A recurring theme in theorizing about human-computer interaction is the duality of the following two views on the computer:
Ad 2. Alternatively, there is a trend, in which goal-oriented and intentional behavior is explicitly introduced in so-called intelligent agents, which behave according to a logic containing ``beliefs'' about the world and about the goals of the user. Synergistic with this trend is the quickly growing capability with which human-like output (speech, faces, movement) can be more or less faithfully produced by computers. Critics of this latter approach pejoratively call it ``animism'' and point out that user-friendly design is much more helped by giving the user the idea that he/she is fully in control of the situation, rather than having to negotiate with a potentially uncooperative anthropomorphic servant. However, the anthropomorphic dialogue may also be compelling, and the metaphors of ``partner'', ``colleague'' or ``coach'' may be used to make clear to the user what are the characteristics of the dialogue , in terms of protocol, attitude, and etiquette.
Rather than deciding a priori for either one of these basic interface philosophies, different types of interaction and different application conditions must be defined within this project which are suited for either the Computer-as-Tool or the Computer-as-Dialogue-Partner approach, possibly also combining aspects of both approaches where possible. As an example: Whereas it may be acceptable to a user to organize the planning of daily activities in a dialogue with an anthropomorphical partner, it is equally unacceptable to use a human-style dialogue in detailed low-level aspects of text editing or computer programming.
Within the project of the concepts introduced here will be substantially refined.