Comparing relational database designing approaches: some managerial implications for database training

摘要

This research conducted two experiments to understand the performance (correctness and efficiency) of novice database designers, and perceptions of ease of use and preferences of two approaches for modeling relational databases: the semantic-oriented approach (top-down, e.g. using the entity-relationship model) and the logical-oriented approach (bottom-up, view decomposition, focusing only on the logical model). The findings indicated that in experiment 1, semantic-oriented treatments performed better in a complex, written-text case; logical-oriented treatments were better In a simple, tabular-form case. The same situation happened in experiment 2 though the differences were not statistically significant.