Open AccessTextual-based DSL for Conceptual Database Modeling: A Controlled Experiment
Author(s) -
Jonnathan Lopes,
Maicon Bernardino,
Fábio Paulo Basso,
Elisa Rodrigues
Publication year - 2021
Language(s) - English
Resource type - Conference proceedings
DOI - 10.5753/sbbd.2021.17875
Subject(s) - digital subscriber line , computer science , relation (database) , context (archaeology) , data modeling , quality (philosophy) , variety (cybernetics) , database , conceptual model , precision and recall , data mining , information retrieval , artificial intelligence , telecommunications , paleontology , philosophy , epistemology , biology
The variety of database system technologies that have become available in recent years makes it difficult to select tools for entity-relationship modeling (ER) in the teaching-learning context. This paper reports a replicated controlled experiment carried out with 33 subjects in order to compare effort spent (time) and quality, using the harmonic average between precision and recall, of the models produced with two different approaches. The models were produced in a proposed tool (ERtext) with a textual-based DSL and in another tool with a graphical approach (brModelo). Briefly, the data obtained indicate: i) both approaches present similar performance in relation to associated effort, and; ii) that there is a statistically significant difference in relation to the quality of the generated models, with a slightly advantage for the textual approach. Therefore, we conclude that the use of a textual-based DSL is feasible and our proposal is an acceptable solution in the context of conceptual database modeling.