Skip to content

ALL-DATA0009 · data_definition_gap_percentage

Measure cataloged assets and fields lacking a business description.

This is a deterministic rule for all languages. Read its implementation.

Treat every asset and every field as one object that ought to carry a description, then divide the objects whose description is empty once trimmed by the number of objects. A catalog without descriptions is a list of names, and the cost lands on whoever has to guess whether amount is gross or net, in what currency, and as of when.

Presence is all this measures. Whether a description is accurate or useful is a judgment a contextual rule makes, and conflating the two would let a catalog full of restated column names score as documented.

Each finding names one asset or field whose description is empty. The value is the percentage of catalog objects carrying no description, and nothing is inferred from a name that looks self-explanatory.

A description of only whitespace reads as absent, since a field holding a space documents nothing. An empty snapshot measures zero rather than one hundred, because there is no undocumented object in it to count. A field whose meaning its name genuinely carries still counts as undocumented, and a project that disagrees is disagreeing with the rule rather than finding an exception to it.

One asset with an empty description holding two fields, one described and one not, has three objects and two gaps, so the value is about 66.7. An asset described together with its one described field returns 0. An empty snapshot returns 0.

  • Cites “DAMA-DMBOK”, metadata management principles
  • Cites “DataHub documentation”, glossary and description metadata