State—and Event-Based Reactive Programming in Shared Dataspaces

Abstract

The traditional Linda programming style, based on the introduction and consumption of data from a common repository, seems not adequate for highly dynamic applications in which it is important to observe modification of the environment which occur quickly. On the other hand, reactive programming seems more appropriate for this kind of applications. In this paper we consider some approaches recently presented in the literature for embedding reactive programming in shared dataspaces, we present a possible classification for these approaches, and we perform a rigorous investigation of their relative advantages and disadvantages.