Code smells are symptoms of poor design and implementation choices. Previous studies empirically assessed the impact of smells on code quality and clearly indicate their negative impact on maintainability, including a higher bug-proneness of components affected by code smells. In this paper, we capture previous findings on bug-proneness to build a specialized bug prediction model for smelly classes. Specifically, we evaluate the contribution of a measure of the severity of code smells (i.e., code smell intensity) by adding it to existing bug prediction models based on both product and process metrics, and comparing the results of the new model against the baseline models. Results indicate that the accuracy of a bug prediction model increases by adding the code smell intensity as predictor. We also compare the results achieved by the proposed model with the ones of an alternative technique which considers metrics about the history of code smells in files, finding that our model works generally better. However, we observed interesting complementarities between the set of buggy and smelly classes correctly classified by the two models. By evaluating the actual information gain provided by the intensity index with respect to the other metrics in the model, we found that the intensity index is a relevant feature for both product and process metrics-based models. At the same time, the metric counting the average number of code smells in previous versions of a class considered by the alternative model is also able to reduce the entropy of the model. On the basis of this result, we devise and evaluate a smell-aware combined bug prediction model that included product, process, and smell-related features. We demonstrate how such model classifies bug-prone code components with an F-Measure at least 13 percent higher than the existing state-of-the-art models.

@en

Code smells are sub-optimal implementation choices applied by developers that have the effect of negatively impacting, among others, the change-proneness of the affected classes. Based on this consideration, in this paper we conjecture that code smell-related information can be effectively exploited to improve the performance of change prediction models, i.e., models having the goal of indicating which classes are more likely to change in the future. We exploit the so-called intensity index—a previously defined metric that captures the severity of a code smell—and evaluate its contribution when added as additional feature in the context of three state of the art change prediction models based on product, process, and developer-based features. We also compare the performance achieved by the proposed model with a model based on previously defined antipattern metrics, a set of indicators computed considering the history of code smells in files. Our results report that (i) the prediction performance of the intensity-including models is statistically better than the baselines and, (ii) the intensity is a better predictor than antipattern metrics. We observed some orthogonality between the set of change-prone and non-change-prone classes correctly classified by the models relying on intensity and antipattern metrics: for this reason, we also devise and evaluate a smell-aware combined change prediction model including product, process, developer-based, and smell-related features. We show that the F-Measure of this model is notably higher than other models.@en

Context. The demand for green software design is steadily growing higher especially in the context of mobile devices, where the computation is often limited by battery life. Previous studies found how wrong programming solutions have a strong impact on the energy consumption. Objective. Despite the efforts spent so far, only a little knowledge on the influence of code smells, i.e.,symptoms of poor design or implementation choices, on the energy consumption of mobile applications is available. Method. To provide a wider overview on the relationship between smells and energy efficiency, in this paper we conducted a large-scale empirical study on the influence of 9 Android-specific code smells on the energy consumption of 60 Android apps. In particular, we focus our attention on the design flaws that are theoretically supposed to be related to non-functional attributes of source code, such as performance and energy consumption. Results. The results of the study highlight that methods affected by four code smell types, i.e.,Internal Setter, Leaking Thread, Member Ignoring Method, and Slow Loop, consume up to 87 times more than methods affected by other code smells. Moreover, we found that refactoring these code smells reduces energy consumption in all of the situations. Conclusions. Based on our findings, we argue that more research aimed at
designing automatic refactoring approaches and tools for mobile apps is needed.@en

Software testing is a key activity to control the reliability of production code. Unfortunately, the effectiveness of test cases can be threatened by the presence of faults. Recent work showed that static indicators can be exploited to identify testrelated issues. In particular test smells, i.e., sub-optimal design choices applied by developers when implementing test cases, have been shown to be related to test case effectiveness. While some approaches for the automatic detection of test smells have been proposed so far, they generally suffer of poor performance: as a consequence, current detectors cannot properly provide support to developers when diagnosing the quality of test cases. In this paper, we aim at making a step ahead toward the automated detection of test smells by devising a novel textual-based detector, coined TASTE (Textual AnalySis for Test smEll detection), with the aim of evaluating the usefulness of textual analysis for
detecting three test smell types, General Fixture, Eager Test, and Lack of Cohesion of Methods. We evaluate TASTE in an empirical study that involves a manually-built dataset composed of 494 test smell instances belonging to 12 software projects, comparing the capabilities of our detector with those of two code metrics-based techniques proposed by Van Rompaey et al. and Greiler et al.
Our results show that the structural-based detection applied by existing approaches cannot identify most of the test smells in our dataset, while TASTE is up to 44% more effective. Finally, we find that textual and structural approaches can identify different sets of test smells, thereby indicating complementarity.@en

Continuous changes applied during software maintenance risk to deteriorate the structure of a system and are a threat to its maintainability. In this context, predicting the portions of source code where specific maintenance operations should be focused on may be crucial for developers to prevent maintainability issues. Previous work proposed change prediction models relying on product and process metrics as predictors of change-prone source code classes. However, we believe that existing approaches still miss an important piece of information, i.e., developer-related factors that are able to capture the complexity of the development process under different perspectives. In this paper, we firstly investigate three change prediction models that exploit developer-related factors (e.g., number of developers working on a class) as predictors of change-proneness of classes and then we compare them with existing models. Our findings reveal that these factors improve the capabilities of change prediction models. Moreover, we observed interesting complementarities among the prediction models. For this reason, we devised a novel change prediction model exploiting the combination of developer-related factors and product and evolution metrics. The results show that such a combined model is up to 22% more effective than the single models in the identification of change-prone classes.@en

Code smells are symptoms of poor design or implementation choices that have a negative effect on several aspects of software maintenance and evolution, such as program comprehension or change- and fault-proneness. This is why researchers have spent a lot of effort on devising methods that help developers to automatically detect them in source code. Almost all the techniques presented in literature are based on the analysis of structural properties extracted from source code, although alternative sources of information (e.g., textual analysis) for code smell detection have also been recently investigated. Nevertheless, some studies have indicated that code smells detected by existing tools based on the analysis of structural properties are generally ignored (and thus not refactored) by the developers. In this paper, we aim at understanding whether code smells detected using textual analysis are perceived and refactored by developers in the same or different way than code smells detected through structural analysis. To this aim, we set up two different experiments. We have first carried out a software repository mining study to analyze how developers act on textually or structurally detected code smells. Subsequently, we have conducted a user study with industrial developers and quality experts in order to qualitatively analyze how they perceive code smells identified using the two different sources of information. Results indicate that textually detected code smells are easier to identify and for this reason they are considered easier to refactor with respect to code smells detected using structural properties. On the other hand, the latter are often perceived as more severe, but more difficult to exactly identify and remove.

@en

Regression testing is performed during maintenance activities to assess whether the unchanged parts of a software behave as intended. To reduce its cost, test case prioritization techniques can be used to schedule the execution of the available test cases to increase their ability to reveal regression faults earlier. Optimal test ordering can be determined using various techniques, such as greedy algorithms and meta-heuristics, and optimizing multiple fitness functions, such as the average percentage of statement and branch coverage. These fitness functions condense the cumulative coverage scores achieved when incrementally running test cases in a given ordering using Area Under Curve (AUC) metrics. In this paper, we notice that AUC metrics represent a bi-dimensional (simplified) version of the hypervolume metric, which is widely used in many-objective optimization. Thus, we propose a Hypervolume-based Genetic Algorithm, namely HGA, to solve the Test Case Prioritization problem when using multiple test coverage criteria. An empirical study conducted with respect to five state-of-the-art techniques shows that (i) HGA is more cost-effective, (ii) HGA improves the efficiency of Test Case Prioritization, (iii) HGA has a stronger selective pressure when dealing with more than three criteria.@en

Code smells are symptoms of poor design and implementation choices that may hinder code comprehensibility and maintainability. Despite the effort devoted by the research community in studying code smells, the extent to which code smells in software systems affect software maintainability remains still unclear. In this paper we present a large scale empirical investigation on the diffuseness of code smells and their impact on code change- and fault-proneness. The study was conducted across a total of 395 releases of 30 open source projects and considering 17,350 manually validated instances of 13 different code smell kinds. The results show that smells characterized by long and/or complex code (e.g., Complex Class) are highly diffused, and that smelly classes have a higher change- and fault-proneness than smell-free classes.

@en

Predicting the areas of the source code having a higher likelihood to change in the future is a crucial activity to allow developers to plan preventive maintenance operations such as refactoring or peer-code reviews. In the past the research community was active in devising change prediction models based on structural metrics extracted from the source code. More recently, Elish et al. showed how evolution metrics can be more efficient for predicting change-prone classes. In this paper, we aim at making a further step ahead by investigating the role of different developer-related factors, which are able to capture the complexity of the development process under different perspectives, in the context of change prediction. We also compared such models with existing change-prediction models based on evolution and code metrics. Our findings reveal the capabilities of developer-based metrics in identifying classes of a software system more likely to be changed in the future. Moreover, we observed interesting complementarities among the experimented prediction models, that may possibly lead to the definition of new combined models exploiting developer-related factors as well as product and evolution metrics.@en

Modeling the power profile of mobile applications is a crucial activity to identify the causes behind energy leaks. To this aim, researchers have proposed hardware-based tools as well as model-based and software-based techniques to approximate the actual energy profile. However, all these solutions present their own advantages and disadvantages. Hardware-based tools are highly precise, but at the same time their use is bound to the acquisition of costly hardware components. Model-based tools require the calibration of parameters needed to correctly create a model on a specific hardware device. Software-based approaches do not need any hardware components, but they rely on battery measurements and, thus, they are hardware-assisted. These tools are cheaper and easier to use than hardware-based tools, but they are believed to be less precise. In this paper, we take a deeper look at the pros and cons of software-based solutions investigating to what extent their measurements depart from hardware-based solutions. To this aim, we propose a softwarebased tool named PETRA that we compare with the hardwarebased MONSOON toolkit on 54 Android apps. The results show that PETRA performs similarly to MONSOON despite not using any sophisticated hardware components. In fact, in all the apps the mean relative error with respect to MONSOON is lower than 0:05. Moreover, for 95% of the analyzed methods the estimation error is within 5% of the actual values measured using the hardware-based toolkit.@en

Refactoring aims at improving the internal structure of a software system without changing its external behavior. Previous studies empirically assessed, on the one hand, the benefits of refactoring in terms of code quality and developers’
productivity, and on the other hand, the underlying reasons that push programmers to apply refactoring. Results achieved in the latter investigations indicate that besides personal motivation such as the responsibility concerned with code authorship, refactoring is mainly performed as a consequence of changes in the requirements rather than driven by software quality. However,
these findings have been derived by surveying developers, and therefore no software repository study has been carried out to corroborate the achieved findings. To bridge this gap, we provide a quantitative investigation on the relationship between different types of code changes (i.e., Fault Repairing Modification, Feature Introduction Modification, and General Maintenance Modification) and 28 different refactoring types coming from 3 open source projects. Results showed that developers tend to apply a higher number of refactoring operations aimed at improving maintainability and comprehensibility of the source code when fixing bugs. Instead, when new features are implemented, more complex refactoring operations are performed to improve code cohesion. Most of the times, the underlying reasons behind the application of such refactoring operations are represented by the presence of duplicate code or previously introduced self-admitted technical debts.@en

Code smells are symptoms of poor design solutions applied by programmers during the development of software systems. While the research community devoted a lot of effort to studying and devising approaches for detecting the traditional code smells defined by Fowler, little knowledge and support isavailable for an emerging category of Mobile app code smells. Recently, Reimann et al. proposed a new catalogue of Androidspecific code smells that may be a threat for the maintainability and the efficiency of Android applications. However, current tools working in the context of Mobile apps provide limited support and, more importantly, are not available for developers interested in monitoring the quality of their apps. To overcome these limitations, we propose a fully automated tool, coined ADOCTOR, able to identify 15 Android-specific code smells from the catalogue by Reimann et al. An empirical study conductedon the source code of 18 Android applications reveals that the proposed tool reaches, on average, 98% of precision and 98% of recall. We made ADOCTOR publicly available.@en

Energy efficiency is a vital characteristic of any mobile application, and indeed is becoming an important factor for user satisfaction. For this reason, in recent years several approaches and tools for measuring the energy consumption of mobile devices have been proposed. Hardware-based solutions are highly precise, but at the same time they require costly hardware toolkits. Model-based techniques require a possibly difficult calibration of the parameters needed to correctly create a model on a specific hardware device. Finally, software-based solutions are easier to use, but they are possibly less precise than hardware-based solution. In this demo, we present PETrA, a novel software-based tool for measuring the energy consumption of Android apps. With respect to other tools, PETrA is compatible with all the smartphones with Android 5.0 or higher, not requiring any device specific energy profile. We also provide evidence that our tool is able to perform similarly to hardware-based solutions.@en

Information Retrieval (IR) approaches are nowadays used to support various software engineering tasks, such as feature location, traceability link recovery, clone detection, or refactoring. However, previous studies showed that inadequate instantiation of an IR technique and underlying process could significantly affect the performance of such approaches in terms of precision and recall. This paper proposes the use of Genetic Algorithms (GAs) to automatically configure and assemble an IR process for software engineering tasks. The approach (named GA-IR) determines the (near) optimal solution to be used for each stage of the IR process, i.e., term extraction, stop word removal, stemming, indexing and an IR algebraic method calibration. We applied GA-IR on two different software engineering tasks, namely traceability link recovery and identification of duplicate bug reports. The results of the study indicate that GA-IR outperforms approaches previously published in the literature, and that it does not significantly differ from an ideal upper bound that could be achieved by a supervised and combinatorial approach. @en

In this paper, we present TACO (Textual Analysis for Code Smell Detection), a technique that exploits textual analysis to detect a family of smells of different nature and different levels of granularity. We run TACO on 10 open source projects, comparing its performance with existing smell detectors purely based on structural information extracted from code components. The analysis of the results indicates that TACO's precision ranges between 67% and 77%, while its recall ranges between 72% and 84%. Also, TACO often outperforms alternative structural approaches confirming, once again, the usefulness of information that can be derived from the textual part of code components.@en

Trace ability recovery allows software engineers to understand the interconnections among software artefacts and, thus, it provides an important support to software maintenance activities. In the last decade, Information Retrieval (IR) has been widely adopted as core technology of semi-automatic tools to extract trace ability links between artefacts according to their textual information. However, a widely known problem of IR-based methods is that some artefacts may share more words with non-related artefacts than with related ones. To overcome this problem, enhancing strategies have been proposed in literature. One of these strategies is relevance feedback, which allows to modify the textual similarity according to information about links classified by the users. Even though this technique is widely used for natural language documents, previous work has demonstrated that relevance feedback is not always useful for software artefacts. In this paper, we propose an adaptive version of relevance feedback that, unlike the standard version, considers the characteristics of both (i) the software artefacts and (ii) the previously classified links for deciding whether and how to apply the feedback. An empirical evaluation conducted on three systems suggests that the adaptive relevance feedback outperforms both a pure IR-based method and the standard feedback.

@en

In this paper, we formalize the defect-prediction problem as a multiobjective optimization problem. Specifically, we propose an approach, coined as multiobjective defect predictor (MODEP), based on multiobjective forms of machine learning techniques - logistic regression and decision trees specifically - trained using a genetic algorithm. The multiobjective approach allows software engineers to choose predictors achieving a specific compromise between the number of likely defect-prone classes or the number of defects that the analysis would likely discover (effectiveness), and lines of code to be analysed/tested (which can be considered as a proxy of the cost of code inspection). Results of an empirical evaluation on 10 datasets from the PROMISE repository indicate the quantitative superiority of MODEP with respect to single-objective predictors, and with respect to trivial baseline ranking classes by size in ascending or descending order. Also, MODEP outperforms an alternative approach for cross-project prediction, based on local prediction upon clusters of similar classes.

@en

Test case prioritization (TCP) is aimed at finding an ideal ordering for executing the available test cases to reveal faults earlier. To solve this problem greedy algorithms and meta-heuristics have been widely investigated, but in most cases there is no statistically significant difference between them in terms of effectiveness. The fitness function used to guide meta-heuristics condenses the cumulative coverage scores achieved by a test case ordering using the Area Under Curve (AUC) metric. In this paper we notice that the AUC metric represents a simplified version of the hypervolume metric used in many objective optimization and we propose HGA, a Hypervolume-based Genetic Algorithm, to solve the TCP problem when using multiple test criteria. The results shows that HGA is more cost-effective than the additional greedy algorithm on large systems and on average requires 36% of the execution time required by the additional greedy algorithm.@en

A way to reduce the cost of regression testing consists of selecting or prioritizing subsets of test cases from a test suite according to some criteria. Besides greedy algorithms, cost cognizant additional greedy algorithms, multi-objective optimization algorithms, and multi-objective genetic algorithms (MOGAs), have also been proposed to tackle this problem. However, previous studies have shown that there is no clear winner between greedy and MOGAs, and that their combination does not necessarily produce better results. In this paper we show that the optimality of MOGAs can be significantly improved by diversifying the solutions (sub-sets of the test suite) generated during the search process. Specifically, we introduce a new MOGA, coined as DIversity based Genetic Algorithm (DIV-GA), based on the mechanisms of orthogonal design and orthogonal evolution that increase diversity by injecting new orthogonal individuals during the search process. Results of an empirical study conducted on eleven programs show that DIV-GA outperforms both greedy algorithms and the traditional MOGAs from the optimality point of view. Moreover, the solutions (sub-sets of the test suite) provided by DIV-GA are able to detect more faults than the other algorithms, while keeping the same test execution cost.@en