DATA DICTIONARY FOR LEARNING DATA ANALYSIS.

A data dictionary system, DADICS, was conceived and developed as an online activity for students of an OU data analysis course. The system holds a highly structured knowledge base of facts about a number of conceptual data models, corresponding modeling rules, and the user's session state. These all form a semantic network of sentences, each with a subject, verb, and one or several objects. By simple commands the user may: choose a model representation - e. g. entities, attributes and relationships; select a local or global model; and list as sentences the relevant facts about any element. Alternatively he may update a model, insert facts as discovered, and command the system to highlight inconsistencies. The implementation is a relatively short general-purpose Pascal program that interprets the stored facts to determine what is required.