Uncoordinated or ad hoc changes in the structure, element names, datatypes, or constraints of a metadata schema across implementations or versions that invalidate validators, mappings, transformations, or downstream processes.

Principle

Automated pipelines and consumers assume a stable schema; when structural changes occur without versioning, deprecation policies or transform rules, ingestion, validation and downstream analytics fail because consumers cannot interpret altered element meanings or shapes.

Demonstration

Situation: An aggregator depends on the field to populate a timeline. Recognition: A supplier renames the field to and starts providing mixed date formats. Action: The aggregator's ingestion and index pipelines run without schema adaptation. Consequence: Dates are misparsed or dropped, timeline visualisations are incomplete, and automated reporting is incorrect until transform rules are implemented.

Misapplication

Blaming data‑processing tools alone for failures. The error is to treat incompatible outputs as tool bugs rather than noticing upstream schema changes; failing to track schema versions obscures the true causal link between change and failure.

Consequence

Broken ingest and validation, silent data loss or corruption during transforms, increased engineering effort to restore compatibility, and delays in services relying on the affected schema; these occur because consumers lack predictable element contracts.

Reversal

When schema evolution follows versioning, deprecation notices, backward‑compatible additions, or maintained transformation scripts, drift is an expected and manageable process rather than a disruptive failure.

Boundary

Clearly within: Same schema used across datasets but field names and constraints change without versioning. Boundary case: A minor optional element is added with clear documentation but without migration of legacy records. Clearly outside: Different schemas intentionally applied to distinct resource types with separate pipelines and clear contracts.

Semantic Tension

Rigidity/Stability ↔ Evolution/Flexibility — enforcing rigid schemas preserves predictable processing, while permitting schema evolution enables feature growth but requires governance to avoid breaking consumers.

Synthesis

Schema drift emphasises the need for schema governance: explicit versioning, deprecation policies, test harnesses and transformation layers convert schema change from a failure mode into a controlled process of evolution.