Chanchal Roy

Morteza Zakeri-Nasrabadi, Saeed Parsa, Mohammad Ramezani et al.

Measuring and evaluating source code similarity is a fundamental software engineering activity that embraces a broad range of applications, including but not limited to code recommendation, duplicate code, plagiarism, malware, and smell detection. This paper proposes a systematic literature review and meta-analysis on code similarity measurement and evaluation techniques to shed light on the existing approaches and their characteristics in different applications. We initially found over 10000 articles by querying four digital libraries and ended up with 136 primary studies in the field. The studies were classified according to their methodology, programming languages, datasets, tools, and applications. A deep investigation reveals 80 software tools, working with eight different techniques on five application domains. Nearly 49% of the tools work on Java programs and 37% support C and C++, while there is no support for many programming languages. A noteworthy point was the existence of 12 datasets related to source code similarity measurement and duplicate codes, of which only eight datasets were publicly accessible. The lack of reliable datasets, empirical evaluations, hybrid methods, and focuses on multi-paradigm languages are the main challenges in the field. Emerging applications of code similarity measurement concentrate on the development phase in addition to the maintenance.

Lamyea Maha, Tajmilur Rahman, Chanchal Roy

The sprint-based iterative approach in the Agile software development method allows continuous feedback and adaptation. One of the crucial Agile software development activities is the sprint planning session where developers estimate the effort required to complete tasks through a consensus-based estimation technique such as Planning Poker. In the Agile software development method, a common unit of measuring development effort is Story Point (SP) which is assigned to tasks to understand the complexity and development time needed to complete them. Despite the benefits of this process, it is an extremely time-consuming manual process. To mitigate this issue, in this study, we investigated if this manual process can be automated using Retrieval Augmented Generation (RAG) which comprises a "Retriever" and a "Generator". We applied two embedding models - bge-large-en-v1.5, and Sentence-Transformers' all-mpnet-base-v2 on 23 open-source software projects of varying sizes and examined four key aspects: 1) how retrieval hyper-parameters influence the performance, 2) whether estimation accuracy differs across different sizes of the projects, 3) whether embedding model choice affects accuracy, and 4) how the RAG-based approach compares to the existing baselines. Although the RAG-based approach outperformed the baseline models in several occasions, our results did not exhibit statistically significant differences in performance across the projects or across the embedding models. This highlights the need for further studies and refinement of the RAG, and model adaptation strategies for better accuracy in automatically estimating user stories.

Al Muttakin, Saikat Mondal, Chanchal Roy

Replication packages are crucial for enabling transparency, validation, and reuse in software engineering (SE) research. While artifact sharing is now a standard practice and even expected at premier SE venues such as ICSE, the practical usability of these replication packages remain underexplored. In particular, there is a marked lack of studies that comprehensively examine the executability and reproducibility of replication packages in SE research. In this paper, we aim to fill this gap by evaluating 100 replication packages published in ICSE proceedings over the past decade (2015 - 2024). We assess the (1) executability of the replication packages, (2) efforts and modifications required to execute them, (3) challenges that prevent executability, and (4) reproducibility of the original findings for those that are executable. We spent approximately 650 person-hours in total to execute the artifacts and reproduce the study findings. Our analysis shows that only 40 of the 100 evaluated artifacts were fully executable. Among these, 32.5% ran without any modification. However, even executable artifacts required varying levels of effort: 17.5% required low effort, while 82.5% required moderate to high effort to execute successfully. We identified five common types of modifications and 13 challenges that lead to execution failure, encompassing environmental, documentation, and structural issues. Among the executable artifacts, only 35% (14 out of 40) reproduced the original results. These findings highlight a notable gap between artifact availability, executability, and reproducibility. Our study proposes three actionable guidelines to improve the preparation, documentation, and review of research artifacts, thereby strengthening the rigor and sustainability of open science practices in SE research.

Saumendu Roy, Banani Roy, Chanchal Roy et al.

Machine learning (ML)-based defect prediction models can improve software quality. However, their opaque reasoning creates an HCI challenge because developers struggle to trust models they cannot interpret. Explainable AI (XAI) methods such as LIME, SHAP, and BreakDown aim to provide transparency, but when used together, they often produce conflicting explanations that increase confusion, frustration, and cognitive load. To address this usability challenge, we introduce XMENTOR, a human-centered, rank-aware aggregation method implemented as a VS Code plugin. XMENTOR unifies multiple post-hoc explanations into a single, coherent view by applying adaptive thresholding, rank and sign agreement, and fallback strategies to preserve clarity without overwhelming users. In a user study, nearly 90% of the participants preferred aggregated explanations, citing reduced confusion and stronger support for daily tasks of debugging and review of defects. Our findings show how combining explanations and embedding them into developer workflows can enhance interpretability, usability, and trust.

Saumendu Roy, Saikat Mondal, Banani Roy et al.

The lack of interpretability is a major barrier that limits the practical usage of AI models. Several eXplainable AI (XAI) techniques (e.g., SHAP, LIME) have been employed to interpret these models' performance. However, users often face challenges when leveraging these techniques in real-world scenarios and thus submit questions in technical Q&A forums like Stack Overflow (SO) to resolve these challenges. We conducted an exploratory study to expose these challenges, their severity, and features that can make XAI techniques more accessible and easier to use. Our contributions to this study are fourfold. First, we manually analyzed 663 SO questions that discussed challenges related to XAI techniques. Our careful investigation produced a catalog of seven challenges (e.g., disagreement issues). We then analyzed their prevalence and found that model integration and disagreement issues emerged as the most prevalent challenges. Second, we attempt to estimate the severity of each XAI challenge by determining the correlation between challenge types and answer metadata (e.g., the presence of accepted answers). Our analysis suggests that model integration issues is the most severe challenge. Third, we attempt to perceive the severity of these challenges based on practitioners' ability to use XAI techniques effectively in their work. Practitioners' responses suggest that disagreement issues most severely affect the use of XAI techniques. Fourth, we seek agreement from practitioners on improvements or features that could make XAI techniques more accessible and user-friendly. The majority of them suggest consistency in explanations and simplified integration. Our study findings might (a) help to enhance the accessibility and usability of XAI and (b) act as the initial benchmark that can inspire future research.

Rodrigo F. Silva, M. Masudur Rahman, Carlos Eduardo Dantas et al.

Developers often depend on code search engines to obtain solutions for their programming tasks. However, finding an expected solution containing code examples along with their explanations is challenging due to several issues. There is a vocabulary mismatch between the search keywords (the query) and the appropriate solutions. Semantic gap may increase for similar bag of words due to antonyms and negation. Moreover, documents retrieved by search engines might not contain solutions containing both code examples and their explanations. So, we propose CRAR (Crowd Answer Recommender) to circumvent those issues aiming at improving retrieval of relevant answers from Stack Overflow containing not only the expected code examples for the given task but also their explanations. Given a programming task, we investigate the effectiveness of combining information retrieval techniques along with a set of features to enhance the ranking of important threads (i.e., the units containing questions along with their answers) for the given task and then selects relevant answers contained in those threads, including semantic features, like word embeddings and sentence embeddings, for instance, a Convolutional Neural Network (CNN). CRAR also leverages social aspects of Stack Overflow discussions like popularity to select relevant answers for the tasks. Our experimental evaluation shows that the combination of the different features performs better than each one individually. We also compare the retrieval performance with the state-of-art CROKAGE (Crowd Knowledge Answer Generator), which is also a system aimed at retrieving relevant answers from Stack Overflow. We show that CRAR outperforms CROKAGE in Mean Reciprocal Rank and Mean Recall with small and medium effect sizes, respectively.

Golam Mostaeen, Banani Roy, Chanchal Roy et al.

A code clone is a pair of code fragments, within or between software systems that are similar. Since code clones often negatively impact the maintainability of a software system, several code clone detection techniques and tools have been proposed and studied over the last decade. To detect all possible similar source code patterns in general, the clone detection tools work on the syntax level while lacking user-specific preferences. This often means the clones must be manually inspected before analysis in order to remove those false positives from consideration. This manual clone validation effort is very time-consuming and often error-prone, in particular for large-scale clone detection. In this paper, we propose a machine learning approach for automating the validation process. Our machine learning-based approach is used to automatically validate clones without human inspection. Thus the proposed approach can be used to remove the false positive clones from the detection results, automatically evaluate the precision of any clone detectors for any given set of datasets, evaluate existing clone benchmark datasets, or even be used to build new clone benchmarks and datasets with minimum effort. In an experiment with clones detected by several clone detectors in several different software systems, we found our approach has an accuracy of up to 87.4% when compared against the manual validation by multiple expert judges. The proposed method also shows better results in several comparative studies with the existing related approaches for clone classification.