Qualification: AiiDA: automated interactive infrastructure and database for computational science
Pizzi, Cepellotti, Sabatini, Marzari, and Kozinsky provide an independent implementation precedent for separating evidentiary completeness from public disclosure. AiiDA records calculations and data as nodes joined by directional, labeled links that preserve the causal chain from input data through intermediate calculations to final results. The system stores queryable node attributes and permits users to inspect the history of a computational result. Its sharing design creates a separate boundary. Each user or group may operate a local instance in which data and calculations remain private. A user may then export only a portion of the database to another instance while retaining the remaining private portion. The relationship is a qualification. The distinction matters. AiiDA shows that detailed causal provenance and selective disclosure can coexist in one implemented system. It does not record a population of autonomous agents or establish Kaal's generation, lineage, fitness, survival, parentage, reputation, and adjudication fields. The paper also does not validate Kaal's execution record or disclose his private apparatus. This use of AiiDA is limited to its privacy and partial-export mechanism. The previously published AiiDA comparison addressed the distinct proposition that a directed graph can preserve an inspectable computational history. Evidentiary integrity does not require universal access to the system that produces the evidence.
research-methodsai-and-agentsreputationscholarly-growth-coveragescholarly-literatureprovenanceprivacyselective-disclosurecomputational-workflows