TICL 2: Knowledge Representation, Associated Theories and Implications for Instructional Systems Dialog on Deep Structures
Joseph M. Scandura, Ken Koedinger, Stellan Ohlsson, Antonija Mitrtovic and Gilbert Parquette
This discussion forum is designed to compare and contrast major alternative deep infrastructures on which progress in TICL depends. This issue has the goal of increasing understanding, specifically by exposing and clarifying basic similarities and differences between three invited articles in TICL Vol. 5, No. 2. Each focuses on a different basic approach to deep infrastructure in TICL. The Ohlsson & Mitrovic article focuses on production systems (PS) and constraint based modeling (CBM), and their use in building ITS. Paquette focuses on instructional systems derived from instructional design principles and relational networks. Scandura focuses on Abstract Syntax Trees (ASTs) and the central role they play in Structural Learning Theory (SLT) and in building adaptive and configurable tutoring systems derived thereon.
The ensuing discussion begins with a dialog involving the three sets of authors, and those they have directly challenged. In all cases the goal has been both clarification and comparison. Initially the dialog focuses on comments and responses about the three core articles, by authors themselves and/or those whose ideas they may have challenged. Once these authors have had their say, probing questions, insightful comments and reasoned criticisms based on or deriving from the articles and ensuing discussion will be welcomed from informed researchers who have read the articles and the above commentary. Questioners will include both other authors in the special triple issue, and invited questioners at a forthcoming 2008 TICL symposium in New York. All are leaders in the field.
Both the initial and this follow-on issue were planned and executed in the belief that true progress in TICL can only, or at least can best be achieved only by clarifying the many similarities and differences that both characterize and retard progress in our field. These similarities and differences have often gone undetected – sometimes for decades – often hidden in different terminology or overlapping ideas. Hopefully, the ensuing dialog will help clear the path to more rapid, cumulative advances in the future.