Skip to content

Where the facts come from

How this data got here.

Provenance means making it traceable where a statement comes from, when it was fetched, and what has not been decided. The unfinished part belongs to that too.

The route the catalogue data takes

  1. Query

  2. Raw hash

  3. Normalisation

  4. Matching

  5. Draft import

Current state of checking

Catalogue records
367
of those unchecked
355
not published
134
sources recorded
59

Documented conflicts

Platform data contradicts the catalogue. For two releases both an institutional record and a platform statement exist. The platform gave a date four and twenty-two months adrift respectively. Checking showed the platform field matches the upload day exactly in every observed case — two tracks were uploaded 43 seconds apart. It is an upload timestamp, not a publication date, and is therefore not used as one.

The consequence: for three releases without a catalogue entry the publication date stays unknown. A plausible value was available; it would have been wrong.

Open decisions

  • Profile attribution

    owner
  • GND linkage

    owner
  • Media rights (partly cleared)

    owner
  • Release matching

    owner

To the source register