Empirical Modelling in support of constructionism: a case study

Conventional programming paradigms have limitations where support for constructionist learning is concerned. This paper illustrates the merits of an alternative approach to giving support for constructionist learning, based on the principles of Empirical Modelling (EM), with reference to an algorithm for determining whether a decomposition of a particular relational schema is lossless. Modelbuilding that is to be effective for constructionist learning has to support activities relating to three distinct roles: that of a student, a teacher and a developer. Our aim is to demonstrate that EM brings far greater conceptual unity to the interactions of the student, the teacher and the developer than is typically found in using a conventional approach to educational software development.