La pauvreté chez les personnes d’origine étrangère

Interactive systems often consider for interaction a single context of use of an able-bodied user with a desktop PC in a stable environment. Conversely, users are heterogeneous, interact with different devices in different environments, and require context-aware adaptation (CAA). Although adaptation has been largely studied since the 90's, its study has been constrained, e.g. by considering one aspect of the context (i.e. user, platform or environment), or by handling dimensions in a limited approach with simple rules, or by adapting one system aspect (as content or presentation).
Moreover, the users benefit not always is a priority, making them lost or without control over the adaptation. Existing frameworks about CAA are often technologically driven, narrow in scope or obsolete. Due to these shortcomings, stakeholders have not enough support during the development of CAA. To address these issues and to bridge the gap between high-level adaptation goals and implementation of adaptation techniques, this thesis presents a conceptual framework for user interface adaptation, joining dimensions that compose the variety of contexts of use through users, platforms, and environments, and the variety of aspects of an interactive system, including contents, presentation and navigation. This framework, named TriPlet, is structured in three elements: a meta-model (CAMM) covering the whole CAA lifecycle, its concepts and properties, a reference framework (CARF) that extensively defines adaptation concepts to support design decisions, and a design space (CADS) for assessing CAA levels with well-defined criteria.