Supporting code and data for "A graded-uncertainty decision layer for global-state freshness in decentralized microservices"

Supporting code and data for the article "A graded-uncertainty decision layer for global-state freshness in decentralized microservices" (H. A. El-Ghareeb). The record contains two archives: nsdsl-software-v2.zip (MIT licence) — the full implementation: the ANTLR-based DSL and rules engine, the decentralized act/abstain decision layer, the graded-uncertainty family operators and scoring panels, the consistency baselines, the simulator, and the table/figure generators, with the reproducibility harness. nsdsl-data-v2.zip (CC BY 4.0) — the aggregated evaluation outputs: the content-hashed calibration file, the canonical per-configuration summaries underlying every figure and table (per_config_summary.csv; pairwise_tests.csv with Holm-corrected paired contrasts and Cohen's d; multiseed_validation.csv; and the per-experiment measurement tables), and the manuscript and supplementary figures. All results are produced under a fixed, content-hashed calibration whose SHA-256 is embedded in every committed results CSV header, and under the recorded random seeds described in the Methods. Given the software archive, the paper's tables and figures regenerate from per_config_summary.csv and pairwise_tests.csv (PYTHONPATH=src:. python -m experiments.make_tables / make_figures). The raw per-trial CSVs (the aggregator's input) are archived by the author and available on request. Concept DOI 10.5281/zenodo.20932988 always resolves to the latest version. Data are released under CC BY 4.0; the bundled software under the MIT licence (see the LICENSE file inside the software archive).