Oracles of Bad Smells: A Systematic Literature Review

• Rafael Prates Ferreira Trindade UFMG
• Mariza Andrade da Silva Bigonha UFMG
• Kecia Aline Marques Ferreira CEFET-MG

Resumo

A bad smell is an evidence of a design problem that may be harmful to the software maintenance. Several studies have been carried out to aid the identification of bad smells, by defining approaches or tools. Usually, the evaluation of these studies' results relies on data of oracles bad smells. An oracle is a set of data of bad smells found in a given software system. Such data serves as a referential template or a benchmark to evaluate the proposals on detecting bad smells. The availability and the quality of bad smell oracles are crucial to assert the quality of detection strategies of bad smells. This study aims to compile the bad smell oracles proposed in the literature. To achieve this, we conducted a Systematic Literature Review (SLR) to identify bad smell oracles and their characteristics. The main result of this study is a catalog of bad smell oracles that may be useful for research on bad smells, especially the studies that propose tools or detection strategies for bad smells.

Palavras-chave: benchmark, bad smell, oracle, systematic literature review, design anomaly, code smell