CodeReviewQA: The Code Review Comprehension Assessment for Large Language Models

State-of-the-art large language models (LLMs) have demonstrated impressive code generation capabilities but struggle with real-world software engineering tasks, such as revising source code to address code reviews, hindering their practical use. Code review comments are often implicit, ambiguous, and colloquial, requiring models to grasp both code and human intent. This challenge calls for evaluating large language models' ability to bridge both technical and conversational contexts. While existing work has employed the automated code refinement (ACR) task to resolve these comments, current evaluation methods fall short, relying on text matching metrics that provide limited insight into model failures and remain susceptible to training data contamination. To address these limitations, we introduce a novel evaluation benchmark, $\textbf{CodeReviewQA}$ that enables us to conduct fine-grained assessment of model capabilities and mitigate data contamination risks. In CodeReviewQA, we decompose the generation task of code refinement into $\textbf{three essential reasoning steps}$: $\textit{change type recognition}$ (CTR), $\textit{change localisation}$ (CL), and $\textit{solution identification}$ (SI). Each step is reformulated as multiple-choice questions with varied difficulty levels, enabling precise assessment of model capabilities, while mitigating data contamination risks. Our comprehensive evaluation spans 72 recently released large language models on $\textbf{900 manually curated, high-quality examples}$ across nine programming languages. Our results show that CodeReviewQA is able to expose specific model weaknesses in code review comprehension, disentangled from their generative automated code refinement results.

Enhancing High-Quality Code Generation in Large Language Models with Comparative Prefix-Tuning

Large Language Models (LLMs) have been widely adopted in commercial code completion engines, significantly enhancing coding efficiency and productivity. However, LLMs may generate code with quality issues that violate coding standards and best practices, such as poor code style and maintainability, even when the code is functionally correct. This necessitates additional effort from developers to improve the code, potentially negating the efficiency gains provided by LLMs. To address this problem, we propose a novel comparative prefix-tuning method for controllable high-quality code generation. Our method introduces a single, property-specific prefix that is prepended to the activations of the LLM, serving as a lightweight alternative to fine-tuning. Unlike existing methods that require training multiple prefixes, our approach trains only one prefix and leverages pairs of high-quality and low-quality code samples, introducing a sequence-level ranking loss to guide the model's training. This comparative approach enables the model to better understand the differences between high-quality and low-quality code, focusing on aspects that impact code quality. Additionally, we design a data construction pipeline to collect and annotate pairs of high-quality and low-quality code, facilitating effective training. Extensive experiments on the Code Llama 7B model demonstrate that our method improves code quality by over 100% in certain task categories, while maintaining functional correctness. We also conduct ablation studies and generalization experiments, confirming the effectiveness of our method's components and its strong generalization capability.

Junior Software Developers' Perspectives on Adopting LLMs for Software Engineering: a Systematic Literature Review

Many studies exploring the adoption of Large Language Model-based tools for software development by junior developers have emerged in recent years. These studies have sought to understand developers' perspectives about using those tools, a fundamental pillar for successfully adopting LLM-based tools in Software Engineering. The aim of this paper is to provide an overview of junior software developers' perspectives and use of LLM-based tools for software engineering (LLM4SE). We conducted a systematic literature review (SLR) following guidelines by Kitchenham et al. on 56 primary studies, applying the definition for junior software developers as software developers with equal or less than five years of experience, including Computer Science/Software Engineering students. We found that the majority of the studies focused on comprehending the different aspects of integrating AI tools in SE. Only 8.9\% of the studies provide a clear definition for junior software developers, and there is no uniformity. Searching for relevant information is the most common task using LLM tools. ChatGPT was the most common LLM tool present in the studies (and experiments). A majority of the studies (83.9\%) report both positive and negative perceptions about the impact of adopting LLM tools. We also found and categorised advantages, challenges, and recommendations regarding LLM adoption. Our results indicate that developers are using LLMs not just for code generation, but also to improve their development skills. Critically, they are not just experiencing the benefits of adopting LLM tools, but they are also aware of at least a few LLM limitations, such as the generation of wrong suggestions, potential data leaking, and AI hallucination. Our findings offer implications for software engineering researchers, educators, and developers.

Generative AI and Empirical Software Engineering: A Paradigm Shift

The widespread adoption of generative AI in software engineering marks a paradigm shift, offering new opportunities to design and utilize software engineering tools while influencing both developers and the artifacts they create. Traditional empirical methods in software engineering, including quantitative, qualitative, and mixed-method approaches, are well established. However, this paradigm shift introduces novel data types and redefines many concepts in the software engineering process. The roles of developers, users, agents, and researchers increasingly overlap, blurring the distinctions between these social and technical actors within the field. This paper examines how integrating AI into software engineering challenges traditional research paradigms. It focuses on the research phenomena that we investigate, the methods and theories that we employ, the data we analyze, and the threats to validity that emerge in this new context. Through this exploration, our goal is to understand how AI adoption disrupts established software development practices that creates new opportunities for empirical software engineering research.

How Developers Interact with AI: A Taxonomy of Human-AI Collaboration in Software Engineering

Artificial intelligence (AI), including large language models and generative AI, is emerging as a significant force in software development, offering developers powerful tools that span the entire development lifecycle. Although software engineering research has extensively studied AI tools in software development, the specific types of interactions between developers and these AI-powered tools have only recently begun to receive attention. Understanding and improving these interactions has the potential to improve productivity, trust, and efficiency in AI-driven workflows. In this paper, we propose a taxonomy of interaction types between developers and AI tools, identifying eleven distinct interaction types, such as auto-complete code suggestions, command-driven actions, and conversational assistance. Building on this taxonomy, we outline a research agenda focused on optimizing AI interactions, improving developer control, and addressing trust and usability challenges in AI-assisted development. By establishing a structured foundation for studying developer-AI interactions, this paper aims to stimulate research on creating more effective, adaptive AI tools for software development.

Can LLMs Replace Manual Annotation of Software Engineering Artifacts?

Experimental evaluations of software engineering innovations, e.g., tools and processes, often include human-subject studies as a component of a multi-pronged strategy to obtain greater generalizability of the findings. However, human-subject studies in our field are challenging, due to the cost and difficulty of finding and employing suitable subjects, ideally, professional programmers with varying degrees of experience. Meanwhile, large language models (LLMs) have recently started to demonstrate human-level performance in several areas. This paper explores the possibility of substituting costly human subjects with much cheaper LLM queries in evaluations of code and code-related artifacts. We study this idea by applying six state-of-the-art LLMs to ten annotation tasks from five datasets created by prior work, such as judging the accuracy of a natural language summary of a method or deciding whether a code change fixes a static analysis warning. Our results show that replacing some human annotation effort with LLMs can produce inter-rater agreements equal or close to human-rater agreement. To help decide when and how to use LLMs in human-subject studies, we propose model-model agreement as a predictor of whether a given task is suitable for LLMs at all, and model confidence as a means to select specific samples where LLMs can safely replace human annotators. Overall, our work is the first step toward mixed human-LLM evaluations in software engineering.

The Impact Of Bug Localization Based on Crash Report Mining: A Developers' Perspective

Developers often use crash reports to understand the root cause of bugs. However, locating the buggy source code snippet from such information is a challenging task, mainly when the log database contains many crash reports. To mitigate this issue, recent research has proposed and evaluated approaches for grouping crash report data and using stack trace information to locate bugs. The effectiveness of such approaches has been evaluated by mainly comparing the candidate buggy code snippets with the actual changed code in bug-fix commits -- which happens in the context of retrospective repository mining studies. Therefore, the existing literature still lacks discussing the use of such approaches in the daily life of a software company, which could explain the developers' perceptions on the use of these approaches. In this paper, we report our experience of using an approach for grouping crash reports and finding buggy code on a weekly basis for 18 months, within three development teams in a software company. We grouped over 750,000 crash reports, opened over 130 issues, and collected feedback from 18 developers and team leaders. Among other results, we observe that the amount of system logs related to a crash report group is not the only criteria developers use to choose a candidate bug to be analyzed. Instead, other factors were considered, such as the need to deliver customer-prioritized features and the difficulty of solving complex crash reports (e.g., architectural debts), to cite some. The approach investigated in this study correctly suggested the buggy file most of the time -- the approach's precision was around 80%. In this study, the developers also shared their perspectives on the usefulness of the suspicious files and methods extracted from crash reports to fix related bugs.

Stop Words for Processing Software Engineering Documents: Do they Matter?

Stop words, which are considered non-predictive, are often eliminated in natural language processing tasks. However, the definition of uninformative vocabulary is vague, so most algorithms use general knowledge-based stop lists to remove stop words. There is an ongoing debate among academics about the usefulness of stop word elimination, especially in domain-specific settings. In this work, we investigate the usefulness of stop word removal in a software engineering context. To do this, we replicate and experiment with three software engineering research tools from related work. Additionally, we construct a corpus of software engineering domain-related text from 10,000 Stack Overflow questions and identify 200 domain-specific stop words using traditional information-theoretic methods. Our results show that the use of domain-specific stop words significantly improved the performance of research tools compared to the use of a general stop list and that 17 out of 19 evaluation measures showed better performance.

She Elicits Requirements and He Tests: Software Engineering Gender Bias in Large Language Models

Implicit gender bias in software development is a well-documented issue, such as the association of technical roles with men. To address this bias, it is important to understand it in more detail. This study uses data mining techniques to investigate the extent to which 56 tasks related to software development, such as assigning GitHub issues and testing, are affected by implicit gender bias embedded in large language models. We systematically translated each task from English into a genderless language and back, and investigated the pronouns associated with each task. Based on translating each task 100 times in different permutations, we identify a significant disparity in the gendered pronoun associations with different tasks. Specifically, requirements elicitation was associated with the pronoun "he" in only 6% of cases, while testing was associated with "he" in 100% of cases. Additionally, tasks related to helping others had a 91% association with "he" while the same association for tasks related to asking coworkers was only 52%. These findings reveal a clear pattern of gender bias related to software development tasks and have important implications for addressing this issue both in the training of large language models and in broader society.

Socialz: Multi-Feature Social Fuzz Testing

Online social networks have become an integral aspect of our daily lives and play a crucial role in shaping our relationships with others. However, bugs and glitches, even minor ones, can cause anything from frustrating problems to serious data leaks that can have far-reaching impacts on millions of users. To mitigate these risks, fuzz testing, a method of testing with randomised inputs, can provide increased confidence in the correct functioning of a social network. However, implementing traditional fuzz testing methods can be prohibitively difficult or impractical for programmers outside of the network's development team. To tackle this challenge, we present Socialz, a novel approach to social fuzz testing that (1) characterises real users of a social network, (2) diversifies their interaction using evolutionary computation across multiple, non-trivial features, and (3) collects performance data as these interactions are executed. With Socialz, we aim to provide anyone with the capability to perform comprehensive social testing, thereby improving the reliability and security of online social networks used around the world.