Table of Contents
Fetching ...

Meerkat: A Distributed Reactive Programming Language with Live Updates

Heng Zhong, Anrui Liu

TL;DR

Meerkat addresses safe, distributed reactive programming with live updates by introducing a type-safe framework that encodes explicit dependencies between state variables and definitions. It extends prior centralized work to support parallel submissions by multiple programmers and leverages a distributed dependency graph maintained across frontend and backend, with Historiographer ensuring strong consistency and glitch-free propagation. The core contributions include a formal typing system that tracks read/write sets and dependency relations, an atomic action semantics, and mechanisms to prevent partial updates and races in a distributed setting. This work enables scalable, collaborative live updates with strong correctness guarantees in distributed reactive applications.

Abstract

We propose a novel type-safe reactive programming language with live updates that extends an existing work to support multiple distributed evolution queues. Dependency sets of definitions are incorporated in the type system to protect the interaction between the frontend user interface and the backend database. Distributed live updates submitted by multiple programmers are ensured strong consistency based on an existing framework for distributed reactive propagation.

Meerkat: A Distributed Reactive Programming Language with Live Updates

TL;DR

Meerkat addresses safe, distributed reactive programming with live updates by introducing a type-safe framework that encodes explicit dependencies between state variables and definitions. It extends prior centralized work to support parallel submissions by multiple programmers and leverages a distributed dependency graph maintained across frontend and backend, with Historiographer ensuring strong consistency and glitch-free propagation. The core contributions include a formal typing system that tracks read/write sets and dependency relations, an atomic action semantics, and mechanisms to prevent partial updates and races in a distributed setting. This work enables scalable, collaborative live updates with strong correctness guarantees in distributed reactive applications.

Abstract

We propose a novel type-safe reactive programming language with live updates that extends an existing work to support multiple distributed evolution queues. Dependency sets of definitions are incorporated in the type system to protect the interaction between the frontend user interface and the backend database. Distributed live updates submitted by multiple programmers are ensured strong consistency based on an existing framework for distributed reactive propagation.
Paper Structure (3 sections, 2 equations)

This paper contains 3 sections, 2 equations.