Following the wide development of distributed and networked information sources, more specifically on the Internet, the advantage of benefiting from reference points, such as “portal” sites that offer facilitated, even personalized access to all available resources became apparent. The personalization of information systems (ISs) aims to provide an adaptive and intelligent human- machine interaction with the goal of improving the efficiency of the interaction as well as the usability (in the sense of [NIE 93]) of systems [MOU 09]. It must offer the possibility of accessing ever greater quantities of information on increasingly varied media and support increasingly different interaction modes. In order to meet the usability criteria and enable the user to easily find the information he wants, personalization appears to be an appropriate solution [MOU 09], [PET 03a]. In addition to the personalization of delivered information, other aspects of the interaction can be subject to personalization in this context, such as the consideration of different interaction modes (vocal, textual, etc.), the consideration of different interaction platforms (PC, personalized digital assistant or PDA, cell phone, etc.) and user assistance. Our objective is to provide a support to the design of personalization systems that covers these different aspects. The consideration of interaction modes and platforms used means, for us, that the personalization system must integrate them into its reasoning in relation to the preferences of the user. The personalization system must be able to know or deduce, for example, that for a given piece of information the user would prefer to listen to it from his cell phone, rather than read it on his PDA [...]