Estudo empírico sobre duplicação de código em aplicações JavaScript

Abstract

Clones are similar or identical code fragments on source code. Code cloning can cause several issues related to code quality and maintenance. Consequently, clone detection is an active research area in Software Engineering. However, few clone detection studies focus on dynamically-typed languages such as JavaScript. In this work, a tool capable of extracting clone candidates in JavaScript files was developed and it uses a minimum informed threshold. The tool identified clone candidates from three different projects; then, we conducted a manual validation step for the identification and classification of clones. The tool showed an average of 68.4% success in detecting clones, finding all classifications of clone types. The conclusion of this work is that a manual analysis step is necessary for the candidate’s validation, since the clones interpretation for the JavaScript language can have different results for each professional during the analysis process.