RESUMEN

The main difficulties in designing collaborative-learning applications are understanding which interactions between learners and teachers are the appropriate ones and then building software that takes such dynamics into account. Some researchers address this challenge by defining reusable components for specific learning tasks. The problems are that developers impose their particular solution, adapting components to requirement changes requires much effort, developers might lack the expertise to identify the appropriate group dynamics, and the resulting software might be used in unexpected ways. An approach based on an interaction model lets designers and educational experts specify the interactions in a learning activity. They can then define a set of reusable interaction components. The resulting flexible model is the basis for a flexible architecture that lets you create, extend, reuse, and compose diverse educational software products.

RESUMEN

Developers typically use software architecture models at design time to capture significant decisions about a software system's organization and to describe and establish a common understanding about the system's abstract properties. Architectural information isn't usually explicitly represented at runtime; developers transform and realize the architectural properties through runtime artifacts. Recently, the introduction of software platforms supporting component plug-in and dynamic binding has facilitated adaptation of software systems at runtime. Preserving the properties described by the architecture model during adaptation is an important task. We propose a self-adaptation approach that exploits architecture models to reason about and control adaptation at runtime. We can derive runtime models from design models facilitating the developers’ task. We developed the approach in the context of mobile computing.This article is part of a focus section on software architecture