P313 · Workflow control

Dependency-ready question rounds

Ask only decisions whose prerequisites have already settled, then recompute the next round.

Editorially reviewed

These examples and illustrative results are independently authored teaching materials, not measured model results.

Use case

A notification system needs retention and cleanup decisions. Teaching retention permission is unresolved, while provider limits are discoverable in current configuration/docs. Cleanup cadence depends on retention rather than an implicit answer in the same round.

Mechanism

Map decision/fact prerequisites and investigate discoverable facts. Ask needed ready decisions, grouping independent questions; recompute after actual answers. Unfinished research blocks dependents only. Review scope when the frontier empties to catch missing branches.

Bad example

Ask for 30-day retention, database and daily cleanup together before policy is established, then implement preselected choices without answers.

Good example

Inspect teaching limits/config first, then resolve notification purpose and necessary retention policy. Do not ask dependent cleanup cadence yet. After actual retention answers, ask compatible cleanup questions with dependencies recorded. Continue independent questions, preserve unresolved downstream choices and never treat 30 days/defaults as confirmed policy.

Why the change matters

Premature questions contain hidden premises. Dependency-ready rounds make decisions answerable from current information and reduce contradictory choices.

Observable expectation

An illustrative first round settles purpose/retention basis; later cleanup follows. Changed retention rechecks downstream choices.

Do not ask users for discoverable limits or mark unanswered decisions complete. An empty frontier needs complete scope review.

Limits

Actual product/compliance owners provide rules; this is no legal-period recommendation. Trees change and need recomputation. Necessary answers cannot be replaced by defaults/silence, and settled facts need no repeated interview.

Sources and evidence

Read the editorial criteria