# The algo records whose domain is test-coverage

This index as JSON: https://banes-lab.com/json/api/facets/algo/domain/test-coverage

## Entries

- [Coverage Workspace](https://banes-lab.com/records/algo/coverage-workspace.md): Establish the unit under test, load the test-surface catalog as the space of what can be wrong, and enumerate every surface the unit can carry, so coverage is measured against the full derivable space rather than an ad-hoc list.
- [Surface Grid Walk](https://banes-lab.com/records/algo/surface-grid-walk.md): Walk the dimension-by-lens grid across the catalog, mark every cell that carries a surface, and surface the empty cells as candidate gaps using the anomaly lens, so a missing aspect is detected structurally rather than by recollection.
- [Uncovered Gap Derivation](https://banes-lab.com/records/algo/uncovered-gap-derivation.md): Reduce the candidate-gap set to the cells this unit can fail in, yielding the required-but-uncovered surfaces that still owe a test, so effort is spent on gaps that can fail rather than the full cartesian complement.
- [Coverage Risk Prioritisation](https://banes-lab.com/records/algo/coverage-risk-prioritisation.md): Score each required-uncovered surface by failure impact and reachability, rank the surfaces, and address the highest-risk gaps first, so limited testing effort removes the most risk.
- [Technique and Invariant Selection](https://banes-lab.com/records/algo/technique-invariant-selection.md): For each prioritised surface, state the invariant that must hold and select the technique whose reasoning mode matches how that surface is observed, so the test asserts the right property by the right method.
- [Test Authoring](https://banes-lab.com/records/algo/test-authoring.md): Realize each surface plan as an executable test: encode the predicate as an assertion and wire the evidence source the technique requires, producing a runnable test that covers the surface.
- [Evidence Verdict](https://banes-lab.com/records/algo/evidence-verdict.md): Run each test, gather its evidence, and issue the verdict: pass or fail when the evidence set is non-empty, unknown when it is empty, so an untested or unrun surface reads as unknown rather than as a silent pass.
- [Coverage Ledger](https://banes-lab.com/records/algo/coverage-ledger.md): Record each surface's verdict in a durable coverage ledger together with the residual uncovered and unknown set, so coverage state is reusable evidence rather than a fact re-derived on every run.
- [Coverage Completion](https://banes-lab.com/records/algo/coverage-completion.md): Mark coverage complete only when every required surface carries a surface, a technique, and an invariant, and its verdict is non-unknown; any required surface still unknown leaves coverage incomplete.
- [Test Coverage Kernel](https://banes-lab.com/records/algo/test-coverage-kernel.md): Load the surface catalog for a unit, walk the dimension-by-lens grid, derive the required-uncovered surfaces, prioritise them by risk, select a mode-matched technique and invariant per surface, author the test, issue an evidence verdict, ledger the state, and terminate only when no required surface remains unknown.
- [<Test Coverage Concern>](https://banes-lab.com/records/algo/test-coverage-concern.md): <Load the surface catalog for a unit> → <Walk the dimension x lens grid> → <Derive the required-uncovered surfaces> → <Prioritise by risk> → <Select a mode-matched technique and its invariant> → <Author the test> → <Issue an evidence verdict> → <Ledger the coverage state> → <Terminate when no required surface is unknown>