Flux - A Data-Flow Programming Language

The goal of this paper is to establish the design requirements of a programming language enabled to extract the maximum parallelism from source code. The structure of imperative and functional languages are analyzed in regards to parallelism. Next, the data-flow paradigm is introduced. Its major obstacles are identified and solutions are provided based on compile-time analysis. The resulting language should enable wide-scale parallelism, scaling from instruction to cluster.