Progressive change in how metadata elements, vocabularies, or syntax are applied across a corpus or system over time, producing temporal heterogeneity that reduces interoperability and complicates consistent interpretation of records.

Principle

Small, incremental alterations in local practice or tool defaults accumulate across time, creating systematic incompatibilities that are harder to correct than isolated errors because the drift is embedded in historical records and workflows.

Demonstration

Situation: A digital archive adopted a new controlled vocabulary in 2015 and a different date format in 2019. Recognition: Searches restricted by legacy subject terms or date patterns return incomplete results. Action: A researcher queries the archive across the full time span. Consequence: The researcher must run multiple queries or rely on a normalization step to retrieve comprehensive results, increasing effort and risk of omission.

Misapplication

Assuming that a single sample of recent records represents the whole collection. The error treats contemporary practice as uniform across time rather than recognizing temporal variation embedded in older records.

Consequence

Search inconsistencies across temporal slices, failed batch processing or analytics that assume homogeneous input, and higher cost for retrospective normalization; these follow because automated processes typically expect consistent element semantics and syntax.

Reversal

When provenance is recorded and versioned schemas or migration tools are applied, drift can be managed: older records are either transformed or queries are rewritten to respect historical practice, restoring interoperability despite temporal heterogeneity.

Boundary

Clearly within: A repository where records created in different years use distinct controlled vocabularies and date encodings. Boundary case: Schema changes are documented but not yet applied to legacy records. Clearly outside: One‑off transcription errors affecting isolated records rather than systematic temporal change.

Semantic Tension

Stability ↔ Evolution — maintaining stable semantics aids long‑term interoperability, while evolution enables adoption of better descriptors but introduces temporal heterogeneity.

Synthesis

Descriptive drift highlights the necessity of provenance, schema versioning and migration strategies: without explicit change records and transformation paths, incremental improvements become cumulative interoperability debts embedded across a corpus.