutes one of the primary challenges in understanding and reasoning about correctness of object-oriented programs. Ownership types provide a principled way of specifying statically enforcable restrictions on object aliasing. Ownership types have been used to aid program understanding and evolution, verify absence of data races and deadlocks in mu...

Types for Safe Locking

A race condition is a situation where two threads manipulate a data structure simultaneously, without synchronization. Race conditions are common errors in multithreaded programming. They often lead to unintended nondeterminism and wrong results. Moreover, they are notoriously hard to diagnose, and attempts to eliminate them can introduce deadlocks. In practice, race conditions and deadlocks ar...

Translucent Abstraction: Safe Views through Invertible Programming

Despite the distinctive advantages of pattern matching in program understanding and reasoning, the tight coupling of interface and implementation has hampered its wider acceptance. Since the first proposal of views two decades ago, significant effort has been invested in tackling this non-modularity; the obvious target has been to decouple datatype implementations from separate interfaces used ...