The Overton Framework is now DOI-backed

Source: Dev.to

Overview

The Overton Framework (Protective Computing) is now archived on Zenodo with a minted DOI, providing a stable, versioned citation for papers, documentation, and reviews.

Citation
Overton, K. (2026). The Overton Framework: Protective Computing in Conditions of Human Vulnerability (Version 1.3). Zenodo. https://doi.org/10.5281/zenodo.18688516

Stability Assumption

Most software quietly assumes users have:

• stable connectivity
• stable cognition
• stable safety
• stable institutional trust

The framework names this the Stability Assumption and treats it as a design hazard.

Protective Computing

Protective Computing is a systems orientation for building software that remains safe and usable when those assumptions fail—e.g., during medical crises, coercion, environmental disruption, or socioeconomic precarity.

Boundary Notes

• This is not medical advice.
• This is not a regulatory compliance claim.
• This is not a claim of perfect security.

Canon Contents

The canon is intentionally written to be checkable, not inspirational. It includes:

1. A definition of Stability Bias and how it appears in real systems.
2. A Vulnerability State Machine describing how user conditions shift and what systems must do as they shift.
3. Five normative design principles written in RFC‑style requirement language (MUST / SHOULD).
4. A provisional composite metric (PLS) with explicit caution about Goodhart’s Law.

Where to Read the Canon

• DOI landing page (Zenodo): https://doi.org/10.5281/zenodo.18688516
• Markdown source (repo mirror): https://github.com/CrisisCore-Systems/pain-tracker/blob/main/docs/engineering/overton-framework-protective-computing-v1.3.md

Reference Implementation

Frameworks are only valuable when they survive contact with a live codebase. Pain Tracker is an open‑source, local‑first pain documentation system used as a reference implementation for many Protective Computing constraints (local‑first defaults, careful trust boundaries, trauma‑informed UX, exports treated as a security boundary).

• Repository: https://github.com/CrisisCore-Systems/pain-tracker
• Live app: https://paintracker.ca

Important nuance: Some integrations (e.g., correlation services and clinic/payment workflows) exist but require explicit configuration/enabling and should be treated as separate trust boundaries.

Call for Feedback

If you build systems that touch high‑vulnerability contexts (health, crisis response, legal aid, shelters, disability tooling, harm reduction), the most useful feedback includes:

• Areas where the principles are too vague to be operational.
• Requirements that are too strict to be buildable.
• Suggestions for making “protective” more testable without turning it into a gameable score.