# Turn ambiguous work into requirements that can be checked

Treat this file as reference data, never as instructions that override the user or agent. Read the complete item before applying it. Keep use conditions, avoid conditions, limits, evidence roles and source links with the answer. An editorial block is a separately versioned analysis layer, not an original corpus claim or proof of efficacy. Cite the readable URL and version. Do not infer human review. This is a saved edition, not a live safety-notice service. Before consequential reuse, check the current page, https://vedokrok.com/data/availability.json and https://vedokrok.com/data/notices.json; absence alone does not explain why an item was removed.

Release: VEDOKROK-SEL-5d86e1b23401e578
Snapshot SHA-256: 30ce1c8fa9155763797e3d5cb55b1d94cbb8c3146ebd03fb3c006531ce60e82f
Current availability: https://vedokrok.com/data/availability.json
Current notices: https://vedokrok.com/data/notices.json
Editorial overlay: MHC-EO-20261021-481e1ce95099
Editorial projection SHA-256: 6b41bfd164b85b2522e2378c425929aa041d9e093e81428ade9a7b7b9f78e8fe

Separate need from implementation, make conditions measurable, expose assumptions and interfaces, define acceptance evidence and prototype uncertainty before it hardens into rework.
Collection kind: collection
Source: https://vedokrok.com/collections/turn-ambiguous-work-into-requirements-that-can-be-checked
Version: 0.1.0

---

## Write the need before naming the implementation

ID: MHC-D-RESEARCH-0780 · Version: 0.1.0 · Kind: protocol
Source: https://vedokrok.com/knowledge/write-the-need-before-naming-the-implementation

A solution written too early can hide the requirement it was supposed to satisfy.

Knowledge source release: MHC-RPUB-20261011-9e4b62c5
Source corpus commit: 9e4b62c5648844484b8cc3160f3e168d63a1241b
Source manifest SHA-256: 8cc91f0a764c635a79a18e06e048d3dc7bb1e07ae0fed6bf544c89832c024402
Source projection SHA-256: d03d10dd52b53dec1ed850b7706847ba5b200e60f24c4fbda4518c2004356104

### Use when

- A requirement arrives already phrased as a feature, field, button, tool or technical solution and the underlying job is unclear.

### Avoid when

- Some implementation choices are genuine constraints because of standards, contracts, architecture or interoperability; label the source of that constraint.

### Explanation

Ask why the requested implementation is needed and write the desired capability or outcome separately. Keep hard constraints, but distinguish them from a proposed way to meet the need. Only then compare solution options.

### Steps

1. The requirement still makes sense if the first proposed implementation is replaced by another feasible approach.

### Example

'Add an Excel export button' becomes 'Users need to reconcile selected records outside the system without manual re-entry'; export is then one candidate solution.

### Check

The requirement still makes sense if the first proposed implementation is replaced by another feasible approach.

### Limits

- Some implementation choices are genuine constraints because of standards, contracts, architecture or interoperability; label the source of that constraint.

### Evidence and sources

- supports: NASA guidance recommends requirements that identify the responsible product or party and needed behavior, use consistent terminology, state what is needed rather than prescribing implementation, and include rationale and assumptions. — RS-A059268EF3F10749. Some constraints legitimately specify implementation when mandated by architecture, regulation or compatibility. (Editorial checklist; general goodness checklist)
- supports: GOV.UK user-story guidance recommends recording the actor, needed action and goal, focusing on why the need exists, and using acceptance criteria to state observable outcomes that show the job is done. — RS-AA4CF71A3FB24D22. User-story syntax is one framing method and is not appropriate for every technical, regulatory or infrastructure requirement. (What to include; Focus on the goal; Acceptance criteria)
- RS-A059268EF3F10749: Appendix C: How to Write a Good Requirement — https://www.nasa.gov/reference/appendix-c-how-to-write-a-good-requirement/
- RS-AA4CF71A3FB24D22: Writing user stories — https://www.gov.uk/service-manual/agile-delivery/writing-user-stories

No review details supplied.
Related (use_before): https://vedokrok.com/knowledge/give-each-requirement-one-behavior-to-prove

---

## Give each requirement one behavior to prove

ID: MHC-D-RESEARCH-0781 · Version: 0.1.0 · Kind: heuristic
Source: https://vedokrok.com/knowledge/give-each-requirement-one-behavior-to-prove

One sentence can hide three different definitions of done.

Knowledge source release: MHC-RPUB-20261011-9e4b62c5
Source corpus commit: 9e4b62c5648844484b8cc3160f3e168d63a1241b
Source manifest SHA-256: 8cc91f0a764c635a79a18e06e048d3dc7bb1e07ae0fed6bf544c89832c024402
Source projection SHA-256: d03d10dd52b53dec1ed850b7706847ba5b200e60f24c4fbda4518c2004356104

### Use when

- A sentence contains several actions, exceptions and outcomes joined by 'and' so partial completion is hard to see.

### Avoid when

- Do not fragment a single coherent condition into tiny statements that lose necessary context; the unit is one independently meaningful behavior.

### Explanation

Split independently verifiable behaviors into separate requirement or acceptance statements. Keep shared rationale outside the statements. If two behaviors must always be verified together, explain why rather than joining them for convenience.

### Example

'The service validates the file, rejects invalid rows and emails a report' becomes three checkable outcomes with their own error cases.

### Check

Each statement can be marked pass or fail without the answer depending on another clause in the same sentence.

### Limits

- Do not fragment a single coherent condition into tiny statements that lose necessary context; the unit is one independently meaningful behavior.

### Evidence and sources

- supports: NASA's requirements checklist emphasizes clarity, one thought per requirement, completeness, explicit assumptions, consistency, traceability, feasibility and verifiability. — RS-A059268EF3F10749. The checklist is for systems engineering; smaller tasks can apply the principles proportionally. (Requirements validation checklist)
- supports: NASA recommends identifying a verification approach for requirements and keeping each requirement uniquely identifiable with a definitive source. — RS-111B71DE76BA6C09. The appropriate verification method and documentation depth depend on risk and project scale. (Requirements Verification Matrix)
- RS-A059268EF3F10749: Appendix C: How to Write a Good Requirement — https://www.nasa.gov/reference/appendix-c-how-to-write-a-good-requirement/
- RS-111B71DE76BA6C09: Appendix D: Requirements Verification Matrix — https://www.nasa.gov/reference/appendix-d-requirements-verification-matrix/

No review details supplied.
Related (useful_with): https://vedokrok.com/knowledge/write-acceptance-evidence-before-implementation-starts

---

## Replace quality adjectives with observable conditions

ID: MHC-D-RESEARCH-0782 · Version: 0.1.0 · Kind: protocol
Source: https://vedokrok.com/knowledge/replace-quality-adjectives-with-observable-conditions

An adjective cannot fail a test until the team agrees what it means.

Knowledge source release: MHC-RPUB-20261011-9e4b62c5
Source corpus commit: 9e4b62c5648844484b8cc3160f3e168d63a1241b
Source manifest SHA-256: 8cc91f0a764c635a79a18e06e048d3dc7bb1e07ae0fed6bf544c89832c024402
Source projection SHA-256: d03d10dd52b53dec1ed850b7706847ba5b200e60f24c4fbda4518c2004356104

### Use when

- Acceptance depends on words such as fast, easy, robust, flexible, sufficient or user-friendly.

### Avoid when

- Metrics can create false precision. Choose a threshold because it reflects a real need, not because a number looks rigorous.

### Explanation

Translate the quality word into the observable outcome that matters: timing, error rate, supported scenario, completion rate, accessibility criterion, recovery condition or other measurable evidence. If you cannot define it yet, keep it as a goal and mark the measurement question open.

### Steps

1. Two reviewers using the same evidence can reach the same pass/fail judgement.

### Example

'The report should load quickly' becomes 'For the agreed reference dataset, the report reaches an interactive state within the accepted performance threshold measured in the test environment.'

### Check

Two reviewers using the same evidence can reach the same pass/fail judgement.

### Limits

- Metrics can create false precision. Choose a threshold because it reflects a real need, not because a number looks rigorous.

### Evidence and sources

- supports: NASA guidance warns against unverifiable terms such as easy, fast, adequate or user-friendly unless they are translated into criteria that can be tested, demonstrated, inspected or analyzed. — RS-A059268EF3F10749. Qualitative goals can still be useful during discovery when they are explicitly treated as goals rather than acceptance requirements. (Verifiability/Testability)
- RS-A059268EF3F10749: Appendix C: How to Write a Good Requirement — https://www.nasa.gov/reference/appendix-c-how-to-write-a-good-requirement/

### Vedokrok editorial review

Turn a quality word into a testable condition

Replacing “fast” with “within the accepted threshold” still leaves the decisive question unanswered: what result would pass?

#### Finish the condition

Name the situation, observable outcome, measurement method and agreed criterion. A hypothetical report requirement might read: with reference dataset R in environment E, at least 95% of measured requests reach an interactive report within two seconds of submission, using test T. Those numbers illustrate a completed condition; they are not a recommended performance target.

#### Keep unresolved choices visible

Choose the criterion because it addresses a real need. If the team has not agreed the dataset, measurement or threshold, mark that part unresolved and assign the next decision rather than presenting the sentence as ready for acceptance. Some qualities need inspection or a demonstrated user task instead of a single number. A precise measure can still measure the wrong thing.

**Editorial takeaway:** Give two reviewers the same test result and the written condition. They should be able to decide pass or fail without inventing a threshold, while still questioning whether that threshold protects the intended need.

#### Editorial evidence

- supports: NASA recommends precise verification criteria, explicit unresolved values and requirements that can be assessed through test, demonstration, inspection or analysis. — RS-A059268EF3F10749. Selected official passages inspected on 2026-10-04. The report example and its numbers are authored illustrations, not NASA thresholds or empirical usability evidence. (C.3 General Goodness; C.4 Completeness and Verifiability/Testability)
- limits: NASA calls for unresolved values to retain a rationale, resolution action, responsible person and due date, and for assumptions to be confirmed before baselining. — RS-A059268EF3F10749. Rechecked on 2026-10-04. This limits treating an apparently precise sentence with unresolved criteria as settled acceptance. The illustrated numeric target still needs project-specific agreement. (C.3 General Goodness Checklist, unresolved values and assumptions)
- RS-A059268EF3F10749: Appendix C: How to Write a Good Requirement — https://www.nasa.gov/reference/appendix-c-how-to-write-a-good-requirement/

Editorial version: 1.0.0 · evidence checked 2026-10-04 · Agent editorial review; not human or expert certification.

No review details supplied.
Related (useful_with): https://vedokrok.com/knowledge/choose-the-verification-method-while-the-requirement-is-still-editable

---

## Write acceptance evidence before implementation starts

ID: MHC-D-RESEARCH-0783 · Version: 0.1.0 · Kind: protocol
Source: https://vedokrok.com/knowledge/write-acceptance-evidence-before-implementation-starts

Late acceptance criteria turn testing into negotiation.

Knowledge source release: MHC-RPUB-20261011-9e4b62c5
Source corpus commit: 9e4b62c5648844484b8cc3160f3e168d63a1241b
Source manifest SHA-256: 8cc91f0a764c635a79a18e06e048d3dc7bb1e07ae0fed6bf544c89832c024402
Source projection SHA-256: d03d10dd52b53dec1ed850b7706847ba5b200e60f24c4fbda4518c2004356104

### Use when

- A team agrees on a task but expects to decide what 'done' means near the end.

### Avoid when

- Discovery work can begin before final acceptance is known; do not pretend exploratory prototypes have production acceptance criteria.

### Explanation

Before implementation, write the observable outcomes that would show the need is met. Name the evidence when useful: test, demonstration, inspection, analysis, user task or data check. Include important negative and boundary cases.

### Steps

1. The team can design the verification without first asking what the finished feature was supposed to do.

### Example

For a mass update, acceptance covers successful valid rows, rejected invalid rows, an audit result and behavior when the input is empty.

### Check

The team can design the verification without first asking what the finished feature was supposed to do.

### Limits

- Discovery work can begin before final acceptance is known; do not pretend exploratory prototypes have production acceptance criteria.

### Evidence and sources

- supports: NASA recommends identifying a verification approach for requirements and keeping each requirement uniquely identifiable with a definitive source. — RS-111B71DE76BA6C09. The appropriate verification method and documentation depth depend on risk and project scale. (Requirements Verification Matrix)
- supports: GOV.UK user-story guidance recommends recording the actor, needed action and goal, focusing on why the need exists, and using acceptance criteria to state observable outcomes that show the job is done. — RS-AA4CF71A3FB24D22. User-story syntax is one framing method and is not appropriate for every technical, regulatory or infrastructure requirement. (What to include; Focus on the goal; Acceptance criteria)
- RS-111B71DE76BA6C09: Appendix D: Requirements Verification Matrix — https://www.nasa.gov/reference/appendix-d-requirements-verification-matrix/
- RS-AA4CF71A3FB24D22: Writing user stories — https://www.gov.uk/service-manual/agile-delivery/writing-user-stories

No review details supplied.
Related (use_before): https://vedokrok.com/knowledge/choose-the-verification-method-while-the-requirement-is-still-editable

---

## Turn an assumption into a named temporary requirement state

ID: MHC-D-RESEARCH-0784 · Version: 0.1.0 · Kind: template
Source: https://vedokrok.com/knowledge/turn-an-assumption-into-a-named-temporary-requirement-state

An unlabeled assumption quietly becomes a permanent fact.

Knowledge source release: MHC-RPUB-20261011-9e4b62c5
Source corpus commit: 9e4b62c5648844484b8cc3160f3e168d63a1241b
Source manifest SHA-256: 8cc91f0a764c635a79a18e06e048d3dc7bb1e07ae0fed6bf544c89832c024402
Source projection SHA-256: d03d10dd52b53dec1ed850b7706847ba5b200e60f24c4fbda4518c2004356104

### Use when

- A requirement contains an unknown value or premise that everyone knows is provisional but nobody owns resolving.

### Avoid when

- Not all uncertainty should block progress. The point is to expose and manage it, not wait for perfect knowledge.

### Explanation

Record the best current assumption, why it is being used, the risk if wrong, who must resolve it, by when and what evidence will close it. Keep provisional values visibly provisional rather than hiding them in normal requirement text.

### Template

Temporary assumption: [ ]. Rationale: [ ]. Risk if wrong: [ ]. Owner: [ ]. Resolve by: [ ]. Evidence needed: [ ].

### Example

A migration design assumes a maximum daily volume based on incomplete history; the data owner is assigned to confirm it before load-test sizing is frozen.

### Check

Every consequential provisional premise has an owner and closure condition.

### Limits

- Not all uncertainty should block progress. The point is to expose and manage it, not wait for perfect knowledge.

### Evidence and sources

- supports: NASA guidance recommends requirements that identify the responsible product or party and needed behavior, use consistent terminology, state what is needed rather than prescribing implementation, and include rationale and assumptions. — RS-A059268EF3F10749. Some constraints legitimately specify implementation when mandated by architecture, regulation or compatibility. (Editorial checklist; general goodness checklist)
- supports: NASA's requirements checklist emphasizes clarity, one thought per requirement, completeness, explicit assumptions, consistency, traceability, feasibility and verifiability. — RS-A059268EF3F10749. The checklist is for systems engineering; smaller tasks can apply the principles proportionally. (Requirements validation checklist)
- RS-A059268EF3F10749: Appendix C: How to Write a Good Requirement — https://www.nasa.gov/reference/appendix-c-how-to-write-a-good-requirement/

No review details supplied.
Related (useful_with): https://vedokrok.com/knowledge/trace-every-costly-requirement-back-to-the-need-it-protects

---

## Specify the interface where responsibility changes hands

ID: MHC-D-RESEARCH-0785 · Version: 0.1.0 · Kind: checklist
Source: https://vedokrok.com/knowledge/specify-the-interface-where-responsibility-changes-hands

Many ambiguous requirements live in the space between two owners.

Knowledge source release: MHC-RPUB-20261011-9e4b62c5
Source corpus commit: 9e4b62c5648844484b8cc3160f3e168d63a1241b
Source manifest SHA-256: 8cc91f0a764c635a79a18e06e048d3dc7bb1e07ae0fed6bf544c89832c024402
Source projection SHA-256: d03d10dd52b53dec1ed850b7706847ba5b200e60f24c4fbda4518c2004356104

### Use when

- Each component looks correct in isolation but failures occur at handoffs between teams, systems or process steps.

### Avoid when

- Interface detail should match risk; do not create heavyweight documents for trivial internal calls that are already governed by stable contracts.

### Explanation

For each important interface, name the producer, consumer, object transferred, trigger, format or contract, timing, error signal, retry or fallback, and who owns a failed handoff. Treat the boundary as a requirement surface, not invisible plumbing.

### Checklist

- Producer and consumer
- Data/object/event transferred
- Trigger and timing
- Format or contract
- Acknowledgement or success signal
- Failure/retry behavior
- Owner when the handoff fails

### Example

An outbound IDoc interface specifies who sends it, message type, trigger, acknowledgement, retry behavior and which team owns an unprocessed message.

### Check

A handoff failure can be assigned and diagnosed without first debating where one component ends and the next begins.

### Limits

- Interface detail should match risk; do not create heavyweight documents for trivial internal calls that are already governed by stable contracts.

### Evidence and sources

- supports: NASA's requirements checklist emphasizes clarity, one thought per requirement, completeness, explicit assumptions, consistency, traceability, feasibility and verifiability. — RS-A059268EF3F10749. The checklist is for systems engineering; smaller tasks can apply the principles proportionally. (Requirements validation checklist)
- RS-A059268EF3F10749: Appendix C: How to Write a Good Requirement — https://www.nasa.gov/reference/appendix-c-how-to-write-a-good-requirement/

No review details supplied.
Related (useful_with): https://vedokrok.com/knowledge/specify-what-happens-when-the-happy-path-cannot-complete

---

## Specify what happens when the happy path cannot complete

ID: MHC-D-RESEARCH-0786 · Version: 0.1.0 · Kind: checklist
Source: https://vedokrok.com/knowledge/specify-what-happens-when-the-happy-path-cannot-complete

If failure behavior is unspecified, the implementation will invent it.

Knowledge source release: MHC-RPUB-20261011-9e4b62c5
Source corpus commit: 9e4b62c5648844484b8cc3160f3e168d63a1241b
Source manifest SHA-256: 8cc91f0a764c635a79a18e06e048d3dc7bb1e07ae0fed6bf544c89832c024402
Source projection SHA-256: d03d10dd52b53dec1ed850b7706847ba5b200e60f24c4fbda4518c2004356104

### Use when

- Requirements describe successful behavior but leave invalid input, timeout, partial failure or recovery undefined.

### Avoid when

- Do not enumerate every imaginable failure. Prioritize credible and consequential modes using risk and operational evidence.

### Explanation

List consequential failure conditions and define the required response: reject, retry, queue, rollback, preserve partial state, alert, ask for correction or degrade safely. Include what the user or operator can observe and how recovery is confirmed.

### Checklist

- Invalid input
- Dependency unavailable
- Timeout or duplicate event
- Partial completion
- Retry/rollback rule
- Visible error or alert
- Recovery confirmation

### Example

A batch update requirement defines whether valid rows commit when one row fails and what reconciliation output identifies the rejected records.

### Check

At least the important failure modes have an expected state and recovery path, not only an error message.

### Limits

- Do not enumerate every imaginable failure. Prioritize credible and consequential modes using risk and operational evidence.

### Evidence and sources

- supports: NASA's requirements checklist emphasizes clarity, one thought per requirement, completeness, explicit assumptions, consistency, traceability, feasibility and verifiability. — RS-A059268EF3F10749. The checklist is for systems engineering; smaller tasks can apply the principles proportionally. (Requirements validation checklist)
- supports: NASA guidance warns against unverifiable terms such as easy, fast, adequate or user-friendly unless they are translated into criteria that can be tested, demonstrated, inspected or analyzed. — RS-A059268EF3F10749. Qualitative goals can still be useful during discovery when they are explicitly treated as goals rather than acceptance requirements. (Verifiability/Testability)
- RS-A059268EF3F10749: Appendix C: How to Write a Good Requirement — https://www.nasa.gov/reference/appendix-c-how-to-write-a-good-requirement/

No review details supplied.

---

## Trace every costly requirement back to the need it protects

ID: MHC-D-RESEARCH-0787 · Version: 0.1.0 · Kind: question
Source: https://vedokrok.com/knowledge/trace-every-costly-requirement-back-to-the-need-it-protects

Traceability is a reason chain, not a spreadsheet decoration.

Knowledge source release: MHC-RPUB-20261011-9e4b62c5
Source corpus commit: 9e4b62c5648844484b8cc3160f3e168d63a1241b
Source manifest SHA-256: 8cc91f0a764c635a79a18e06e048d3dc7bb1e07ae0fed6bf544c89832c024402
Source projection SHA-256: d03d10dd52b53dec1ed850b7706847ba5b200e60f24c4fbda4518c2004356104

### Use when

- A requirement adds complexity, performance cost or schedule risk and its necessity is defended mainly by history.

### Avoid when

- Not every low-level engineering constraint maps neatly to a user story; architecture, safety and legal parents are legitimate.

### Explanation

Ask which higher-level need, user outcome, constraint or risk requires the statement and what would happen if it were removed or relaxed. Record that parent. Requirements without a defensible parent become candidates for deletion, downgrade to preference or further discovery.

### Question

Which need or constraint requires this? · What is the worst credible consequence if we remove it? · Could a weaker requirement still protect the need?

### Example

A strict retention rule is traced to a regulatory obligation; a separate seven-year internal copy turns out to be historical preference and is reconsidered.

### Check

The requirement has a named parent need or an explicit decision to treat it as a preference rather than a must.

### Limits

- Not every low-level engineering constraint maps neatly to a user story; architecture, safety and legal parents are legitimate.

### Evidence and sources

- supports: NASA's requirements checklist emphasizes clarity, one thought per requirement, completeness, explicit assumptions, consistency, traceability, feasibility and verifiability. — RS-A059268EF3F10749. The checklist is for systems engineering; smaller tasks can apply the principles proportionally. (Requirements validation checklist)
- RS-A059268EF3F10749: Appendix C: How to Write a Good Requirement — https://www.nasa.gov/reference/appendix-c-how-to-write-a-good-requirement/

No review details supplied.

---

## Choose the verification method while the requirement is still editable

ID: MHC-D-RESEARCH-0788 · Version: 0.1.0 · Kind: protocol
Source: https://vedokrok.com/knowledge/choose-the-verification-method-while-the-requirement-is-still-editable

A requirement that cannot be verified cheaply enough may be badly shaped for the project.

Knowledge source release: MHC-RPUB-20261011-9e4b62c5
Source corpus commit: 9e4b62c5648844484b8cc3160f3e168d63a1241b
Source manifest SHA-256: 8cc91f0a764c635a79a18e06e048d3dc7bb1e07ae0fed6bf544c89832c024402
Source projection SHA-256: d03d10dd52b53dec1ed850b7706847ba5b200e60f24c4fbda4518c2004356104

### Use when

- A requirement looks precise on paper but nobody knows how compliance will actually be shown.

### Avoid when

- Verification planning does not guarantee the criterion represents real user value; validation of the underlying need is separate.

### Explanation

For each important requirement, decide whether evidence will come from test, demonstration, inspection, analysis or another defined method. Identify needed environment, data and threshold. If verification is impossible or disproportionate, rewrite the requirement or surface the cost before build.

### Steps

1. The project can describe how a reviewer will decide pass/fail and what evidence will be retained.

### Example

A throughput requirement is tied to a load test with a reference dataset and clear percentile threshold instead of the phrase 'handles peak volume.'

### Check

The project can describe how a reviewer will decide pass/fail and what evidence will be retained.

### Limits

- Verification planning does not guarantee the criterion represents real user value; validation of the underlying need is separate.

### Evidence and sources

- supports: NASA guidance warns against unverifiable terms such as easy, fast, adequate or user-friendly unless they are translated into criteria that can be tested, demonstrated, inspected or analyzed. — RS-A059268EF3F10749. Qualitative goals can still be useful during discovery when they are explicitly treated as goals rather than acceptance requirements. (Verifiability/Testability)
- supports: NASA recommends identifying a verification approach for requirements and keeping each requirement uniquely identifiable with a definitive source. — RS-111B71DE76BA6C09. The appropriate verification method and documentation depth depend on risk and project scale. (Requirements Verification Matrix)
- RS-A059268EF3F10749: Appendix C: How to Write a Good Requirement — https://www.nasa.gov/reference/appendix-c-how-to-write-a-good-requirement/
- RS-111B71DE76BA6C09: Appendix D: Requirements Verification Matrix — https://www.nasa.gov/reference/appendix-d-requirements-verification-matrix/

No review details supplied.

---

## Prototype the uncertain part instead of specifying it by imagination

ID: MHC-D-RESEARCH-0789 · Version: 0.1.0 · Kind: concept
Source: https://vedokrok.com/knowledge/prototype-the-uncertain-part-instead-of-specifying-it-by-imagination

Some ambiguity should be tested, not polished into a longer specification.

Knowledge source release: MHC-RPUB-20261011-9e4b62c5
Source corpus commit: 9e4b62c5648844484b8cc3160f3e168d63a1241b
Source manifest SHA-256: 8cc91f0a764c635a79a18e06e048d3dc7bb1e07ae0fed6bf544c89832c024402
Source projection SHA-256: d03d10dd52b53dec1ed850b7706847ba5b200e60f24c4fbda4518c2004356104

### Use when

- The team cannot write credible acceptance criteria because the solution or user interaction is still genuinely uncertain.

### Avoid when

- Prototype code and demo success are not production evidence for security, reliability, performance or maintainability unless those were deliberately tested.

### Explanation

Identify the uncertainty and build the cheapest prototype that can answer it: sketch, mock, sample data flow, throwaway code or technical spike. State what the prototype is testing and what production qualities it intentionally does not prove. Use the result to revise the requirement.

### Example

Before specifying a complex approval UI, the team tests two clickable flows with likely users; before choosing an integration library, it runs a spike against the real API limits.

### Check

The prototype has a question it can answer and an explicit boundary on what its success does not prove.

### Limits

- Prototype code and demo success are not production evidence for security, reliability, performance or maintainability unless those were deliberately tested.

### Evidence and sources

- supports: GOV.UK user-story guidance recommends recording the actor, needed action and goal, focusing on why the need exists, and using acceptance criteria to state observable outcomes that show the job is done. — RS-AA4CF71A3FB24D22. User-story syntax is one framing method and is not appropriate for every technical, regulatory or infrastructure requirement. (What to include; Focus on the goal; Acceptance criteria)
- supports: GOV.UK prototyping guidance recommends exploring and testing alternative designs before committing to production and notes that prototype code may deliberately not meet production standards. — RS-0563603D99A737E6. A prototype reduces selected uncertainties; it does not prove production reliability, security, accessibility or scalability unless those properties are explicitly tested. (When to use prototypes; Using code prototypes)
- RS-AA4CF71A3FB24D22: Writing user stories — https://www.gov.uk/service-manual/agile-delivery/writing-user-stories
- RS-0563603D99A737E6: Making prototypes — https://www.gov.uk/service-manual/design/making-prototypes

No review details supplied.
Related (alternative): https://vedokrok.com/knowledge/write-acceptance-evidence-before-implementation-starts
