ABSTRACT
Intelligent tutoring systems (ITS) are tools for the 21st century. An ITS capable of helping students learn complex problem-solving tasks requires extremely flexible human-machine interfaces. Trends in knowledge design for ITSs today suggest a balanced implementation of declarative facts and procedural skills. Domain knowledge is most often represented operationally. Although most ITS design is centered on the user interface, the domain obviously has evolutionary implications for how a tutor is conceptualized and how a student views the instruction. Most importantly the knowledge acquisition process for domain representation will become the major research issue in tutoring development. Instructors have knowledge of domains, certainly, but instructional effectiveness is more often achieved because of the relationships good teachers have with individual students. Good teachers or tutors are able to take in the instructional context such that they know what the student does not know in terms of knowledge, and perceive what the student can and cannot do in terms of skill.
