#Decision-Relevant Least Action and Operational Parsimony

A.11.OP · Stable · Part A pragmatic principle pattern · Normative unless marked informative · Part A - Kernel Architecture Cluster

#About this pattern

This is a generated FPF pattern page projected from the published FPF source. It is canonical FPF content for this ID; it is not a FPF Reference product feature page.

#How to use this pattern

Read the ID, status, type, and normativity first. Use the content for exact wording, the relations for adjacent concepts, and citations to keep active work grounded without pasting the whole specification.

Type: Part A pragmatic principle pattern Class: Prag Status: Stable Normativity: Normative unless marked informative

Plain name. Keep only work that changes a substantive choice or result, or protects a condition on which the use relies.

Primary reader. A practitioner or designer deciding whether one proposed action or supporting apparatus should be mandatory.

Use this when. Use this pattern when someone proposes making an action or apparatus mandatory and a plausible question remains: does this requirement change the subject work, or does it only make the route look controlled?

#Relations

#Content

#Problem frame

Use this when. Use this pattern when someone proposes making an action or apparatus mandatory and a plausible question remains: does this requirement change the subject work, or does it only make the route look controlled?

The primary EntityOfConcern is one proposed mandatory requirement under one declared use and one substantive horizon. The pattern screens the requirement for a substantive contribution. A contribution is necessary, but does not alone make the work obtainable, worthwhile or obligatory; action, apparatus, requirement, and horizon keep their ordinary meanings.

First useful result. Return one of two short answers:

• retain the requirement for this use and horizon because it has a named substantive contribution and its direct choice, realization, assurance or authority basis justifies requiring the work; or
• remove the requirement or leave it optional because it has no such contribution, or because a contribution does not justify its burden for this use.

For a proposed inquiry whose worth is still open, use C.11.DUA to compare what can be gained with feasible effort, delay, opportunity cost and downside; use C.11 when a current chooser and options need a local choice. Finish with the supported answer and feasible continuation, which may retain the current action, narrow a claim or decline an unsupported use.

Ordinary use needs no score or separate record. Name the receiving decision, result, reliance, or recovery condition in the same sentence as the disposition.

Three recognition cases.

• A team has added a second status update before a repair decision. Every possible status leaves the same repair action, and no later user relies on the duplicate update.
• A laboratory considers a bounded probe that leaves today's setup unchanged but can determine which of two methods will be used next week.
• A release route contains both a deterministic build step that creates the selected publication and an assurance check whose evidence is consumed by the release decision.

These are one recurring problem across unlike situations: mandatory effort can be ceremonial, immediately productive, decision-relevant only later, or necessary because another use relies on the assurance or recovery condition it preserves.

What goes wrong if missed. Requirements accumulate because each sounds prudent in isolation, while their possible results change no substantive choice and produce no selected result. The opposite error removes exploration, deterministic realization, safety evidence, recovery support, or a small discriminating cue merely because it does not change the next administrative state.

What this buys. The practitioner can remove ceremony without treating the fewest steps as the goal. Useful exploration, realization work, assurance, option preservation, and recovery remain when their receiving use is named.

Not this pattern when.

• When the question is the force or applicability of an instituted obligation, use its direct authority. Apply A.11.OP only to discretionary apparatus inside the space it leaves. If the present question concerns the obligation's merits, use C.11.DUA to examine its protective contribution, burden and feasible amendment; that appraisal does not cancel its current force.
• When several already qualifying alternatives need comparison, use their direct choice, apparatus, architecture, or Method Engineering pattern.
• When the question is whether a new durable ontology value should exist, use A.11.
• When the question is how to use an already selected pattern, use E.11.PUA or E.11.PUR.
• When ongoing Work needs one next action chosen from current facts, use A.15.7.
• When an available direct-kind apparatus is already being configured for a declared use, use C.19.2 for that application question.

#Problem

Methods and their administrative or support arrangements can accumulate apparatus, such as duplicate reads, checks, and handoffs. Each addition can be defended by a possible future benefit. If hypothetical usefulness is enough, every requirement survives. Attention and elapsed time then move from the subject result to the route's own states and receipts.

Simple minimization fails in the other direction. A deterministic assembly step may have no rival outcome yet still create the selected result. A decision maker may use new information to change a later policy while leaving the immediate action unchanged. A safety check may return the expected result while supplying evidence on which release reliance depends. A practitioner may use a recovery cue to avoid continuing from the wrong place. Counting branches, steps, documents, or minutes does not distinguish those cases.

The problem is therefore not how to minimize action in general. It is how to admit one proposed mandatory requirement only when a materially plausible difference reaches a named substantive use, without weakening the direct authority that governs the action.

#Forces

#Solution

Apply one bounded admission question before making the proposed action or apparatus mandatory.

Admission rule. An author or method designer MUST NOT make a proposed action or apparatus mandatory unless at least one materially plausible result can change a named substantive decision or branch within the declared horizon, the action realizes an already selected transformation or required subject result, or removing it changes a named assurance or recoverability condition on which the declared use relies.

Passing one branch establishes only a substantive contribution for this use and horizon. It does not establish that the work can be obtained or is worth requiring. Complete any live worth or choice question through its direct owner before selecting the requirement.

#Name the use and nearest substantive horizon

1. Name the proposed requirement and the declared use for which mandatory status is being considered.
2. End the horizon at the nearest named substantive decision, receiving use, selected transformation result, assurance use, or recovery use that can justify the requirement.
3. Name the possible result or removal consequence that reaches that horizon. Do not use the requirement's own status, completion flag, receipt, or other administrative transition as its receiver.

The nearest substantive horizon is not necessarily the next event. It may include a later decision when the dependency from the present result to that decision is stated. End it before any further use whose receiver and dependency have not been named.

#Compare keeping and removing through three branches

Compare the concrete situation with and without the requirement. A passing branch removes the objection that the work contributes nothing. Retain it only as far as its direct basis justifies requiring it. An already selected transformation or established reliance can supply that basis without another comparison.

When a proposed inquiry could matter but its worth remains open, apply C.11.DUA to the actual demand and receiving question. Identify the attainable observation, what it could change, and its whole cost within the receiving horizon. Compare available continuations through C.11 when a local choice is needed. A useful possible result can still arrive too late, require unavailable means, or cost more than its contribution. Keep the presently supported answer or select a feasible alternative under its actual limits. Do not invent an OptionSet or an inquiry merely to certify that none is needed.

If no branch passes, remove the requirement or leave it as an optional convenience. Convenience and prior investment do not supply the missing receiving difference.

#Judge material plausibility through the subject claim

Materially plausible means more than logical possibility and less than certainty. The direct owner of the claimed consequence supplies its standard of evidence. A low-probability result can remain material when its consequence changes exposure or the admissible policy. A large information volume is material only when some possible result changes a named receiving use.

When the branches cannot be distinguished, name the exact claim and missing basis and return them to that claim's direct owner. Keep the qualified answer already supported. A bounded experiment is one possible continuation only when an attainable result and worthwhile contribution justify its whole burden under §4.2. Unresolved usefulness alone does not select an experiment or create permanent mandatory status.

#Return authority and claims to their direct owners

Apply this screen only inside the space left by every applicable direct authority. The direct owner establishes the obligation or floor and resolves disputes about its basis or applicability.

When the requirement itself is being appraised, use C.11.DUA to compare the protected bearer and interest, the threshold and horizon, the causal contribution claimed, and who bears the burden. Identify who can amend the requirement and whether that amendment is feasible in time. Keep its merits and present force distinct: neither a protective label nor a burdensome rule settles the merits, and an unfavorable appraisal supplies no unilateral waiver. The legal, ethical and domain claims remain with their direct owners.

A passing branch establishes only the named contribution. Every downstream claim remains with the direct pattern named in Relations; obtain the required result by value instead of treating this screen as its substitute.

#Keep the result light and reopenable

For ordinary use, say:

Keep <requirement> for <declared use> until <nearest substantive horizon> because <named contribution and the basis for requiring this work>.

or:

Remove or demote <requirement> for <declared use> because keeping and removing it produce the same substantive decision and result and change no relied-on assurance or recovery condition.

If a proposed inquiry has a contribution but is unavailable or not worthwhile, finish with the current supported answer and selected continuation. Keep a short reason or limitation in that result when the recipient needs it; add no empty probe fields or separate omission account.

A named later use that must cite, compare, audit, or rely on the disposition records it in the existing record kind appropriate to that use. Otherwise the one-sentence result is complete.

Reopen the disposition when the horizon, plausible results, selected transformation, direct duty, assurance floor, recovery reliance, or burden-bearing alternative changes.

#Keep framework layers distinct

FPF owns this cross-domain admission principle. A Method Engineering DPF may use it when deciding which requirements should be mandatory in a named Method situation; that DPF still owns the Method-specific design. A local practice framework may bind the principle to its own execution and assurance mechanisms. Those mechanisms retain local scope, and the FPF admission condition must still be established for the declared use.

#Archetypal Grounding

#Duplicate status update and deterministic publication build

A publication repair route asks for a second status update immediately before the repair decision. The update has the same possible values as the first one. None changes repair, stop, or publish, no receiver cites it, and removing it changes no assurance or recovery condition. The second update passes no branch, so the route removes it or leaves it as an optional convenience.

The same route runs a deterministic build after the sources and publication form have been selected. The build has no decision-changing outcome by design, but it assembles the selected sources into the required publication. It passes selected realization and remains. Any claim about the assembled publication or its release still needs its direct owner.

#Exploration whose value appears in a later decision

A maintenance team must choose next week between Method A and Method B for a recurring seal failure. A bounded probe performed today can return one of three observations: evidence favoring A, evidence favoring B, or an unresolved result that triggers a hold. Today's immediate action is unchanged, but every possible probe result has a named effect on the later Method-selection decision.

The probe passes the contribution screen. Its horizon ends at that named selection and window. The team still has to decide whether to obtain it.

For a constructed comparison, suppose A, B and holding are all available within the applicable operating constraints. The team minimizes expected hours of later rework or deferral over the same maintenance horizon. Its current model gives three conditions with weights 0.4, 0.4 and 0.2:

These are illustrative planning inputs. On this basis, A and B each cost five expected hours; holding costs four. Suppose the probe distinguishes the three conditions in time. Choosing A, B or holding after its result costs 0.8 expected hours before probe effort. At one hour for the whole probe, the total is 1.8 hours, so the team selects the probe and the stated conditional continuation. At five hours for the same probe, still available before selection, the total is 5.8 hours: the completed advice is to hold on the current basis. If the probe cannot return before selection, its information does not serve this horizon. The operating constraints and uncertainty remain visible in each answer.

If instead the team establishes that every materially plausible observation leads to Method A and no other reliance changes, the probe fails even the contribution screen for that use. The team can choose A directly, without inventing a study and then recording why it was omitted.

#Assurance evidence with an unchanged operating decision

A pressure-system release check is expected to confirm the current operating decision. The release authority nevertheless relies on its evidence, and omission changes the accepted exposure for release. The check passes assurance preservation even when its most likely result leaves the operating branch unchanged.

B.3, the applicable evidence pattern, and the release authority set the assurance floor and disposition. A.11.OP returns only that the check is non-ceremonial for this named release reliance. A candidate check qualifies here only when the relying condition, exposure change, and direct owner are known.

Now consider a local rule requiring a copy of the same accepted check record. In this constructed case, every relying reader already has the original, and retyping adds no independent verification. It consumes the technician's only hour available for correcting an identified defect. The local rule owner can amend this copying requirement today while retaining the required check and its accessible evidence. The useful amendment is to use the original record and recover that hour for correction. The protected people and pressure-system condition, the lost correction opportunity and the feasible authority to amend supply the comparison. If the amendment cannot be obtained in time, the technician follows the applicable requirement or authorized hold route; a favorable merits comparison alone does not authorize omission.

#Recovery cue and discriminating language

After an interrupted multi-part analysis, a small progress marker identifies the last closed item and the next item. Removing it can cause the practitioner to repeat completed work or resume from the wrong branch. The marker passes recovery preservation while that continuation use relies on it. If the task is short and the next item is already unambiguous, the marker becomes optional.

In another case, a sentence must distinguish a reusable Method from dated Work that enacted it. The distinction changes whether the practitioner repairs a description claim or a performance claim. The requirement to preserve that distinction passes the decision-changing-result branch for this use. Other terminology remains informative unless a named interpretation or action depends on it.

#Speculative compliance without a receiver

A team proposes generating a compliance packet for possible future reuse. No applicable authority or named receiving use has been established for the packet, so it passes no branch and is not mandatory.

If an applicable duty or relying receiver is later established, reopen the disposition under that direct authority. The earlier speculative possibility was not evidence that the duty already existed.

#Bias-Annotation

Scope: Universal for the cross-domain admission question governed by this pattern.

#Conformance Checklist

#Common Anti-Patterns and How to Avoid Them

#Consequences

The pattern changes practice before a requirement is installed. A designer names the receiving horizon and checks what keeping or removing the requirement changes. Duplicate status work becomes removable without making “less paperwork” a universal argument. Deterministic transformations remain because they produce the selected result. Exploration, assurance, recovery, and small cues have a substantive reason to remain when their delayed or relied-on consequence is explicit. Their direct basis determines whether to require them; a relevant but excessive inquiry can give way to a qualified current answer.

#Rationale

Operational parsimony is about relevance, not abstract minimization. The fewest-step method can be wrong when one additional action realizes the chosen result, changes a later policy, or preserves a relied-on condition. The longest method can also be wrong when its extra actions have no substantive receiver. Comparing keeping and removing one proposed requirement makes that difference visible without inventing a global cost function.

The three branches distinguish contributions that a step-count screen would confuse. Decision-changing result recognizes exploration and discrimination. Selected realization recognizes deterministic work. Assurance or recoverability preservation recognizes a named relied-on condition. None alone establishes obtainability, net value or an obligation. Existing selection or reliance can settle the need; otherwise the direct choice or demand method completes it. This preserves useful work while allowing a probe that could matter to lose against an available continuation.

The horizon must be substantive and bounded. A next-event horizon hides delayed information value; an indefinite horizon lets hypothetical future usefulness justify everything. The nearest named receiver is the smallest horizon that can carry the reason and the smallest reopen boundary when the use changes.

The rule coordinates existing decisions, transformations, results, evidence, assurance, and recovery uses. Keeping these objects under their direct patterns preserves FPF layering while giving practitioners one discoverable admission question.

#SoTA-Echoing

#Relations

• Classified by: E.3 as one Prag principle. It primarily advances P-1 Cognitive Elegance, P-7 Pragmatic Utility, P-10 Open-Ended Evolution, and P-11 State-of-the-Art Alignment while respecting the other Pillars.
• Coordinates with: A.11, which governs admission of ontology additions. Namespace adjacency makes the two parsimony questions discoverable; their EntitiesOfConcern remain distinct.
• Coordinates with: E.11.PUA and E.11.PUR, which govern use, recommendation, coordination, and reuse after a pattern has been selected. A.11.OP asks whether an extra mandatory requirement belongs in the first place.
• Coordinates with: C.11.DUA for the merits and feasible continuation of an advice or evidence demand, including a single requirement with no live OptionSet; C.11 for a current local choice and probe-worthiness; and C.19.2, A.19, and Method Engineering for their apparatus, architecture or Method comparisons. A.11.OP supplies the contribution screen. These direct owners complete the applicable demand, selection or configuration question.
• Coordinates with: E.13 for proxy-to-value repair and E.23 for operations inside repeated evaluated improvement. A.11.OP retains the initial action-admission question.
• Coordinates with: A.3.1 and A.3.2 for Method and MethodDescription identity, A.15.1 for dated Work, and A.15.7 for next-action choice during ongoing Work. A.11.OP governs design-time admission of the requirement.
• Constrained by: applicable law and regulation, E.5 Guard-Rails, A.10 reliance boundaries, B.3 assurance floors, and any other direct subject or authority pattern for the use.
• Consumed by: Method Engineering and other DPFs when they decide whether domain-specific requirements or support apparatus deserve mandatory status. Those frameworks retain their domain decisions and evidence.
• May be specialized by: local practice frameworks for concrete execution and assurance. Their local arrangements remain local rather than FPF authority.

#A.11.OP:End

---

Last Updated: 2026-09-10 — upstream FPF commit a87d0ef4 (github.com/ailev/FPF)