Most adaptive systems adapt to one parameter and assume that everything else remains unchanged. A recommender system like TV Scout assumes that the user’s interests do not change by watching programs on TV. A learning system assumes that the learner’s knowledge about concepts in a course is the only thing that changes. And systems that adapt to the output device or network do not at the same time consider adaptation to the user’s knowledge or interest.
We will now consider systems or applications that do take more than one type of adaptation into account. We will illustrate this second-order adaptation by considering learning applications that perform adaptation not only to the user’s increasing knowledge about the course, but also to the user’s learning style. Considering several aspects of adaptation is inherently difficult because the aspects may influence each other.