Victor Angelier BSc.

This paper examines the limitations of AI detection tools in academic settings, highlighting bias against non-native English speakers, students with disabilities, and collaborative forms of writing. Empirical studies show that AI text detectors exhibit an accuracy-fairness trade-off: attempts to improve detection rates can increase false positives, particularly for authors who deviate from expected language norms (Liang et al., 2023; Elkhatat, Elsaid and Almeer, 2023). Building on institutional… Show more

This paper examines the limitations of AI detection tools in academic settings, highlighting bias against non-native English speakers, students with disabilities, and collaborative forms of writing. Empirical studies show that AI text detectors exhibit an accuracy-fairness trade-off: attempts to improve detection rates can increase false positives, particularly for authors who deviate from expected language norms (Liang et al., 2023; Elkhatat, Elsaid and Almeer, 2023). Building on institutional decisions by universities such as Vanderbilt, Curtin, and several New Zealand institutions to disable or restrict AI detection in high-stakes assessment (Vanderbilt University, 2023; Curtin University, 2024; RNZ, 2025), we argue that detector-centric policies risk undermining fairness and trust. Drawing on anecdotal user-reported tests-including a cyber dependency preprint, team-written projects, and AI-translated professional texts-we illustrate how detectors often penalise stylistic features and translation artefacts rather than substantive originality. In dialogue with Edward Snowden's defence of privacy as a precondition for free expression, we contend that restricting AI assistance can disproportionately burden non-native and neurodivergent writers who benefit most from language-support tools (Snowden, 2015; Zhao, Cox and Cai, 2024). We conclude by advocating for process-oriented assessment, oral defences, and AI literacy as more inclusive alternatives to unreliable AI detection, situating detector-centric policies within wider human-rights debates on reasonable accommodation and non-discrimination in education. Show less

Het position paper van Stichting LibreKAT levert een inhoudelijk relevante bijdrage aan het debat over digitale soevereiniteit door expliciet te erkennen dat Nederlandse overheidsorganisaties structureel afhankelijk zijn geworden van externe digitale infrastructuur en leveranciers. Het paper doorbreekt terecht de wijdverbreide illusie dat compliance, certificering en governance op zichzelf voldoende zijn om deze afhankelijkheid te neutraliseren.

Deze peer review betoogt echter dat het… Show more

Het position paper van Stichting LibreKAT levert een inhoudelijk relevante bijdrage aan het debat over digitale soevereiniteit door expliciet te erkennen dat Nederlandse overheidsorganisaties structureel afhankelijk zijn geworden van externe digitale infrastructuur en leveranciers. Het paper doorbreekt terecht de wijdverbreide illusie dat compliance, certificering en governance op zichzelf voldoende zijn om deze afhankelijkheid te neutraliseren.

Deze peer review betoogt echter dat het LibreKAT-paper, ondanks zijn scherpe diagnose, blijft opereren binnen hetzelfde onderliggende paradigma dat het bekritiseert. Afhankelijkheid wordt zichtbaar gemaakt, maar niet normatief verworpen. Bovendien wordt digitale soevereiniteit expliciet benaderd als een gradueel fenomeen, waardoor zij wordt gereduceerd tot een schaalbaar risicobegrip in plaats van een afdwingbare machtsconditie.

Als alternatief introduceert deze review Bring Your Own Data and Keys (BYODK), niet als technische optimalisatie, maar als normatief en architecturaal grenscriterium voor vitale overheidsdata en -systemen, uitgewerkt via het principe Computing-to-Data. Show less

Digital sovereignty is increasingly invoked in European policy debates, yet the term is often diluted into compliance or risk‑mitigation language. This paper argues that such dilution obscures the fundamental problem: sovereignty is about exclusive and enforceable control. Where a foreign jurisdiction can exert legal, technical, or economic coercion, sovereignty is structurally absent, even when systems are secure, certified, or hosted in the EU.

Dit document presenteert een definitief en normatief kader voor digitale soevereiniteit, gericht op bestuurders, politici en topambtenaren. Het betoogt dat digitale soevereiniteit geen vorm van compliance, certificering of risicobeheersing is, maar een machtsbegrip dat draait om onvoorwaardelijke en afdwingbare controle over digitale systemen, data, cryptografische sleutels en kennis. Het rapport formuleert een heldere definitie van digitale soevereiniteit, onderscheidt deze expliciet van… Show more

Dit document presenteert een definitief en normatief kader voor digitale soevereiniteit, gericht op bestuurders, politici en topambtenaren. Het betoogt dat digitale soevereiniteit geen vorm van compliance, certificering of risicobeheersing is, maar een machtsbegrip dat draait om onvoorwaardelijke en afdwingbare controle over digitale systemen, data, cryptografische sleutels en kennis. Het rapport formuleert een heldere definitie van digitale soevereiniteit, onderscheidt deze expliciet van privacy en cybersecurity, en analyseert waarom Europese certificeringskaders zoals EUCS en SecNumCloud onvoldoende zijn om soevereiniteit te garanderen. Daarmee biedt het document een begripsmatig fundament voor strategische besluitvorming over digitalisering, staatsveiligheid en publieke infrastructuur. Show less

This report examines a fundamental shift in Software Engineering Project Management (SEPM) from traditional user-centred evaluation (UCE) towards signal-based governance. It introduces Epistemic Signal Governance (ESG) as an analytical construct that synthesises contemporary practices in infrastructure observability, developer productivity measurement, and automated decision-making within modern software engineering environments.

Drawing on established research, including the SPACE… Show more

This report examines a fundamental shift in Software Engineering Project Management (SEPM) from traditional user-centred evaluation (UCE) towards signal-based governance. It introduces Epistemic Signal Governance (ESG) as an analytical construct that synthesises contemporary practices in infrastructure observability, developer productivity measurement, and automated decision-making within modern software engineering environments.

Drawing on established research, including the SPACE framework for developer productivity, evolving software process models, and advances in distributed, cloud-native, and edge computing architectures, the report argues that evaluative authority in SEPM is increasingly relocating from subjective perception towards system-generated behavioural signals. These signals—derived from telemetry, CI/CD pipelines, incident management systems, and usage analytics—enable real-time, longitudinal assessment of quality, risk, and organisational performance.

The analysis critically evaluates the limitations of user-centred evaluation in large-scale, continuously deployed systems, highlighting issues of scalability, cognitive bias, and cost. It then explores the implications of signal-based governance for core SEPM functions, including quality assurance, release gating, risk management, resource allocation, and the evolving professional role of the project manager.

The report also addresses key risks associated with signal-based governance, such as metric gaming, loss of minority user voices, infrastructure dependency, and algorithmic bias, and proposes mitigation strategies grounded in disciplined governance design and ethical oversight. An implementation roadmap is provided to support organisations seeking to transition towards signal-based SEPM practices.

This work is intended for software engineering project managers, technical leaders, and researchers interested in the future direction of SEPM. Show less

This report analyses the structural risks to the Netherlands’ digital sovereignty arising from DigiD’s dependence on US-based infrastructure. It examines the legal impact of the CLOUD Act and FISA 702, the architectural role of Kyndryl, and the implications for confidentiality, oversight, and national security. Clear conclusions and actionable recommendations outline how the Dutch government can regain control over its critical digital identity infrastructure.

This preprint investigates the expressive boundaries of Behaviour-Driven Development (BDD) in security-critical software engineering. It argues that while Gherkin-based BDD enhances clarity and early validation of functional requirements, it cannot adequately represent non-functional or architectural security constraints such as encryption, key management or compliance auditing. Synthesising evidence from academic and industrial research—including large-scale case studies by Irshad et al… Show more

This preprint investigates the expressive boundaries of Behaviour-Driven Development (BDD) in security-critical software engineering. It argues that while Gherkin-based BDD enhances clarity and early validation of functional requirements, it cannot adequately represent non-functional or architectural security constraints such as encryption, key management or compliance auditing. Synthesising evidence from academic and industrial research—including large-scale case studies by Irshad et al. (2021) and architectural findings by Olmsted (2024)—the paper defines a complexity threshold beyond which declarative scenario languages fail. It concludes with a hybrid framework that combines BDD, UML security stereotypes, formal verification, and AI-assisted requirement analysis, providing a practical roadmap for achieving genuine security-by-design assurance in modern software systems. Show less

Story Point Standard Recommendation (SPSR): A Governance Framework for Data-Driven Decision-Making in Agile Software Engineering Project Management

This document presents a practical governance framework aimed at enhancing data integrity in agile software development. It addresses five critical failure modes that undermine the reliability of estimation practices.

Core Problem:
Data-driven decision-making in agile projects fails not due to the sophistication of analytical tools,… Show more

Story Point Standard Recommendation (SPSR): A Governance Framework for Data-Driven Decision-Making in Agile Software Engineering Project Management

This document presents a practical governance framework aimed at enhancing data integrity in agile software development. It addresses five critical failure modes that undermine the reliability of estimation practices.

Core Problem:
Data-driven decision-making in agile projects fails not due to the sophistication of analytical tools, but because of weak data governance. The SPSR framework operationalises governance principles by embedding disciplined data collection into estimation practices.

The Five Failure Modes:

Unreliable / Biased Data – Inconsistent story-point definitions contaminate datasets.

Insufficient / Obsolete Data – Lack of variance analysis fosters false confidence.

Automation Bias – Blind reliance on burndown charts without Definition of Done validation.

Absence of Feedback Loops – Sprints without retrospectives perpetuate bias.

Governance Gaps – Absence of enforceable authority and accountability structures.

The SPSR Solution:
A five-step governance protocol integrating standardised story-point definitions, baseline calibration, peer review, variance tracking, and feedback loops.

Proven Results:
Teams with inconsistent story-point definitions exhibited velocity variance of ±20% (25–30% forecast error). After implementing SPSR, this reduced to ±8% within two sprint cycles.

Application:
This framework aligns with the broader context of data management challenges in agile development and is consistent with recent research identifying data quality as a critical success factor in software engineering governance. Show less

Abstract
This paper examines why data-driven decision-making frequently fails within Software Engineering Project Management (SEPM) despite its theoretical advantages. Drawing on governance theory, cognitive psychology, and data-governance frameworks, the study identifies four systemic failure modes: unreliable or biased data, insufficient or obsolete datasets, automation bias, and structural self-validation bias ("marking one's own homework"). The analysis demonstrates that these failures… Show more

Abstract
This paper examines why data-driven decision-making frequently fails within Software Engineering Project Management (SEPM) despite its theoretical advantages. Drawing on governance theory, cognitive psychology, and data-governance frameworks, the study identifies four systemic failure modes: unreliable or biased data, insufficient or obsolete datasets, automation bias, and structural self-validation bias ("marking one's own homework"). The analysis demonstrates that these failures are not primarily technical, but organisational and cognitive. Even with accurate models, projects collapse when the same teams build, validate, and deploy the analytical tools they rely on. The paper proposes a governance-centred framework in which data validation, model use, and decision-making are institutionally separated and supported by independent audit functions, escalation mechanisms, and explicit confidence bounds. Rather than treating data-driven SEPM as a purely technical enhancement to the SDLC, it should be managed as a parallel governance system with its own accountability structures and quality controls. The argument concludes that sustainable data-driven SEPM requires structural separation of concerns, continuous model auditing, and recognition that effective data use increases organisational complexity rather than reducing it. Show less

In an era increasingly defined by technological dependency and geopolitical tensions, digital sovereignty has emerged as a critical imperative for national security, public trust, and democratic resilience. This essay investigates the necessity of a sovereign Dutch government cloud, arguing that current reliance on foreign cloud providers—particularly American hyperscalers like Amazon and Microsoft—exposes the Netherlands to legal, operational, and ethical vulnerabilities. Central to these… Show more

In an era increasingly defined by technological dependency and geopolitical tensions, digital sovereignty has emerged as a critical imperative for national security, public trust, and democratic resilience. This essay investigates the necessity of a sovereign Dutch government cloud, arguing that current reliance on foreign cloud providers—particularly American hyperscalers like Amazon and Microsoft—exposes the Netherlands to legal, operational, and ethical vulnerabilities. Central to these concerns is the extraterritorial reach of the US CLOUD Act, which threatens the privacy and autonomy of Dutch citizens by allowing American authorities to access data stored on foreign soil.

Through a case study of the Ministry of the Interior and its executive agency DICTU, the essay analyses the benefits and challenges of establishing a sovereign cloud infrastructure. It applies analytical frameworks such as PESTLE, TOE, SWOT, and the NIST Cloud Security Framework to evaluate strategic feasibility, legal compliance, and technical requirements. Drawing on comparative insights from initiatives like Gaia-X in Germany and X-Road in Estonia, the essay outlines a phased implementation strategy to ensure operational continuity, minimise disruption, and support capacity-building.

The study finds that a sovereign government cloud can strengthen national control over data, enhance transparency, reduce vendor lock-in, and align digital infrastructure with Dutch public values. However, it also acknowledges substantial challenges related to cost, legacy integration, and staff readiness. By addressing these through coordinated governance, open-source adoption, and investment in digital skills, the Netherlands can chart a path toward greater digital autonomy. Show less