software testing

C2GRAPH: A Repository-Level Governance Package for Disciplined, Measurable AI Coding Agents

AI coding agents increasingly operate as long-horizon software actors: they inspect mutable repositories, call tools, preserve state across interruptions, and may eventually modify artifacts whose failure modes are difficult to reverse. This paper presents C2GRAPH (Convert to Graph Engineering), a repository-side governance architecture that treats these problems as explicit state and authority management rather than as prompt engineering alone. The central design separates evidence acquisition, semantic scope, exact mutation authorization, rollback and freshness, continuity, model/runtime adaptation, and diagnostic recovery. Four extensions—bounded emergency authorization, adaptive context representation, cognitive containment, and progress-liveness self-audit—are deliberately defined as non-authority planes unless they satisfy the same canonical mutation predicates as ordinary work. We formulate this requirement as authority non-interference and evaluate it with process-isolated executable replay, exhaustive enumeration over declared finite abstractions, boundary/equivalence models, and implementation-surface diagnostics. No counterexample was observed in the enumerated authority, liveness, numeric, representation, bounded-read, or adapter domains. The evaluation also discovered two implementation defects that conventional green-path testing had not exposed; both were repaired and converted into regression tests. The results support compositional repository governance within the stated artifact and model boundaries, but do not establish universal correctness, live-provider performance, crash consistency, or human-authentication guarantees.

(Over)Reliance on Test Agents in AI-Assisted Software Testing

citation-cliques: Replication package for Citation-Network Cohesion Across Journal-Impact Strata

Version 4 contains the verified full-cohort replication workflow, tests, manuscript sources, generated aggregate results, and compiled PDFs for Citation-Network Cohesion Across Journal-Impact Strata. The canonical rebuild uses every eligible identified author in the fixed source snapshot and produces 102,626 matched pairs across five subjects. The release archive does not redistribute the approximately 128 GB raw Crossref snapshot or the generated 17.5 GB analysis database; both are reproducible from the documented workflow. Results are aggregate associational evidence and do not establish intent, manipulation, or misconduct. Software is MIT-licensed; source-data conditions are documented in DATA_LICENSE.md.

Active Inference and Digital Twins for dynamic reconfiguration of software systems

In the following, the artefacts used during the experiments described in the article titled "Active Inference and Digital Twins for dynamic reconfiguration of software systems" are provided. The experiments were designed to evaluate and compare a Digital Twin enhanced with Active Inference (AIF-enhanced DT) proposal with another resource manager, called Proactive Latency-aware Adaptation (PLA). The article focuses on the development of the DT and the comparison between both reconfiguration managers with the aim of determining which of the two provides a better parameter configuration for the system in which they have been tested. The artifacts included are the followings: tuning_hyperparameters_runs.xlsx: This file contains the results of the 30 runs performed for each combination of Active Inference hyperparameters. Each sheet represents a different configuration. tuning_means_runs.xlsx: This file contains the average of the 30 runs performed for each combination of Active Inference hyperparameters. final_results_trace_0.xlsx: This file contains the average results of the 30 runs performed for each reconfiguration manager evaluated in the times series evaluation section of the paper. The Adapt variable distinguishes between them, with a value of 0 for the AIF-enhanced DT and 1 for PLA. Traces_1_2_3_results.xlsx: This file contains the average results of the 30 runs performed for each reconfiguration manager evaluated in the statistical comparison section of the paper. The Adapt variable distinguishes between them, with a value of 0 for the AIF-enhanced DT and 1 for PLA. B_matrix_complete: This file includes the complete version of the B matrix for the Active Inference Algorithm. This paper is part of the Grant PID2022-140907OB-I00 funded by MICIU/AEI/10.13039/501100011033 and ERDF, EU. It has also been partially supported by the Junta de Comunidades de Castilla-La Mancha/ERDF (SBPLY/24/180225/000020), by the University of Castilla-La Mancha (2025-GRIN-38441) and by the Cátedra Ciudad de Albacete (13585/2025). Elena Pretel holds a FPU21/02679 Scholarship from Spanish \textit{Ministerio de Educación y Formación Profesional} and holds a EST25/00141 scholarship from Spanish Ministerio de Ciencia, Innovación y Universidades. Dustdar‘s work and equipment has been supported by CNS2023-144359 financed by MICIU/AEI/10.13039/501100011033 and the European Union NextGeneration EU/PRTR. Many thanks to Gabriel A. Moreno for his help in developing the project.

PyFgsea: GSEA in Python and Rust

PyFgsea 0.2.0 Release date: 2026-09-02. PyFgsea 0.2.0 updates the statistical core used by the high-level Python API and narrows the trajectory interface to functions that are implemented and tested. The paper remains reproducible with its original software pair; current comparisons use a separate, newer reference pair. Highlights mode="aligned" and score_type="std" are the high-level defaults. Aligned mode uses exact pathway sizes (bin_width=0). Gene-ID ordering makes tied ranks deterministic. Unresolved and failed pathways remain visible in the result table. Results include tail-depth, error, termination, approximation, pathway-size, ranking, and algorithm-revision fields. The supported trajectory API is run_trajectory_gsea; unfinished public helpers and the incomplete trajectory CLI have been removed. DPT computation requires an explicit root and preserves root identity when cells are filtered or subsetted. The low-level build_tail_curve and query_tail_curve helpers remain available but are deprecated. They implement the approximate legacy-empirical-tail-v1 method, not the aligned estimator. Reference comparisons The two reference lanes answer different questions and should not be pooled: | Lane | PyFgsea | R | Bioconductor | fgsea | Purpose | | --- | --- | --- | --- | --- | --- | | publication | 0.1.4 | 4.4.3 | 3.20 | 1.32.2 | reproduce the comparison reported in the paper | | current | 0.2.0 | 4.6.0 | 3.23 | 1.38.0 | evaluate the current implementation | On the 100-pathway publication input, the publication lane reproduced the reported values after rounding: ES RMSE 3.84e-16; NES Pearson 0.999967, Spearman 0.999676, and RMSE 0.01333; transformed-p Pearson 0.997199, Spearman 0.988318, and RMSE 0.16960. On the same input, the current lane gave ES RMSE 3.21e-16; NES Pearson 0.999952, Spearman 0.999652, and RMSE 0.01567; transformed-p Pearson 0.999069, Spearman 0.996628, and RMSE 0.11826. A separate 60-pathway tie-heavy case exposes a 0.1.4 limitation. Its legacy NES RMSE was 0.29868; the current implementation reduced it to 0.01558 while restoring machine-precision ES agreement. This sensitivity case was not the publication input. Across 20 fixed score-permutation replicates, the current installed package returned all 2,000 planned null pathways. The pooled mean and median nominal p-values were 0.4991 and 0.5010; the proportions below 0.01, 0.05, and 0.10 were 0.0065, 0.0430, and 0.0935. These are descriptive summaries, not a fitted acceptance threshold. The original speed and memory values are not carried forward because the current run did not measure all tools under one matched scope. Figure 2 scope The accepted trajectory parameters are window_size=500, step=50, nperm_nes=2000, score_type="std", exact pathway sizes, bin_width=0, and NES caching off. The reassembled figure uses a processed 3,576-cell erythroid subset from GSE155254 with precomputed UMAP coordinates and DPT. Panel D ranks genes from scaled adata.X; the frozen object has no separate raw or named log1p layer. Public assignment tables match 2,097 control and 816 gata307mut cells, while 663 cells remain unmatched. The figure pools these groups and is descriptive, not a control-versus-mutant comparison. Benjamini-Hochberg adjustment is performed across pathways within each window and is not trajectory-wide error control. Migration from 0.1.4 The defaults intentionally differ: | Setting | 0.1.4 | 0.2.0 | | --- | --- | --- | | high-level score type | two_sided_abs | std | | estimator | implicit legacy behavior | mode="aligned" | | pathway-size binning | None | 0 (exact size) | | trajectory step | 100 | 50 | | trajectory NES permutations | 100 | 2000 | | trajectory NES cache | on | off | For new analyses, state the numerical contract explicitly: result = pyfgsea.run_gsea( data=ranks, gmt=pathways, gene_col="gene", score_col="score", mode="aligned", score_type="std", bin_width=0, seed=42, ) Install trajectory dependencies with: python -m pip install "pyfgsea[trajectory]" score_type="two_sided_abs" remains available for a bounded legacy-like calculation, but it is approximate and is not equivalent to fgseaMultilevel(scoreType="std"). The Python API does not accept an n_threads keyword. Packaging and paper impact Cargo.toml is the single version source for the Rust crate and Python distribution. Release packages are built as source distribution and platform wheels, installed into fresh environments, and tested through the installed Python package and native core. The two R reference images are also started and checked at runtime. The manuscript review is classified as a numerical and methods correction whose central software-concordance conclusion remains intact. The detailed decision note and claim matrix identify the text and values to keep, qualify, update, or remove. Limits The formal comparison covers the publication input and one predeclared tie-heavy sensitivity case; it is not a universal equivalence claim. Correlation and pathway overlap do not establish algorithmic identity or biological correctness. The Figure 1 run does not provide matched current benchmarks against GSEApy or BlitzGSEA. Runtime and peak memory remain hardware- and scope-dependent. Fast mode and the legacy empirical-tail helpers are approximate.

A Unified Taxonomy of Modeling Paradigms: Linear, Nonlinear, Context-Variable, and Self-Directed Causal Frameworks with Theoretical Foundations and Empirical Validation

This manuscript develops a unified taxonomy for modeling paradigms in complex, non-stationary systems, organizing linear, nonlinear, context-variable (regime-switching), and self-directed causal frameworks under a single structural causal model (SCM) formalism. Building on Markov-switching structural vector autoregression (MS-SVAR) theory, we present explicit assumptions, identifiability theorems, and a novel hybrid construct, the Causal Markov-Switching SCM (CMS-SCM), together with expanded, self-contained proofs. Every quantitative claim in this manuscript is the direct output of code executed during preparation of the manuscript: a Hamilton filter/Kim smoother expectation-maximization (EM) estimator for two- and three-regime MS-AR(1) models is implemented from first principles in NumPy/SciPy (no proprietary or unverifiable software), and is used to (i) recover regime parameters from synthetic data with 95% block-bootstrap confidence intervals, (ii) quantify one-step-ahead forecasting accuracy against both a linear AR(1) baseline and a genuine multilayer-perceptron neural-network baseline, (iii) recover regime-dependent causal structure in a four-variable synthetic system with a mean edge-recovery F1 of 0.993 ± 0.021 under oracle regime labels versus 0.708 ± 0.042 for a regime-blind pooled linear fit, (iv) characterize the empirical size and power of a moment-based linearity test, and (v) conduct a five-parameter global (Saltelli/Sobol) sensitivity analysis together with heavy-tailed-innovation and regime-order misspecification robustness checks, including a case in which information criteria favor an under-specified two-regime model over the true three-regime generating process. The regime-switching estimator reduces held-out mean-squared forecasting error by 12.1% relative to the linear baseline and 9.3% relative to the neural-network baseline. We further provide explicit Scientific and Technical Risk Assessment and Roadmap/Falsifiability sections that specify, for each central claim, the observation that would confirm or refute it.

Real-world GitHub Issue Duration Benchmark for Software PM Simulation Study

This dataset contains real-world software task duration data (n=346) collected via the GitHub REST API. Each record represents a closed-and-completed GitHub issue with its creation and closing timestamps, sampled across four categories (documentation, backend, frontend, testing) to serve as an empirical benchmark for software project duration estimation research. Collected in September 2026.

Real-world GitHub Issue Duration Benchmark for Software PM Simulation Study

This dataset contains real-world software task duration data (n=346) collected via the GitHub REST API. Each record represents a closed-and-completed GitHub issue with its creation and closing timestamps, sampled across four categories (documentation, backend, frontend, testing) to serve as an empirical benchmark for software project duration estimation research. Collected in September 2026.

The Influence of the Work Environment on Employee Work Productivity at the Krebet Baru Sugar Factory

Employee work productivity is one of the important factors in supporting the success of the company. One of the factors suspected to influence work productivity is the work environment, both from physical and non-physical aspects. This study aims to analyze the influence of the work environment on employee work productivity at Krebet Baru Sugar Factory. The study used a quantitative approach with a descriptive method. The study population was approximately 2,000 employees, while a sample of 50 respondents was determined using the Yamane formula with a simple random sampling technique. Data collection was carried out through a Likert-scale questionnaire, then analyzed using validity tests, reliability tests, classical assumption tests, simple linear regression analysis, t-tests, and coefficients of determination (R²) using SPSS software. The results showed that the work environment had a positive and significant effect on employee work productivity. The regression equation obtained was Y = 13.048 + 0.272X, with a t-value of 3.503 and a significance value of 0.001 (<0.05), so the research hypothesis was accepted. The coefficient of determination (R²) of 0.204 indicates that the work environment contributes 20.4% to employee productivity, while 79.6% is influenced by factors outside the study. This finding suggests that the better the work environment, both physical and non-physical, the higher employee productivity. Therefore, companies need to continuously improve the quality of the work environment to support optimal productivity.

Karabük ilinde sağlık kuruluşlarına başvuran 0-6 yaş çocukluk çağı aşıları ile ilgili bilgi, tutum ve aşı reddini etkileyen faktörler

This study was conducted to determine parents’ knowledge levels regarding childhood vaccinations for the 0–6 age group, their attitudes toward vaccination, and the factors influencing vaccine refusal. The study was conducted between February 2025 and April 2025 with parents who had children aged 0–6 years and who visited Primary Family Health Centers in the central district of Karabük and the district of Safranbolu. A total of 252 parents participated in the study. Data were collected by the researcher under the researcher’s supervision using a questionnaire consisting of 47 questions developed by the researcher based on a literature review. The sample size was determined to be a minimum of 252 based on calculations performed for the One-Way ANOVA F test. Statistical analyses were performed using the SPSS 27.0 software package; data were summarized using percentage distributions. Pearson Chi-square, Continuity Correction, and Fisher’s Exact tests were applied to examine the relationships between variables; a p<0.05 value was considered statistically significant. Mothers constituted 71.4% of the participants, while fathers constituted 28.2%. While 94.0% believed that childhood vaccinations were beneficial, 95.6% regularly followed the national vaccination schedule. The percentage of those who had heard of the concept of vaccine refusal was 79.4%, and it was determined that the most common reason for vaccine refusal (74.2%) was concerns about the reliability and protective nature of vaccines. When examining the relationship between parents’ sociodemographic characteristics (employment status, education level, occupation, place of residence, and number of children) and their vaccination status, no statistically significant difference was found in general. However, the existence of a positive perception of the benefits of vaccines was noteworthy. No significant difference was observed in parents’

Karabük ilinde sağlık kuruluşlarına başvuran 0-6 yaş çocukluk çağı aşıları ile ilgili bilgi, tutum ve aşı reddini etkileyen faktörler

This study was conducted to determine parents’ knowledge levels regarding childhood vaccinations for the 0–6 age group, their attitudes toward vaccination, and the factors influencing vaccine refusal. The study was conducted between February 2025 and April 2025 with parents who had children aged 0–6 years and who visited Primary Family Health Centers in the central district of Karabük and the district of Safranbolu. A total of 252 parents participated in the study. Data were collected by the researcher under the researcher’s supervision using a questionnaire consisting of 47 questions developed by the researcher based on a literature review. The sample size was determined to be a minimum of 252 based on calculations performed for the One-Way ANOVA F test. Statistical analyses were performed using the SPSS 27.0 software package; data were summarized using percentage distributions. Pearson Chi-square, Continuity Correction, and Fisher’s Exact tests were applied to examine the relationships between variables; a p<0.05 value was considered statistically significant. Mothers constituted 71.4% of the participants, while fathers constituted 28.2%. While 94.0% believed that childhood vaccinations were beneficial, 95.6% regularly followed the national vaccination schedule. The percentage of those who had heard of the concept of vaccine refusal was 79.4%, and it was determined that the most common reason for vaccine refusal (74.2%) was concerns about the reliability and protective nature of vaccines. When examining the relationship between parents’ sociodemographic characteristics (employment status, education level, occupation, place of residence, and number of children) and their vaccination status, no statistically significant difference was found in general. However, the existence of a positive perception of the benefits of vaccines was noteworthy. No significant difference was observed in parents’

Development and implementation of a simulation software for the irradiation control system in a carbon ion therapy nozzle

The proposed framework provides an effective laboratory platform for ICS functional verification, system integration testing, and subsequent debugging, and results indicate that the proposed framework provides an effective laboratory platform for ICS functional verification, system integration testing, and subsequent debugging.