Refactoring effect on internal quality attributes: What haven’t they told you yet?

Automated Refactoring of Legacy JavaScript Code to ES6 Modules

“…Fowler [10] introduced a series of basic refactorings for handling common code smells in object-oriented code and specialized them on the JavaScript language in the second edition of his book [11]. Recent studies and systematic literature reviews focus on the relationship between code smells and refactoring, especially refactoring towards mitigating architectural problems [12,13], and their impact on software quality [14,15,16,17]. Ongoing research expands the application of refactoring techniques to address diverse requirements, beyond design improvements in code or other software artifacts, e.g.…”

Aligning Developer Quality Concerns, Refactoring Applications, and Their Effects

“…Thus, this means that incomplete composite refactorings have a mostly neutral effect on internal quality, which might mean that there are factors related to composite refactorings that might affect refactoring effectiveness. [21] evaluated how composites compare to single refactorings in improving cohesion, (code) complexity, coupling, inheritance, and size of affected elements. They analyzed refactoring operations on 23 software projects, on a set in which nearly 50% of it was comprised of rerefactorings.…”

“…They found that most incomplete composites have a neutral effect on internal quality -neither increasing nor decreasing code quality. Also on the context of composite refactorings, Fernandes et al [21] evaluated how composites compare to single refactorings in improving cohesion, (code) complexity, coupling, inheritance and size of affected elements. Among the analyzed refactorings, 65% improved attributes associated with their refactoring types, while 35% kept them unaffected.…”

Revealing Interacting Factors in Decay of Software Design

“…Finally, size represents the physical size of a code structure (e.g., a method or a class), often measured by the number of lines of code or by the number of statements. Even though many previous studies [44,30,45] associate the size attribute with code quality, due to its connection to bad code structures, such as long methods and long classes, the values of the metrics related to size vary wildly throughout a project's evolution -thus, care should be taken when considering size changes as being a positive or negative influence to the code's quality. Previous studies have also concluded that, in isolation, size metrics may not reflect what developers consider as quality in practice [35].…”

“…For example, LCOM2 and TCC measure the lack of cohesion from different viewpoints and are each only computed by one of these tools. Moreover the selection of these tools was driven by the fact that they implement already validated metrics for design decay in previous studies [21,30,45,53,52].…”