Developing Distributed Systems with Multiparty Functional Reactive Programming.

Functional Reactive Programming (FRP) is a programming paradigm that describes a system as a data flow between time-varying values, a concept that abstracts values that change over time. It helps describe embedded systems and GUIs. FRP is also useful for describing distributed applications, but avoiding glitches is more complicated. Glitches are inconsistencies between time-varying values caused by differences in update timing and are unavoidable in distributed systems where communication delays are inevitable. Existing methods to avoid glitching incur high communication overhead to ensure consistency among the system’s time-varying values. To address this problem, we propose the concept of multiparty FRP and a language MPFRP based on it. A party is a subset of the time-varying values that make up the system and are updated independently. We propose avoiding glitches with small communication overhead by guaranteeing consistency within each party.