SCOOP: concurrency made easy

  • Authors:
  • Volkan Arslan;Patrick Eugster;Piotr Nienaltowski;Sebastien Vaucouleur

  • Affiliations:
  • Chair of Software Engineering Swiss Federal Institute of Technology Zurich, Zurich, Switzerland;Chair of Software Engineering Swiss Federal Institute of Technology Zurich, Zurich, Switzerland;Chair of Software Engineering Swiss Federal Institute of Technology Zurich, Zurich, Switzerland;Chair of Software Engineering Swiss Federal Institute of Technology Zurich, Zurich, Switzerland

  • Venue:
  • Dependable Systems
  • Year:
  • 2006

Quantified Score

Hi-index 0.00

Visualization

Abstract

The metaphor of objects as entities encompassing both logic and state, simplifying the design and development of particularly large-scale applications, is well established in the industry. However, large applications are rarely monolithic components that carry out a single sequential task; most applications are composed of many components running in parallel. Yet, the vast majority of such applications are built in a rather ad-hoc manner, typically by making use of threading libraries and explicit synchronization through low-level mechanisms such as semaphores, locks, or monitors layered on top of objects. The Simple Concurrent Object-Oriented Programming (SCOOP) model strives for a higher-level abstraction for concurrency, naturally woven into “traditional” object-oriented constructs. Thanks to the full support for contracts and other object-oriented mechanisms and techniques — inheritance, polymorphism, dynamic binding, genericity, and agents — SCOOP offers the programmer a simple yet powerful framework for efficient development of concurrent systems. This paper presents a survey of SCOOP, including (1) the foundations of the SCOOP paradigm, its computation and synchronization models (focusing on simplicity), and our more recent developments. These are (2) an extended type system for eliminating synchronization defects (improving safety), (3) support for transactional semantics for subcomputations (enforcing atomicity), and (4) an event library for programming real-time concurrent tasks (allowing for predictability).