The user interface is a crutial element for any interactive system. In fact, from the users' point of view, the quality of the whole system is highly dependent to the quality of the interface, existing many real exmples of excelent systems that fail due to bad interface conceptions. In addition, the quality of a user interface is not an objective concept, as it depends on the user's subjective criterion, and therefor on its characteristics and preferences. The consequence is the impossibility of developing a universal interface that is suitable and adecúate for each of the users of a system.The previous has motivated the launching of many research lines that share the goal of achieving interfaces that are adequate for every target user. Most of the efforts ha ve been placed into two topics: the creation of user models and the definition os adaptation procedures. The main problem is that both are usercentered approaches, and therefor their proposals are dependent to some extent of the application domain and on the channel in v^hich the interaction is performed. This makes it hard to apply their results in systems of different characteristics to the ones used as a reference during research.The situation gets even worst if we take into account the proliferation of new channels that are used, due to the popularization of devices such as PDAs or cellular phones, and environments such as Internet or digital TV. This has forced industry to provide especific developments of their systems for each new channel, deriving again in an increase of the effort needed for crating new interfaces.