[Context] In Brazil, 41% of companies use machine learning (ML) to some extent. However, several challenges have been reported when engineering ML-enabled systems, including unrealistic customer expectations and vagueness in ML problem specifications. Literature suggests that Requirements Engineering (RE) practices and tools may help to alleviate these issues, yet there is insufficient understanding of RE’s practical application and its perception among practitioners. [Goal] This study aims to investigate the application of RE in developing ML-enabled systems in Brazil, creating an overview of current practices, perceptions, and problems in the Brazilian industry. [Method] To this end, we extracted and analyzed data from an international survey focused on ML-enabled systems, concentrating specifically on responses from practitioners based in Brazil. We analyzed the cluster of RE-related answers gathered from 72 practitioners involved in data-driven projects. We conducted quantitative statistical analyses on contemporary practices using bootstrapping with confidence intervals and qualitative studies on the reported problems involving open and axial coding procedures. [Results] Our findings highlight distinct RE implementation aspects in Brazil’s ML projects. For instance, (i) RE-related tasks are predominantly conducted by data scientists; (ii) the most common techniques for eliciting requirements are interviews and workshop meetings; (iii) there is a prevalence of interactive notebooks in requirements documentation; (iv) practitioners report problems that include a poor understanding of the problem to solve and the business domain, low customer engagement, and difficulties managing stakeholders expectations. Our analysis suggests that development methodology plays a role in these challenges. Agile methods appear to facilitate the management of customer expectations compared to traditional approaches; however, they also appear to introduce greater difficulties in problem understanding and customer involvement. [Conclusion] These results provide an understanding of RE-related practices and challenges in the Brazilian ML industry, helping to guide research and initiatives toward improving the maturity of RE for ML-enabled system projects.

Developing the necessary skills in Software Engineering students to conduct effective requirements elicitation interviews is a complex challenge. Immersive role playing has emerged as a promising educational strategy, enabling students to simulate realistic interviews, receive real-time feedback, and improve their performance under pressure. This approach blends traditional role playing with immersive learning environments, providing engaging and authentic experiences that better prepare students for industry demands. This article presents an experience report on the use of an immersive role playing to teach the interview technique for requirements elicitation. Conducted with 86 undergraduate students enrolled in a Requirements Engineering course, the study offers a broader perspective on the effectiveness and challenges of this approach. The findings suggest that immersive activities foster reflection, help identify areas for improvement, and emphasize the importance of emotional regulation in real-world interactions. These insights reinforce that mastering Requirements Engineering requires not only technical proficiency, but also strong interpersonal and emotional skills.

Software development is an activity characterized by the continuous search for information by programmers. General-purpose search engines, such as Google, Bing, and Yahoo, are widely used by programmers to find solutions to their problems. However, the best solutions are not always among the initial pages of search results that are ranked according to the search engine's algorithms. This study aims to investigate the influence that the order of the pages returned by the ranking of search engines exerts on programmers' performance when performing programming tasks. We designed an empirical within-subject study with programmers to understand and evaluate their performance when solving programming tasks using a ranked list of pages returned by the Google search engine and artificially modified with two different ranking quality levels (Higher Quality Ranking and Lower Quality Ranking). Moreover, for each entry in the ranking of pages, the most frequent methods mentioned on the respective page were listed in the ranking visualization. Analysis of participants' recorded videos was conducted through a mixed-methods research approach to provide insights into the results. We found that programmers spent approximately 8 minutes longer resolving tasks associated with a Lower Quality Ranking, spending more time on irrelevant pages compared to relevant ones, due to efforts to fix problematic code or new searches for another page. The addition of a list of frequent methods in the ranking visualization could help programmers to skip irrelevant pages and reduce time wastage. The ranking quality influences the programmers' performance during the development of programming tasks. Therefore, we suggest the development of filters aimed at improving the quality of results delivered by search engines. Moreover, the results may encourage the adaptation of this study for other approaches that require information foraging, such as chatting with LLMs.

Code smells are indicators of poor design implementation and decision-making that can potentially harm the quality of software. Therefore, detecting these smells is crucial to prevent such issues. Some studies aim to comprehend the impact of code smells on software quality, while others propose rules or machine learning-based approaches to identify code smells. Previous research has focused on labeling and analyzing code snippets that significantly impair software quality using machine learning techniques. These snippets are classified as Clean, Smelly, Buggy, and Harmful Code. Harmful Code refers to smelly code segments that have one or more reported bugs, whether fixed or not. Consequently, the presence of a Harmful Code increases the risk of introducing new defects and/or design issues during the remediation process. We perform our study as an extension of the previous study, with the scope of 5 smell types. The total number of commits across all four tables (Java, C++, C#, and Python projects) is 641,736, versions of 91 open-source projects, 17,022 bugs, and 24,737 code smells. The findings revealed promising transferability of knowledge between Java and C# in the presence of various code smell types, while C++ and Python exhibited more challenging transferability. Also, our study discovered that a sample size of 32 demonstrated favorable outcomes for most harmful codes, underscoring the efficiency of transfer learning even with limited data. Moreover, the exploration of transfer learning between bugs and code smells represents a not-very-ineffective avenue within the realm of software engineering.

Application Programming Interfaces (APIs) are everyday tools for every software professional. When creating an API, a designer typically abstracts the tasks the API intends to perform. The API user, on the other hand, should use the API to perform specific-propose tasks. During the development of an API, the designer needs to write dialogues with which the user will interact with the API, performing a conversation between them. In this way, an API capable of offering effective dialogues to its users is called a conversational API. A conversational API is the one that is capable of communicating to its users its form of use and its internal logic of operation, making it clear the design decisions when abstracting concepts and tasks. In this paper, we present Colloquy, an evidence-based method to support the design of conversational APIs. Colloquy was conceived based on the lessons learned from an action research we conducted to identify effective strategies for designing conversational APIs. For six months, we followed the redesign of an existing API developed at an R&D laboratory of a large IT company. In this paper, we also present the first empirical evaluation of Colloquy. We conducted a case study in which the method was employed to design an API for refactoring source code in the Java programming language. Among other benefits, we found that Colloquy was effective in developing empathy with users and modeling conversations with them. Besides, Colloquy contributed to identifying new requirements and created models as API documentation.

Peer-to-peer (P2P) systems, characterized by their decentralized architecture, have become crucial to various applications such as file sharing, content distribution, and blockchain technologies. Due to their dynamic and heterogeneous nature, ensuring their reliability and robustness presents challenges. Such an issue has motivated researchers to develop testing techniques for these systems. Despite advances in research about software testing for distributed systems, there is a lack of data about their applicability and which features from P2P systems are highlighted by the proposed testing approaches. This paper provides a comprehensive overview of the state-of-the-art software testing for P2P systems. We systematically examine the inherent challenges and situations. Our findings show that some aspects of P2P systems can make the software testing activity difficult, such as validation of global properties and consensus, peer churn, a large volume of data, malformed data, reproducibility, and non-determinism. Several testing techniques have been used in this context, such as model-based testing, fault injection, fuzzing, stress testing, differential testing, conformance testing, and concurrency testing. We highlight fuzzing for testing malformed data, global properties, parallel events, and model-based testing as a powerful technique to abstract the complexities and varieties of P2P applications during the testing activities. Moreover, blockchain applications are the most addressed P2P application explored by the studies, where most papers were published in 2023, highlighting the theme’s relevance.

Technical debt is a concept used to describe the lack of good practices during software development, leading to several problems and costs. Identification and management strategies can help reduce these difficulties. In a previous study, Alves et al. (2016) analyzed the research landscape of such strategies from 2010 to 2014. This paper replicates and updates their study to explore the evolution of technical debt identification and management research landscape over a decade, including literature from 2010 until 2022. We analyzed 117 papers from the ACM Digital Library, IEEE Xplore, Science Direct, and Springer Link. Newly suggested strategies include automatically identifying admitted debt in comments, commits, and source code. Between 2015 and 2022, more empirical evaluations have been performed, and the general research focus has changed to a more holistic approach. Therefore, the research area evolved and reached a new level of maturity compared to previous results from Alves et al. (2016). Not only are code aspects considered for technical debt, but other aspects have also been investigated (e.g., models for the development process).

In recent years, the use and investigations about Grey Literature (GL) increased, in particular, in Software Engineering (SE) research. However, its understanding is still scarce and sometimes controversial, such as interpreting GL types and assessing their credibility. This study aimed to understand the credibility aspects that SE researchers consider in assessing GL and its types. To achieve this goal, we surveyed 53 SE researchers (who answered that they have used GL in our previous investigation), receiving a total of 34 valid responses. Our main findings show that: 1) GL source produced or cited by a renowned source is the main credibility criteria used to assess GL, 2) most of the GL types tend to have a Low to Moderate level of Control and Expertise, 3) there is a positive statistical correlation between the level of Control and Expertise for most GL types, and 4) the different respondent profiles shared similar opinions about the credibility criteria. Our investigation contributes to helping future SE researchers that intend to use GL with more credibility. Additionally, shows the need for future studies to better understand the GL types in SE research.

Search-based algorithms have been successfully employed in Product Line Architecture (PLA) design in the seminal approach named Multi-Objective Approach for Product-Line Architecture Design (MOA4PLA). This approach generates a set of alternative PLA designs, which optimize different architectural properties. In addition to these properties, the alternative PLA designs should have as few architectural smells as possible. Architectural smells can negatively impact PLA variability, PLA extensibility, SPL maintainability, and other non-functional attributes. However, one of the main findings of a previous study is that the tool that automates the application of MOA4PLA adversely introduces architectural smells in the automatically generated solutions. In this work, we present OPLA-Tool-ASP, which is a tool that implements guidelines to detect and prevent the architectural smells Unused Interface, Unused Brick, Concern Overload, and Link Overload in the context of MOA4PLA. An empirical study was carried out to assess the effectiveness of OPLA-Tool-ASP in preventing the aforementioned smells in the resulting PLA designs. The obtained results pointed out that the proposed tool is effective in both preventing the smells and improving the architectural properties selected for optimization.

Communication is one of the leading challenges faced by teams working in a distributed setting; yet, little has been theorized about how communication occurs in such context. Our long-term research goal is to construct a Communication Theory in Distributed Software Development, aiming to propose a theoretical foundation for future academic studies on the topic of communication and a reference for industry practitioners. To achieve this goal, we are using Grounded Theory, including an Exploratory Literature Review before the theory construction, to confirm the research gap. In this paper, we present a further preliminary version of the Communication Theory comprising six theoretical categories and 31 subcategories. The theory brings, up to know, a consolidated body of knowledge and points out the main concepts that define what communication is in distributed software teams.