sourceseparated

Sourceseparated is a term used to describe a publishing or data-management approach in which the materials that constitute a work’s sources are kept separate from the derived content, with a formal mechanism for linking the two. In this model, the source materials—such as original articles, datasets, or code—are stored in distinct repositories or metadata layers, while the derived content, including narrative text or analyses, is stored independently but openly linked to the sources. The separation aims to improve provenance, licensing clarity, and reuse.

Principles include separation of concerns, explicit provenance, and reproducibility. By decoupling sources from derived content, readers and machines can verify claims, assess reliability, and reassemble the work under different licensing conditions. Attribution is managed through metadata that records source identifiers, authors, publication dates, and licenses.

Implementation often relies on metadata standards and linked-data technologies. Common tools include bibliographic metadata (Dublin Core, BIBFRAME), provenance models (PROV-O), and persistent identifiers (DOIs, ARKs). A sourceseparated workflow might publish a content module with its derived text and a separate source module with the original materials, connected in a source graph or manifest.

Applications span academic publishing, open data portals, document management, and knowledge bases. Benefits include clearer licensing, easier content reuse, and better audit trails. Challenges can include increased tooling requirements, synchronization between source and content, and potential accessibility concerns.

See also: provenance, data lineage, attribution, licensing, open science.