discripsi

Discripsi is a term used to describe a framework for systematically describing, diagnosing, and documenting discrepancies among related data artifacts, including datasets, records, and textual sources. It focuses on producing structured descriptors, capturing provenance and context, to support data integration, auditability, and scholarly critique.

Etymology and origins are informal, with the term arising in late 2010s in discussions within information science and digital humanities about managing inconsistencies across sources. There is no universally recognized standard, and usage varies by discipline and project.

Core concepts and components include a taxonomy of discrepancy types (such as omission, addition, alteration, mislabeling, and timing or ordering issues), a set of qualifiers (certainty, provenance, timestamp, confidence), and a standardized descriptor syntax. Many implementations also include a scoring system to indicate the significance of each discrepancy and guidance for remediation.

Methodology and workflow typically involve data collection, artifact mapping (aligning items across sources), automated anomaly detection, manual classification, and the generation of descriptive records that document the nature and context of each discrepancy. Remediation or reconciliation actions are carried out as needed, with the descriptive records maintained to support future reference and verification.

Applications span data integration projects, digital archives, bibliographic databases, and version control for documents, as well as quality assurance, governance, and textual criticism. Advantages include creating auditable records of differences and aiding prioritization of reconciliation efforts; challenges involve resource requirements, potential subjectivity in classification, and the need for cross-domain conventions and governance.

See also: discrepancy theory, data provenance, data quality, textual criticism, audit trail.