[Google AI Overview (expanded), 2026-10-14, incognito, signed out. The paste carries the header 'AI Mode Conversation' and echoes the query twice; citation clusters are reduced to their numbers, links to their text; the raw paste keeps the /goto URLs.] You said: persistent identifier erosion observatory The Platform Erosion Observatory (also known as the PID Erosion Observatory) is an empirical research initiative and framework designed to measure whether persistent identifiers (PIDs)—such as DOIs—actually persist, and to study what happens when scholarly digital objects are removed from institutional or data repositories. [1, 2, 3] The project addresses the reality that PIDs are not permanent strings, but rather maintained relations that are susceptible to "erosion" or deletion over time. [1] Core Focus & Instruments * Measuring "Custody Separation": The observatory tracks the actual population of platform-driven scholarly record removals. It evaluates how platforms handle metadata after a deletion occurs, specifically evaluating "tombstone" compliance rates (whether a placeholder remains) and the "shape of absence". [1, 2, 3] * The Zenodo Removal Census: Acting as its primary case study instrument (Case 001), it tracks a specific cohort of over 14,000 DOIs to establish a baseline analysis of how, why, and when records vanish from the Zenodo repository. [1, 2] * Cross-Platform Tracking: It outlines a forward roadmap to extend this tracking across multiple major scholarly platforms to map structural enforcement actions and account-level deletions. [1] Key Ecosystem Concepts The project is part of a broader, sovereign alternative tracking architecture spearheaded by researcher Lee Sharks. It acts as the research counterpart to infrastructure initiatives, which include: [1] * The Severance Registry / Resolution Index: A system designed to map severed identifiers and resolve them outside of the primary platform's control. [1, 2] * The Capture Registry: An engine designed to capture asymmetry in how corrections and representation changes are documented across repositories. [1] For open-source code and dataset specifics, you can view the project directly on the Platform Erosion Observatory GitHub Repository. [1] If you'd like to dive deeper, let me know if you want to explore: * The methodology behind the Zenodo Removal Census * What tombstone compliance rates reveal about repository permanence * How a Resolution Index handles severed identifiers www.persistentidentifiers.org