Who should take care of the personalization?

In this paper, a new approach is proposed for the personalization of the access to different information sources (servers, Web pages, services, etc.) distributed across the Internet. In contrast to the traditional approaches, with personalization software owned and controlled by the owner of the information source, we propose to use the software that is prepared and controlled directly by the end-users. Our approach is based on software agent technology and imperative programming of agent code. The main goal of using personalization agents is twofold. First, such agents act as information brokers, adjusting both contents and format of the information to individual user requirements, restrictions imposed by the end-user hardware and communication means, current situation, etc. Second, the agents are used as autonomous monitors, individually informing agent owners about “important” (from the particular owner point of view) information changes. The fact of using source-independent personalization agents makes it possible to personalize access to such traditionally closed and fixed (i.e., unmanageable from the end-user point of view) information sources and systems, as e-banks, public Web portals and information servers, etc. Due to the fact the agents are prepared by (or at least for) particular users, the expenses related with the development of the agent code are in the major part incurred by these users.