Complete can name several different events. A program may successfully register a reservation while publication time is still in the future. Reviewing Northstar's status copy made the business stage more important than the fact that an executor returned successfully.
Executor termination and business completion are separate axes
RunNames displays running, succeeded, failed, needs_review and cancelled. Reservation results then receive another interpretation: plan preview, reservation accepted/waiting, or all work checked complete. The outer executor can succeed while the inner business state remains queued.
Without that distinction, a person could interpret accepted scheduling as already published. Execution status supervises the program; business status explains progress toward the customer's purpose. Keep both and state the level of completion in the sentence. A green check is useful only when the operator can tell what was checked before choosing the next action.
A preview should name the action that has not happened
A reservation preview says the plan was checked and not yet submitted. Social results distinguish preview prepared/not published from an actual checked result. A visible output card must not imply that its external effect already occurred.
Uploads also distinguish private saving from scheduled visibility. Such copy needs evidence about the observed service state. Do not replace a start response with the customer's final objective. Copy is both a readable summary and an expression of an execution boundary. Naming an action not yet performed helps prevent duplicate work and inaccurate reporting.
Do not force review-needed into failure or completion
Needs_review appears as Result needs review. Some steps can finish while final outcome remains unsettled. That can differ from a simple unsent failure: read delivery and external effects before deciding what comes next.
The shared copy module demonstrates the distinction using separate declarations for delivery, observed result and final record. It neither verifies original evidence nor clears operational state. Contradictory success is rejected and wording grants no replay authority. An optimistic message cannot repair a still-unresolved operational record. Visible language should preserve the journal's meaning rather than conceal it.
Move history away from current progress without deleting it
A historical runtime audit addressed an old repair failure repeatedly presented as a current error. Terminal events were folded into dated history while current progress remained expanded. The change moved where failure was read without deleting it or turning it into success.
The current product removed embedded repair runtime, so that historical UI is not advertised as active automatic repair. The reusable principle distinguishes current activity from a past event. The synthetic history flag suggests aria-live off but constitutes no screen-reader test. Preserving time context prevents an old record from impersonating a new fault.
Attach the next check to the status
Running points to the same execution, waiting points to its existing result, and review-needed points to original delivery evidence. Reappearance of an enabled button does not establish authority to send the same input again. Visibility and execution permission cannot replace each other.
This article reads actual presentation code and historical checks without creating another reservation or publication. Module tests cover declared synthetic states only. The copy principle connects known facts to the next evidence worth reading rather than produce an optimistic completion sentence. A customer should understand what was accepted and what remains without memorizing internal status names. Choosing the next check becomes easier when the summary preserves that factual distinction.