MultiChor: Census Polymorphic Choreographic Programming with Multiply Located Values

TL;DR

MultiChor is a Haskell library enabling polymorphic choreographies that adapt to varying numbers of participants, simplifying reasoning and deadlock detection in distributed programming.

Contribution

It introduces MultiChor, a novel library that supports polymorphic choreographies without extra communication, expanding the expressiveness of choreographic programming in Haskell.

Findings

Supports polymorphism over participant count

Eliminates excess communication for Knowledge-of-Choice

Enables static deadlock checking

Abstract

Choreographic programming is a concurrent paradigm in which a single global program called a choreography describes behavior across an entire distributed network of participants. Choreographies are easier to reason about than separate programs running in parallel, and choreographic programming systems can check for deadlocks statically. We present MultiChor, a library for writing and running choreographies as monadic values in Haskell. Unlike prior Haskell implementations, MultiChor does not require excess communication to handle Knowledge-of-Choice. Unlike all prior general-purpose choreographic languages, MultiChor can express choreographies that are polymorphic over the number of participants.