Abstract

The concept of code smell was introduced in the late 90’s and is a way of referring to certain subjective characteristics in the source code that could have repercussions on operation and maintenance. In order to correct these problems, a whole study area called refactoring has been developed within Software Engineering, which focuses mainly on strategies for correcting this type of anomalous code. The objective of this article is to unveil the current status of studies related to code smells, considering mainly their detection. A review was conducted based on the systematic mapping protocol. The search strategies retrieved a set of 215 documents, of which 30 were selected for analysis. Four categories of classification were defined: detection method, contribution type, programming language and code smell category. The results of the review indicate a predominance of studies that detected code smells in Java. It was found that the literature analyzed lacks studies that report methods, tools and detection strategies in the categories ”object-oriented abusers” and ”change preventers”, while the highest concentration of articles are in the code smells categories called ”bloaters” and ”dispensable”, whose main detection methods correspond to metrics and log analysis.