Goal
Eliminate hand-maintained non-test inventory cardinalities from durable workshop-rs documentation where the canonical catalog, dataset, manifest, or generated report already owns the underlying facts.
Context
A broader WrightKit documentation audit found duplicated catalog/dataset counts that have already drifted. The root README currently claims all 255 canonical Workshop values, while docs/language-support.md claims 258 supported values & expressions. The same support documentation also hard-codes current totals for events, actions, and enum domains, and docs/language-support/enums.md repeats per-domain member totals beside the enumerated content.
The repository also contains useful dynamic coverage metrics in provenance documentation, such as locale mapping coverage. Those metrics are valuable only when they remain owned by a deterministic catalog/data pipeline; manually copying them into prose creates another synchronization point.
This work is separate from workshop-rs#104, which remains scoped to test/fixture/corpus counts. Organization-wide guidance for non-test derived counts is tracked in wrightkit/.github#12.
Scope
- Audit durable README/docs for mutable catalog, gameplay-dataset, locale-coverage, and support-inventory counts.
- Remove cardinalities that merely restate a list, table, catalog, dataset, or registry already present in the repository.
- Reconcile the current
255 vs 258 value-count contradiction by removing the duplicated hand-maintained claims rather than choosing another prose total.
- Remove redundant totals from the language-support index and component inventories unless the document is made a deterministic generated projection of the canonical source.
- For genuinely useful dynamic coverage metrics, prefer the canonical generator/check output or a generated/validated projection instead of manually synchronized prose.
- Preserve semantic numeric contracts such as Workshop element costs, exit codes, resource limits, versions, and other values whose number is itself meaningful.
Non-goals
- Do not change workshop-rs#104 or absorb its test/fixture/corpus-count scope.
- Do not change catalog membership, gameplay data, locale mappings, support state, or semantic behavior merely to make documentation totals agree.
- Do not remove numeric Workshop semantics or element-count formulas.
- Do not introduce a blanket numeric-text linter.
- Do not duplicate canonical catalog/data ownership into a new documentation-only manifest.
Acceptance Criteria
- Root README and language-support documentation no longer contain contradictory hand-maintained totals for the same catalog surface.
- Mutable action/value/event/enum/hero/member cardinalities are not manually duplicated when their canonical inventory is already available.
- Any retained dynamic coverage cardinality is generated or validated from its canonical source and its ownership is clear.
- Per-domain enum documentation describes members without redundant manually maintained
N members / N total members bookkeeping unless that presentation is generated from the catalog.
- Semantic numeric contracts, including the element-count model, remain intact.
- No test/fixture/corpus cleanup from workshop-rs#104 is folded into this issue.
Dependencies
Goal
Eliminate hand-maintained non-test inventory cardinalities from durable
workshop-rsdocumentation where the canonical catalog, dataset, manifest, or generated report already owns the underlying facts.Context
A broader WrightKit documentation audit found duplicated catalog/dataset counts that have already drifted. The root README currently claims
all 255 canonical Workshop values, whiledocs/language-support.mdclaims258 supported values & expressions. The same support documentation also hard-codes current totals for events, actions, and enum domains, anddocs/language-support/enums.mdrepeats per-domain member totals beside the enumerated content.The repository also contains useful dynamic coverage metrics in provenance documentation, such as locale mapping coverage. Those metrics are valuable only when they remain owned by a deterministic catalog/data pipeline; manually copying them into prose creates another synchronization point.
This work is separate from workshop-rs#104, which remains scoped to test/fixture/corpus counts. Organization-wide guidance for non-test derived counts is tracked in wrightkit/.github#12.
Scope
255vs258value-count contradiction by removing the duplicated hand-maintained claims rather than choosing another prose total.Non-goals
Acceptance Criteria
N members/N total membersbookkeeping unless that presentation is generated from the catalog.Dependencies