OPS.2:4.1 - Pattern-Use Unfolding

Recover the operating input. Start from the OPS.1 operating System, continuing result, commitments, subjects, units, horizon, and decision. Stop if the mode question is only a label preference.
List the viewpoint questions. Name each user, decision, subject, grain, horizon, and missing information. Separate planning, recurring execution, case continuation, aggregate waiting/capacity, and control questions.
Generate serious view candidates. Include project or programme, process, case, queueing, and control-structure views only where their questions are current. Include the incumbent as the no-change option, and a simpler view when it is a plausible alternative.
Compare fit to the current Work. Examine predictability, recurrence, subject identity, decision latency, runtime discretion, plannability, shared-resource coupling, exception frequency, evidence refresh, and authority. These are comparison prompts, not a universal scoring formula.
Select the coordination Method separately. State which reusable coordination Method is used with each view, its applicability, burden, evidence, and stop. A notation, board, model, or view does not establish a Method.
Select actual structures independently. Use A.22 when a transformation-flow, queue/resource-dependency, commitment, event, case-state, control, or other actual structure changes the decision. A mathematical graph or diagram remains a representation.
Use the control branch conditionally. When observation, decision, actuation, supervision, feedback, or rate separation changes coordination, use A.22 and C.30.LCA to identify one selected control structure and recover the needed view. Use B.2.5 only for an obtaining two-sided supervisor relation and C.27.TA, C.27, or A.3.3 for the exact rate, temporal, or dynamics claim. Stop without a control view when no control relation changes the decision.
Compose views explicitly. State subject correspondences, scope and grain, governing commitments, state correspondence, handoff or coexistence conditions, decision rights, update rules, conflicts, and stop conditions. Do not rely on a “hybrid” label.
Test with a changing case. Ask what happens when new evidence defeats the plan, a case crosses a process boundary, queue delay changes a commitment, or control evidence arrives at a different rate. Repair only the affected view or correspondence.
Return and reopen. Record selected and rejected views, coordination Methods, unresolved correspondences, next subject-account need, and one observation that reopens the decision.