Marco Aurélio Graciotto Silva

Context: Managing issues in open-source software projects is challenging and costly, as many developers are casual and/or newcomers. On the one hand, maintainers must ensure the quality of issue descriptions and their labels and create mechanisms for recommending and assigning issues. On the other hand, to complete the issue, contributors must understand it and locate the artifacts related to a given functionality or the defect to be fixed. Objectives: This work aimed to conduct a comparative study of different models for recommending similar issues that could help developers with their contributions. Methods: We collected data on issues and pull requests from 35 open-source projects hosted on GitHub. We used the Term Frequency Inverse Document Frequency (TF-IDF), Sentence BERT (SBERT), and Word2Vec techniques to recommend similar issues and source code to assist newcomers' contributions. Results: The models based on the SBERT and TF-IDF techniques yielded better results in the recommendations generated than Word2Vec in the two evaluated scenarios (general issues and those marked as good for newcomers). SBERT was able to recommend past issues where the code used in the solution was approximately 17% similar to the actual solution of the issue used as a query to evaluate the models, reaching results similar to those of GPT 3.5 and GPT 4. Conclusion: Based on the empirical results obtained, we hope to take the next steps in transferring the knowledge gained to software projects and developers, especially by supporting newcomers developers during their first contribution.

Atualmente os modelos grandes de linguagem (LLM) conseguem resolver com sucesso problemas tipicamente utilizados em disciplinas introdutórias de Computação (CS1). No entanto, de modo geral os estudos consideraram problemas e prompts especificados em inglês. O objetivo deste trabalho é avaliar os efeitos dos idiomas português e inglês utilizados na especificação do problema para a geração automática de respostas em problemas aplicáveis ao contexto de disciplinas de introdução à programação (CS1). Desta forma, observa-se que a taxa de acerto para o ChatGPT e Bard são elevadas, tanto para problemas em português ou inglês, enquanto o HuggingChat não alcançou bons resultados. Quanto à estrutura do prompt, nenhuma das aplicações LLM conseguiu acertar a resposta apenas com o enunciado. No entanto, ao fornecer mais informações, como formatação, dicas e casos de teste, geralmente se observa um desempenho melhor do ChatGPT e Bard.

Este estudo aborda a seleção de exemplos trabalhados para Engenharia de Software, com foco na melhoria da qualidade de código no contexto de Sistemas Distribuídos. Propõe-se uma abordagem, que emprega uma heurística baseada em métricas para selecionar exemplos de código que demonstrem melhorias significativas em aspectos como coesão e acoplamento. A análise é realizada em projetos de software livre amplamente reconhecidos, como Apache Kafka e ZooKeeper, utilizando as ferramentas CK, PyDriller e RefactoringMiner. Nossos resultados preliminares sugerem que tal abordagem pode auxiliar na seleção de códigos para a criação de exemplos trabalhados, fornecendo recursos práticos aos educadores, aplicados a domínios complexos como Sistemas Distribuídos. Dessa forma, espera-se contribuir para um aprendizado mais efetivo de Engenharia de Software aplicada naquele domínio.

O objetivo deste trabalho foi avaliar a integração de atividades de programação no ensino de teste de software usando elementos de jogos. Foi executado um estudo sobre implementação de programas e testes de unidade em Java com a ferramenta Code Defenders, avaliando-se o efeito do jogo e a qualidade dos programas. Como resultado, aferiu-se que a qualidade do Code Defender para ensino de teste é boa, destacando-se pelo interesse dos estudantes. Quanto à qualidade dos programas, a maioria dos estudantes conseguiu implementar com êxito os programas especificados.

Background: Gender diversity is an important feature for softwaredevelopment. Studies show that teams with more diversity are moreproductive, too. Objective: This study is a systematic review thatanalyzes women’s motivations to contribute to software development,focusing on a sense of belonging. Method: A protocol wasestablished and the works retrieved by the applied search string.It was analyzed according to the inclusion and exclusion criteria.Results: The results showed that the main factors that motivatewomen to contribute to software projects are Kinship, Career andLearning. However, the sense of belonging, related to the termKinship, was more important in the analysis of groups composedonly of women, compared to heterogeneous groups. Conclusions:Thus, it was found that there are differences in motivations betweengenders, and the sense of belonging is more relevant forwomen than for men. Therefore, to increase diversity in softwareprojects, an important factor to be improved is the belonging of thecontributors.

Regression testing is an important phase to deliver software with quality. However, flaky tests hamper the evaluation of test results and can increase costs. This is because a flaky test may pass or fail non-deterministically and to identify properly the flakiness of a test requires rerunning the test suite multiple times. To cope with this challenge, approaches have been proposed based on prediction models and machine learning. Existing approaches based on the use of the test case vocabulary may be context-sensitive and prone to overfitting, presenting low performance when executed in a cross-project scenario. To overcome these limitations, we investigate the use of test smells as predictors of flaky tests. We conducted an empirical study to understand if test smells have good performance as a classifier to predict the flakiness in the cross-project context, and analyzed the information gain of each test smell. We also compared the test smell-based approach with the vocabulary-based one. As a result, we obtained a classifier that had a reasonable performance (Random Forest, 0.83) to predict the flakiness in the testing phase. This classifier presented better performance than vocabulary-based model for cross-project prediction. The Assertion Roulette and Sleepy Test test smell types are the ones associated with the best information gain values.