Code Changes Research Articles

Explaining Explanations: An Empirical Study of Explanations in Code Reviews

Code reviews are central for software quality assurance. Ideally, reviewers should explain their feedback to enable authors of code changes to understand the feedback and act accordingly. Different developers might need different explanations in different contexts. Therefore, assisting this process first requires understanding the types of explanations reviewers usually provide. The goal of this paper is to study the types of explanations used in code reviews and explore the potential of Large Language Models (LLMs), specifically ChatGPT, in generating these specific types. We extracted 793 code review comments from Gerrit and manually labeled them based on whether they contained a suggestion, an explanation, or both. Our analysis shows that 42% of comments only include suggestions without explanations. We categorized the explanations into seven distinct types including rule or principle, similar examples, and future implications. When measuring their prevalence, we observed that some explanations are used differently by novice and experienced reviewers. Our manual evaluation shows that, when the explanation type is specified, ChatGPT can correctly generate the explanation in 88 out of 90 cases. This foundational work highlights the potential for future automation in code reviews, which can assist developers in sharing and obtaining different types of explanations as needed, thereby reducing back-and-forth communication.

Multiple independent genetic code reassignments of the UAG stop codon in phyllopharyngean ciliates.

The translation of nucleotide sequences into amino acid sequences, governed by the genetic code, is one of the most conserved features of molecular biology. The standard genetic code, which uses 61 sense codons to encode one of the 20 standard amino acids and 3 stop codons (UAA, UAG, and UGA) to terminate translation, is used by most extant organisms. The protistan phylum Ciliophora (the 'ciliates') are the most prominent exception to this norm, exhibiting the grfeatest diversity of nuclear genetic code variants and evidence of repeated changes in the code. In this study, we report the discovery of multiple independent genetic code changes within the Phyllopharyngea class of ciliates. By mining publicly available ciliate genome datasets, we discovered that three ciliate species from the TARA Oceans eukaryotic metagenome dataset use the UAG codon to putatively encode leucine. We identified novel suppressor tRNA genes in two of these genomes which are predicted to decode the reassigned UAG codon to leucine. Phylogenomics analysis revealed that these three uncultivated taxa form a monophyletic lineage within the Phyllopharyngea class. Expanding our analysis by reassembling published phyllopharyngean genome datasets led to the discovery that the UAG codon had also been reassigned to putatively code for glutamine in Hartmannula sinica and Trochilia petrani. Phylogenomics analysis suggests that this occurred via two independent genetic code change events. These data demonstrate that the reassigned UAG codons have widespread usage as sense codons within the phyllopharyngean ciliates. Furthermore, we show that the function of UAA is firmly fixed as the preferred stop codon. These findings shed light on the evolvability of the genetic code in understudied microbial eukaryotes.

CPT 2025: Reporting a Brief Communication Technology-Based Service

This article includes an expanded discussion of new code 98016 (brief communication technology-based evaluation and management [E/M] service) that was introduced in the first article in this series, “CPT 2025: An Overview of Telemedicine Code Changes.” Topics covered in this article are as follows: Code 98016 and similar codesAssigned relative value units (RVUs)Essentials of the brief communication technology-based E/M serviceExamples: coding 98016 services

CPT 2025: Audio-Video Telemedicine Visits

This continues a series of articles on telemedicine code changes that will be implemented for services provided on and after January 1, 2025. This installment focuses on a subcategory of codes for telemedicine evaluation and management (E/M) services provided via synchronous audio-video communications technology. Topics include the following: Review of audio-video telemedicine codes 98000–98007Assigned relative value units (RVUs)General instructions for reporting 98000–98007Coding examples for audio-video telemedicine services

Mining Fine-Grained Code Change Patterns Using Multiple Feature Analysis

Maintaining high code quality is a crucial concern in software development. Existing studies demonstrated that developers frequently face recurrent bugs and adopt similar fix measures, known as code change patterns. As an essential static analysis technique, code pattern mining supports various tasks, including code refactoring, automated program repair, and defect prediction, thus significantly improving software development processes. A prevalent approach to identifying code patterns involves translating code changes to edit actions into a Bag-of-Words (BoW) model. However, when applied to open-source projects, this method exhibits several limitations. For instance, it overlooks function call information and disregards feature word order. This study introduces MIFA, a novel technique for mining code change patterns using multiple feature analysis. MIFA extends existing BoW methods by incorporating analysis of function calls and overall changes in the Abstract Syntax Tree (AST) structure. We selected 20 popular Python projects and evaluated MIFA in both intra-project and cross-project scenarios. The experimental results indicate that: (1) MIFA achieved higher silhouette coefficients and F1 scores compared to other state-of-the-art methods, demonstrating a superior accuracy; (2) MIFA can assist developers in detecting unique change patterns more earlier, with an efficiency improvement of over 40% compared to random sampling. Additionally, we discussed critical parameters for measuring the similarity of code changes, guiding users to apply our method effectively.

Continuous Integration, Continuous Delivery Pipeline with AI Integration

In today’s fast-paced tech world, dropping apps quickly while keeping them top-tier is non-negotiable. This project dives into fixing the grind of traditional, error-prone software development by rolling out a Continuous Integration and Continuous Deployment (CI/CD) pipeline on Amazon Web Services (AWS). The idea is simple: ditch the old-school manual hustle and use automation tools to totally revamp the software game. From coding to deployment, everything gets faster, smarter, and way more reliable—no more messy releases or unnecessary delays. Back in the day, manual processes were a total buzzkill for development teams. Code integration took forever, testing was a headache, and deployments were straight-up chaos, leading to bugs and downtime. As apps got bigger and more complex, these issues only stacked up, slowing down the entire system. Our project uses AWS tools like CodePipeline, CodeBuild, and CodeDeploy to sort this out. These tools handle code integration, testing, and deployment on autopilot, speeding up the delivery game while cutting down on human errors. Plus, QA teams can breathe easy with less grunt work, and developers can focus on creating cool features instead of firefighting bugs. On top of being faster and more accurate, this pipeline levels up teamwork with instant feedback on code changes. Developers can squash bugs on the fly, and everyone from project managers to stakeholders stays in the loop about what’s going down. This creates a vibe of constant improvement—each update builds on the last, making the software sharper every time. Thanks to AWS’s scalable setup, handling bigger projects or adding more team members is no biggie. It’s like having a system that grows with you without ever feeling clunky. And let’s not forget how this aligns perfectly with modern DevOps goals. By automating repetitive and boring tasks, teams can focus on the stuff that actually matters, like innovation and problem-solving. With AWS’s rock-solid tools, security and performance stay on point, no matter how hectic things get. This project is proof that tech done right can help teams stay ahead of the curve and deliver epic software that meets today’s high standards. Key Words: AWS CI/CD Pipeline, AWS CodeCommit, AWS CodeBuild, AWS CodeDeploy, AWS CodePipeline, Amazon Elastic Beanstalk

Rarely categorical, always high-dimensional: how the neural code changes along the cortical hierarchy.

The brain is highly structured both at anatomical and functional levels. However, within individual brain areas, neurons often exhibit very diverse and seemingly disorganized responses. A more careful analysis shows that these neurons can sometimes be grouped together into specialized subpopulations (categorical representations). Organization can also be found at the level of the representational geometry in the activity space, typically in the form of low-dimensional structures. It is still unclear how the geometry in the activity space and the structure of the response profiles of individual neurons are related. Here, we systematically analyzed the geometric and selectivity structure of the neural population from 40+ cortical regions in mice performing a decision-making task (IBL public Brainwide Map data set). We used a reduced-rank regression approach to quantify the selectivity profiles of single neurons and multiple measures of dimensionality to characterize the representational geometry of task variables. We then related these measures within single areas to the position of each area in the sensory-cognitive cortical hierarchy. Our findings reveal that only a few regions (in primary sensory areas) are categorical. When multiple brain areas are considered, we observe clustering that reflects the brain's large-scale organization. The representational geometry of task variables also changed along the cortical hierarchy, with higher dimensionality in cognitive regions. These trends were explained by analytical computations linking the maximum dimensionality of representational geometry to the clustering of selectivity at the single neuron level. Finally, we computed the shattering dimensionality (SD), a measure of the linear separability of neural activity vectors; remarkably, the SD remained near maximal across all regions, suggesting that the observed variability in the selectivity profiles allows neural populations to maintain high computational flexibility. These results provide a new mathematical and empirical perspective on selectivity and representation geometry in the cortical neural code.

Indoor radon trends with building code change in two Canadian cities

Radon studies were conducted in two Canadian cities, in Halifax, Nova Scotia, and Winnipeg, Manitoba, to evaluate trends in indoor radon before and after the 2010 National Building Code of Canada was adopted into the legally binding provincial building codes in 2011. Participants were recruited in neighbourhoods characterized by newer housing developments. A postcard campaign in each city offered free radon testing to every house in the target areas, and free testing kits were mailed to study participants. Indoor radon measurements and house questionnaires were completed by 225 eligible participants in Halifax and 171 eligible participants in Winnipeg, using alpha-track radon detectors deployed for three months during the heating season in 2021–2022. Multivariate logistic regression analyses were conducted to evaluate the association between indoor radon and the period of construction, the area in contact with the ground, the number of storeys, the type of heating system, the water source, and the type of ventilation. These analyses were focussed on the detached study houses because the majority of the participants lived in a detached house, and significant associations were identified for the period of construction and the floor where the radon detector was located. An odds ratio of 1.91 (1.04–3.50) for the detached Halifax study houses built after 2011 was associated with having a higher than geometric mean radon concentration (p = 0.033), nearly double the likelihood. There was no evidence of significant change in the indoor radon after 2011 in the detached Winnipeg study houses. A lower likelihood – almost half - for measurement conducted on a main/upper floor compared to in the basement was associated with a radon concentration above the respective geometric mean for each city: an odds ratio of 0.48 (0.27–0.86) for the detached Halifax study houses (p = 0.012), and of 0.45 (0.32–0.64) for the detached Winnipeg study houses (p = 0.022). Radon is the second most important cause of lung cancer, after smoking, and the results of this study support strengthening the radon preventive measures required in new low-rise housing to reduce the associated lung cancer burden in Canada.

Vehicular Traffic Flow Detection and Monitoring for Implementation of Smart Traffic Light: A Case Study for Road Intersection in Limeira, Brazil

This paper proposes the development of a smart traffic light prototype based on vehicular traffic flow measurement in the stretch between two avenues in the city of Limeira, SP, Brazil, focusing on the stretch towards UNICAMP’s School of Technology. To this end, we initially developed a Python code using the OpenCV library in order to detect and count vehicles. With the counting in operation, programming logic was inserted, aiming at preparing traffic light timers based on vehicular traffic. Finally, the traffic lights were added to display video via a code change to show the ongoing color changes, also obtaining a code for identifying vehicles and flow, in addition to the virtual traffic light system itself in the system. Vehicle counting accuracy was 75% for large vehicles, 90% for passenger cars, and 100% for motorcycles. The simulation of a smart traffic light implementation worked satisfactorily according to the post-processing of the video recorded for validation.

Towards Better Comprehension of Breaking Changes in the NPM Ecosystem

Code evolution is prevalent in software ecosystems, which can provide many benefits, such as new features, bug fixes, security patches, etc., while still introducing breaking changes that make downstream projects fail to work. Breaking changes cause a lot of effort to both downstream and upstream developers: downstream developers need to adapt to breaking changes and upstream developers are responsible for identifying and documenting them. In the NPM ecosystem, characterized by frequent code changes and a high tolerance for making breaking changes, the effort is larger. For better comprehension of breaking changes in the NPM ecosystem and to enhance breaking change detection tools, we conduct a large-scale empirical study to investigate breaking changes in the NPM ecosystem. We construct a dataset of explicitly documented breaking changes from 381 popular NPM projects. We find that 95.4% of the detected breaking changes can be covered by developers’ documentation, and 19% of the breaking changes cannot be detected by regression testing. Then in the process of investigating source code of our collected breaking changes, we yield a taxonomy of JavaScript and TypeScript-specific syntactic breaking changes and a taxonomy of major types of behavioral breaking changes. Additionally, we investigate the reasons why developers make breaking changes in NPM and find three major reasons, i.e., to reduce code redundancy, to improve identifier names, and to improve API design, and each category contains several sub-items. We provide actionable implications for future research, e.g., automatic naming and renaming techniques should be applied in JavaScript projects to improve identifier names, future research can try to detect more types of behavioral breaking changes. By presenting the implications, we also discuss the weakness of automatic renaming and breaking change detection approaches, such as the lack of support for public identifiers and various types of breaking changes.

Quantifying Plastic and Reconstructive Surgery Engagement in the Evolution of ICD-10 Codes.

International Classification of Diseases (ICD)-10 code specificity is important for detailed diagnostic documentation. Insufficient coding granularity in plastic surgery may hinder accurate clinical documentation, impacting education, research, and patient care. This study examines ICD-10 code modification efforts by plastic surgeons and plastic surgery organizations within the last decade, subsequently detailing the formal proposal process and highlighting potential areas where coding modifications may be explored. A retrospective review of all modification proposals presented to the ICD Coordination and Maintenance Committee from March 2013 to March 2023 was conducted. Plastic surgery-related proposals were identified and evaluated for successful implementation. These results were compared to those of other specialties using Fisher exact tests and 2-tailed t tests. A total of 472 unique proposals were identified and reviewed. Surgeons and/or surgical organizations contributed to 111 (23.5%) novel proposals, 41 (36.9%) of which were initiated by surgeons. Nine surgical specialties and 14 supporting organizations were represented. The success rate for all surgical-related proposals was 85.6%, with an average time to implementation of 24.7 months. Only 2 (1.8%) of all surgical-related proposals pertained to plastic surgery, both of which were successfully implemented. Though a few plastic surgeons have taken advantage of the ICD code revision process within the last decade, the high overall success rate for surgical-related proposals suggests that this seems to be a feasible method by which plastic surgeons and plastic surgery organizations can bring about useful coding changes that meaningfully impact clinical practice.

CPT 2025: An Overview of Telemedicine Code Changes

This is the first in a series of articles discussing Current Procedural Terminology (CPT®) changes that will be implemented for reporting telemedicine evaluation and management (E/M) services performed by physicians and other qualified health care professionals (QHPs) on and after January 1, 2025, with emphasis on the following topics: Background for telemedicine E/M code changes Defining telemedicine in procedure coding Telemedicine services to new and established patients Audio-video telemedicine codes Audio-only telemedicine codes A brief virtual check-in code

Speech perception and language comprehension in primary progressive aphasia

Primary progressive aphasia (PPA) is a neurodegenerative disorder characterized by progressive loss of speech and language. Although speech perception and language comprehension deficits are observed in individuals with PPA, these deficits have been understudied relative to production deficits. Recent work has examined receptive language processing at sublexical, lexical, and semantic levels in PPA; however, systematic investigation of these levels of processing within a single PPA cohort is lacking. The current study sought to fill this gap. Individuals with logopenic, nonfluent, and semantic variants of PPA and healthy, age-matched controls completed minimal pairs syllable discrimination, auditory lexical decision, and picture-word verification tasks to assess sublexical, lexical, and semantic processing. Distinct profiles were observed across PPA variants. Individuals with logopenic variant PPA had impaired performance on auditory lexical decision and picture-word verification tasks, with a trend toward impaired performance on the syllable discrimination task. Individuals with nonfluent and semantic variant PPA had impaired performance only on auditory lexical decision and picture-word verification. Evaluation of the types of errors made on the picture-word verification task (phonological and semantic) provided further insight into levels of deficits across the variants. Overall, the results indicate deficits in receptive processing at the lexical-phonological, lexical-semantic, and semantic levels in logopenic variant PPA, with a trend toward impaired sublexical processing. Deficits were observed at the lexical-semantic and semantic levels in semantic variant PPA, and lexical-phonological deficits were observed in nonfluent PPA, likely reflecting changes both in lexical-phonological processing as well as changes in predictive coding during perception. This study provides a more precise characterization of the linguistic profile of each PPA subtype for speech perception and language comprehension. The constellation of deficits observed in each PPA subtype holds promise for differential diagnosis and for informing models of intervention.

Carving Out Control Code: Automated Identification of Control Software in Autopilot Systems

Cyber-physical systems interact with the world through software controlling physical effectors. Carefully designed controllers, implemented as safety-critical control software, also interact with other parts of the software suite, and may be difficult to separate, verify, or maintain. Moreover, some software changes, not intended to impact control system performance, do change controller response through a variety of means including interaction with external libraries or unmodeled changes only existing in the cyber system (e.g., exception handling). As a result, identifying safety-critical control software, its boundaries with other embedded software in the system, and the way in which control software evolves could help developers isolate, test, and verify control implementation, and improve control software development. In this work we present an automated technique, based on a novel application of machine learning, to detect commits related to control software, its changes, and how the control software evolves. We leverage messages from developers (e.g., commit comments), and code changes themselves to understand how control software is refined, extended, and adapted over time. We examine three distinct, popular, real-world, safety-critical autopilots—ArduPilot, Paparazzi UAV, and LibrePilot to test our method demonstrating an effective detection rate of 0.95 for control-related code changes.

MALWARE OBFUSCATION MODEL USING MACHINE LEARNING

The paper presents the results of a research of the use of a generative-adversarial network for software obfuscation. The comparison was carried out with deterministic obfuscators Dotfuscator CE, Net Reactor, ProGuard. The research of changes in the source code of the software athena, abc, cheeba, dyre, december_3, engrat, surtr, stasi, otario, dm, v-sign, tequila, flip, grum, mimikatz was carried out using the IDA Pro disassembly tool. Detection of obfuscated malware was carried out based on the piecewise context fuzzy hashing method. The research was conducted to use of a generative-adversarial network as an obfuscator on pre-specified datasets. Simulation of the developed method was carried out in the Hyper-V virtual environment.

From GPU to CPU (and Beyond): Extending Hardware Support in GPUSPH Through a SYCL‐Inspired Interface

ABSTRACTWhile most software is originally designed for serial or parallel execution on CPU, and porting to GPU comes later in its development, GPUSPH was designed from the ground up to run on GPUs using CUDA. Making it accessible to a wider audience by introducing support for other computational hardware, and in particular CPUs, poses challenges that are complementary to the ones normally faced when porting CPU code to GPU. We present the approach we have adopted to support CPUs as computational devices in GPUSPH with minimal code changes and low developer effort. Detailed benchmarks illustrating the performance of the implementation and its scalability across multiple cores in both single‐socket and NUMA configurations show good strong and weak scaling.

Continuous Integration and Continuous Deployment (CI/CD) Optimization

The advent of Continuous Integration and Continuous Deployment (CI/CD) has fundamentally altered the landscape of software development, enabling teams to deliver updates with unprecedented speed and reliability. By automating the integration of code changes from multiple developers into a central repository, CI/CD practices ensure that software is continuously tested and deployed. This ongoing cycle not only facilitates quicker release cycles but also enhances collaboration among team members and fosters a culture of shared responsibility for code quality. Despite these advancements, organizations face significant challenges in optimizing their CI/CD pipelines. As software systems grow in complexity, the demand for swift and dependable deployments intensifies. This paper explores various techniques and strategies for optimizing CI/CD processes to minimize deployment times while maintaining system reliability. Key optimization methods discussed include: Parallelization of Build Processes: This technique involves breaking down the build process into smaller, independent tasks that can be executed concurrently. By leveraging distributed computing resources, organizations can significantly reduce build times, allowing for faster iterations and deployments. Dependency Caching: Caching dependencies can drastically decrease build times by reusing previously downloaded components. This approach not only speeds up the build process but also minimizes network load and enhances the overall efficiency of the CI/CD pipeline. Incremental Builds: Unlike full builds that compile the entire codebase, incremental builds focus on compiling only the changes made since the last build. This strategy reduces the amount of work needed for each build, accelerating the overall development process. The paper also delves into advanced rollback mechanisms such as blue-green deployments and canary releases. Blue-green deployments allow teams to maintain two identical production environments, enabling smooth transitions and quick rollbacks in case of issues. Canary deployments, on the other hand, introduce new features to a small subset of users before a full rollout, allowing teams to monitor the impact and catch potential failures early. Additionally, automated rollback mechanisms play a vital role in maintaining system reliability, ensuring that any failed deployments can be reverted swiftly to avoid downtime and user disruption. The impact of automation tools on deployment speed and error reduction is another critical aspect examined in this research. Automation frameworks can streamline various stages of the CI/CD process, from code integration to testing and deployment, minimizing human error and ensuring consistent, repeatable processes. By implementing robust automation strategies, organizations can not only accelerate their deployment cycles but also improve overall software quality.

The upper bound of information diffusion in code review

BackgroundCode review, the discussion around a code change among humans, forms a communication network that enables its participants to exchange and spread information. Although reported by qualitative studies, our understanding of the capability of code review as a communication network is still limited.ObjectiveIn this article, we report on a first step towards understanding and evaluating the capability of code review as a communication network by quantifying how fast and how far information can spread through code review: the upper bound of information diffusion in code review.MethodIn an in-silico experiment, we simulate an artificial information diffusion within large (Microsoft), mid-sized (Spotify), and small code review systems (Trivago) modelled as communication networks. We then measure the minimal topological and temporal distances between the participants to quantify how far and how fast information can spread in code review.ResultsAn average code review participants in the small and mid-sized code review systems can spread information to between 72 % and 85 % of all code review participants within four weeks independently of network size and tooling; for the large code review systems, we found an absolute boundary of about 11 000 reachable participants. On average (median), information can spread between two participants in code review in less than five hops and less than five days.ConclusionWe found evidence that the communication network emerging from code review scales well and spreads information fast and broadly, corroborating the findings of prior qualitative work. The study lays the foundation for understanding and improving code review as a communication network.

Historical and Current Perspectives on the Criminal Codes as Guarantees of Legal Security in Slovakia

The author describes two crucial periods in Slovak legal history from the perspective of the principle of legal security in criminal law. Firstly, the 19th-century Criminal Code, known as the Csemegi Code, is at the center of attention as it ended the centuries-long legal insecurity resulting from the lack of criminal law codification. The author analyses its leading principles as the safeguard of legal security. In contrast, the second part of the article focuses on the second codification of criminal law in Slovak legal history, the Criminal Code of 1950. The author demonstrates how the totalitarian regime destroyed the principle of legal security in criminal law in Slovakia, only slowly rebuilding it after the death of Joseph Stalin through amendments in 1956 and new legislation in 1961. The current Slovak Criminal Code and Criminal Procedure Code have been in effect since 2006. In conclusion, the author draws attention to how the latest changes in criminal law codes caused turbulences and worries over the principle of legal security in Slovakia.

Influencia de las propiedades mecánicas del acero de alta resistencia en el comportamiento de muros de concreto

The behavior of reinforced concrete walls with high-strength steel (yield strength greater than or equal to 550 MPa [80 ksi]) is evaluated. The influence of yield stress (fy) and the tensile-to-yield strength (ft/fy) are analyzed by means of tests on T-shaped concrete walls. Additionally, the effects of uniform elongation (Esu) and fracture elongation (Esf) are evaluated. Two walls reinforced with Grade 690 or 830 steel (100 or 120) (fy = 690 or 830 MPa [100 or 120 ksi]) and with concrete strength of 55 MPa (8 ksi). Test results were compared with previous tests to validate changes in the Concrete Design Code (ACI 318). The data obtained suggest that walls with high-strength steel have a deformation capacity similar to walls reinforced with conventional steel (fy = 420 MPa [60 ksi]). However, the steel must satisfy ft/fy≥1.2,Esu≤6%yEsf≥10% for the walls to exhibit satisfactory behavior. ILIA: Investigaciones Latinoamericanas en Ingeniería y Arquitectura, No. 01, 2024: 91-99.