An Optimised Flow for Futures: From Theory to Practice

Chappe, Nicolas; Henrio, Ludovic; Maillé, Amaury; Moy, Matthieu; Renaud, Hadrien

doi:10.22152/programming-journal.org/2022/6/3

Cited by 3 publications

(3 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In particular pure calculi are not able to represent cycles of futures [10]. On the other hand, a concurrent λ-calculus with futures [18] and the DeF calculus [5] feature imperative aspects but no cooperative scheduling, which is crucial to many active objects languages. Additionally, DeF separates cleanly global state and local variables and uses a notion of functions closer to object methods instead of λ-calculus.…”

Section: Formal Model For Futures and Effectsmentioning

confidence: 99%

“…or await, this would involve unwrapping and rewrapping the promises. Dealing efficiently with delegation in asynchronous invocations is a well-studied problem [6,7,5]. In [6], one construct called forward was suggested for such delegations; it was then shown that directly forwarding an asynchronous invocation (return(async(e))) could be efficiently and safely implemented using promises.…”

Section: Forwardmentioning

confidence: 99%

“…Hiding Semantics and Middle-step Reductions Proof of correctness of translations between languages and calculi often reduce to simulation or bisimulation proofs [6,5,16] between a source program and a transformed program. Often, it is however necessary for the transformed program to do more steps than the original one.…”

Section: Deep and Shallow Handlersmentioning

confidence: 99%

See 2 more Smart Citations

Active Objects Based on Algebraic Effects

Andrieux,

Henrio,

Radanne

2024

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

Algebraic effects are a long-studied programming language construct allowing the implementation of complex control flow in a structured way. With OCaml 5, such features are finally available in a mainstream programming language, giving us a great opportunity to experiment with varied concurrency constructs implemented as simple libraries. In this article, we explore how to implement concurrency features such as futures and active objects using algebraic effects, both in theory and in practice. On the practical side, we present a library of active objects implemented in OCaml, with futures, cooperative scheduling of active objects, and thread-level parallelism. On the theoretical side, we formalise the compilation of a future calculus that models our library into an effect calculus similar to the primitives available in OCaml; we then prove the correctness of the compilation scheme.

show abstract

Section: Formal Model For Futures and Effectsmentioning

confidence: 99%