Jump to text
On this page
Copy and source
Published source confirmed at last check

Source changed 2026-11-18 21:11:39 UTC · snapshot created 2026-11-18 21:12:35 UTC · last check 2026-11-18 22:45:10 UTC

OPS.2:5 - Archetypal Grounding — PumpWorks Mixed Operating Views

OPS.2:5.1 - Identify the useful views

PumpWorks-ControlServiceOps coordinates several unlike questions:

Current questionUseful viewBoundary
How should FieldIncident-I73 continue after new field evidence?Case view over the field installation, incident state, commitments, evidence, authority, and permissible next Work.The ticket is a record; safety and product decisions retain their owners.
How is release evidence repeatedly produced and checked?Process view of the recurring validation Method and its input/result relations.A process description does not establish that an occurrence happened or conformed.
How is a time-bounded provider-platform migration coordinated?Project and, when several projects are coordinated, programme views of commitments, WorkPlan content, decisions, allocations, and closure.The plan is not performed Work, adoption, or benefit.
How is scarce rig access coordinated across eligible requests?Queueing view over request membership, order/service relation, waiting, resource access, and policy.Establish request membership, service order, and resource access for this queueing view; a board column alone does not establish them.
Which observations and interventions matter at unlike rates?Conditional control-structure views for the product-side plant/controller/observer/supervisor relations and the operating-supervision relation.FPF owns the generic structures and rate claims; the diagram is neither the account nor the Operations Method.

The views preserve correspondences to ReleaseCandidate-R42, FieldIncident-I73, SafetyQuestion-S19, ProviderChange-P8, the test rig, commitments, evidence results, and decisions. The incident case can request a validation-process occurrence; that occurrence can consume rig capacity; its evidence can inform the weekly release decision; a platform-migration project can change provider access. None of these relations makes the views one structure.

The product-control structure has FieldPumpInstallation-P4 as plant, FieldTelemetryObserver-O4, DeployedController-C17, and FieldModeSupervisor-S1 only where their direct relations obtain. The operating-supervision structure has ReleaseSupervisorTeam-S2 receive operating-state reports and return release, hold, rollback, or permitted-mode constraints to the operating System. The timing claims remain separate: sub-second product control, minute-to-hour field observation, daily incident triage, weekly release, and slower provider change. Select the timing or control relations needed for this coordination decision. Establish feedback closure, stability, safety or authority separately when the decision relies on such a claim.

OPS.2:5.2 - Choose and combine actual coordination rules

In a constructed 09:00–11:00 window, PumpWorks permits at most two active engineering work packages under its existing staffing and recovery policy. T8 is already running. Incident diagnosis I73-D and the older provider-preparation package P8-P are both ready and permitted; both would need engineer E18. I73 has a field-contact window closing at 11:00, while P8-P can wait until the afternoon without violating a promise. The recurring T8 validation procedure is adequate and remains unchanged.

The incumbent rule starts the oldest ready package when a slot and performer are free. Its virtue is predictable order with little coordination. The incident rule allows a ready package to precede that order when a named response window would otherwise be lost, provided the incident lead authorizes the exception, the resource owner confirms it, displaced work remains feasible and the active limit is respected. These are supplied local rules, not universal properties of either a case view or Kanban.

Using OPS.17:4.2–OPS.17:4.4, the coordinator compares the two rules at this decision. Starting the noninterruptible two-hour P8-P would remove E18 from the incident’s remaining evidence-and-response interval. The exception instead starts I73-D, defers P8-P to an authorized afternoon review and leaves T8 running. Preserving the field-contact window costs one confirmation and a follow-up on the deferred package. Under the combined rules, readiness and permission govern eligibility, the incident exception governs order among eligible packages, the two-package limit still governs starts, and validation controls its own safe completion. No extra active slot is created.

At 09:50 the field contact becomes unavailable. The exception no longer justifies reserving E18 for live verification. The incident remains a case with a partial diagnostic result, but the live continuation returns for access and permission. T8’s completed evidence and the independently permitted T9 validation remain usable; the coordinator reconsiders ready demand without reopening every view. OPS.5:5.1, OPS.6:5.1 and OPS.7:5.1 develop the admission, observation and commitment decisions for this same constructed matter. They can also be used independently.

Keep the existing validation process where its rule fits. Add the case, queue or commitment relation when it changes the decision or the grounds for relying on it.

Reopen the mode decision when a new subject cannot be related across views, a plan repeatedly fails after new facts, queue policy changes a commitment, a control relation becomes decisive, or maintaining one view costs more than the decision value it supplies.