An emotion-aware method KALEIDOSCOPE is presented in the paper. It consists of a visualization approach, a method for information transfer-by-sense and applied approaches to maintain a natural-style dialog. The combination is purposed to provoke and keep positive emotions in the user and to maintain his interest in the problem. The main goal of this paper is to show how different machine-done visualized patterns lead to student’s perceptions that can’t be represented or estimated by the machine and how this is helpful during the educational process. It is shown that dynamic information presentations are superior to static ones. Original evolving (dynamic) ontology applications have been introduced. It is shown that the creation and maintenance of such ontologies may be a rather complicated process. Essential principles of an educational process using KALEIDOSCOPE method are introduced. It is shown that high quality education systems are far away from monotonic presentation forms, and that they are rather close to contemporary research systems. Both education and research systems use a clear and thoroughly understandable dialog depending on user’s knowledge level. The considered education system is an active part of the human-computer dialog process. Instead of trying to estimate the user’s emotions we apply a set of logical data mining methods aiming at making the human-computed interaction emotion-aware. The educational part of the system is domain independent. Aiming to show domain independent approaches, the examples are introduced from Number Theory to language expressions and nonclassical logic applications. Applications are presented in the domain of Number Theory and IT information security systems.