sourcesthat

Sourcesthat is a conceptual framework for recording, organizing, and evaluating the provenance and reliability of information sources. It is designed to accompany claims or data points with explicit references to the originating sources and any transformations they undergo, enabling traceability and accountability across documents, datasets, and knowledge bases.

The main goal of sourcesthat is to improve traceability, verifiability, and reusability of information in contexts such as academic research, journalism, fact-checking, and AI data curation. By documenting not only a source but also how it was used, transformed, or synthesized, sourcesthat supports transparent citation practices and audit trails.

Core components include a sourcesthat record with fields such as id, type (publication, website, dataset, etc.), title, authors, publication date, publisher or source, URL, access date, license, and a reliability or credibility rating. Provenance information—who added the source, when it was added, and any summarizations, translations, or extractions performed on the content—is stored alongside. Data models may be relational or graph-based to reflect relationships between sources, claims, and outputs, enabling networks of provenance.

Workflows for sourcesthat involve creation, validation, updating, deprecation, and version control. Applications span scholarly citation, digital libraries, journalism, and AI training data curation, where sourcesthat entries help document how information was obtained and used. Challenges include subjective reliability assessments, fluctuating source quality, licensing and privacy considerations, and managing scale. Example entries typically capture core metadata plus notes on reliability and use.