Almost Continuous Transformations of Software and Higher-order Dataflow Programming

TL;DR

This paper introduces a novel approach to higher-order dataflow programming using almost continuous transformations of dataflow graphs, enabling dynamic evolution of programs during execution, exemplified by the Fluid system.

Contribution

It presents a new concept of almost continuous transformations in dataflow graphs and a system, Fluid, for programming with evolving dataflow structures.

Findings

Demonstrates the feasibility of evolving dataflow graphs during execution.

Introduces the Fluid system for dynamic dataflow programming.

Shows potential for flexible, adaptive stream-based computation.

Abstract

We consider two classes of stream-based computations which admit taking linear combinations of execution runs: probabilistic sampling and generalized animation. The dataflow architecture is a natural platform for programming with streams. The presence of linear combinations allows us to introduce the notion of almost continuous transformation of dataflow graphs. We introduce a new approach to higher-order dataflow programming: a dynamic dataflow program is a stream of dataflow graphs evolving by almost continuous transformations. A dynamic dataflow program would typically run while it evolves. We introduce Fluid, an experimental open source system for programming with dataflow graphs and almost continuous transformations.