Next Generation Context-oriented Programming: Embracing Dynamic Generation of Adaptations

By: Nicolás Cardozo, Ivana Dusparic

Abstract

Context-oriented Programming (COP) first appeared in 2005 as a way to enable the dynamic adaptation of software systems to specific situations in their surrounding environment. Multiple COP languages have since been proposed, and used in numerous adaptive systems areas, enabling dynamic swapping and composition of adaptive behavior at run-time. However, until recently, all approaches relied on the offline pre-definition of adaptive behavior, limiting the adaptations to only those foreseen at design time. Auto-COP recently emerged as an approach to shift adaptation definition to run-time, if and when the need for adaptations to new contexts arises, by utilizing reinforcement learning techniques. In this paper, we use Auto-COP as a starting point to discuss the research path to achieve a completely dynamic adaptive system. We discuss the potential benefits of such an automated AI-based approach, present several application domain categories where dynamic adaptation definition would enable adaptivity breakthroughs, and discuss open challenges in developing such a fully automated approach.

Keywords

Context-oriented programming, Reinforcement learning, Dynamic software adaptation.

Cite as:

Nicolás Cardozo, Ivana Dusparic, “Next Generation Context-oriented Programming: Embracing Dynamic Generation of Adaptations”, Journal of Object Technology, Volume 21, no. 2 ( 2022), pp. 1-6, doi:10.5381/jot.2022.21.2.a5.

PDF | DOI | BiBTeX | Tweet this | Post to CiteULike | Share on LinkedIn

This article is accompanied by a video realized by the author(s).

The JOT Journal   |   ISSN 1660-1769   |   DOI 10.5381/jot   |   AITO   |   Open Access   |    Contact