# Complete authored final-project example — Operation Lamplight

## Example status

This is a complete **authored teaching submission package** for an alternate operation at fictional Halcyon Observatory. It is not a real team's work, a record of student playtesting or an answer to Operation Last Light. The event records describe the narrated example's domain states. No physical equipment test or delivered presentation is claimed.

## Mission contract

**Objective:** recover usable research matching the current signed star-atlas reference while preserving the original records and an explicit custody trail.

**Initial preference:** supported physical recovery. This is a method, not an unconditional objective.

**Constraints:** verify content identity; verify support before movement; retain a named custodian; use the current published route and permission rules.

**Stop conditions:** an unresolved identity mismatch, support below the load requirement, an unavailable route or an unconfirmed custody handoff blocks the dependent action.

**Resolution alternatives:** supported original transfer, verified digital recovery, or stabilisation pending handover. Each requires a defensible match to the stated contract.

## Six areas

1. Arrival desk — observation, recall and initial contract.
2. Workshop — connector and lift.
3. Power bay — support-circuit diagnosis.
4. Record dome — atlas comparison and permission charter.
5. Instrumented gallery — route prediction, closure and revision.
6. Dispatch balcony — agreement, receipt and debrief.

## Team agreement and contribution record

| Person | Principal role | Own evidence | Cross-role dependency reviewed |
| --- | --- | --- | --- |
| Lena | Observer / navigator | Arrival record, kit recall, route prediction and trace | Omar's support verification before route recommendation |
| Omar | Systems specialist | Connector mapping, gear/readiness record, powered and isolated circuit trace | Priya's distinction between identity and physical requirements |
| Priya | Investigator | Both manifests, copy comparisons and permission expectations | Jonah's objective and custody agreement |
| Jonah | Coordinator | Contract, explicit handoffs, council agreement, preserved plans and receipt | Priya's current-reference result and Omar's support limit |

Any member can stop a dependent action when its published prerequisite is unresolved. A result is handed to its recipient explicitly rather than assumed to be shared.

## Complete evidence register

- Arrival: 06:45, blue beacon, open hatch, cracked outer case. “All plates intact” remains a claim; outer damage alone does not establish internal loss.
- Kit: Wedge, Lens, Spool, Beacon at Gate, Basin, Workshop, Dome. The first retrieval swapped the middle items; the corrected retrieval and written checklist were both retained.
- Connector: initial north/west at level 0; clockwise 270 degrees gives west/south; translation to level 1 preserves orientation.
- Lift: driver 24, follower 36, input 6, output 4 opposite. Interlock released, cam 270, spring attached. Ratio and readiness checked separately.
- Circuit: powered source/fuse/cable/lamp 9/9/0/0 V; open cable diagnosed; isolated replacement; restored 9/9/9/9 V plus lamp light.
- Initial atlas requirement: 30 sections, H-3. Aster 30/H-3; Cinder 30/H-8; Dawn 29/H-1. Aster selected under the original authority.
- Permission charter: every role reads; maintainer services; custodian approves; all other combinations deny. Positive and negative cases checked.
- Environmental-pack handoff: West Store / 3 / AZURE, complete read-back and confirmation.
- Capacity: original mass 9; bare carrier capacity 7; checked cradle-supported capacity 11.
- Baseline route: 20 → 21 → 22 → 23 → 18 → 13 → 8 → 3. Sensor set 6, 7, 12. Predicted contacts 0; model trace contacts 0.

## Route verification records

Cells use the same five-by-five, row-major grid in every record: 0–4 are the top row and 20–24 the bottom row. Moves join orthogonally adjacent cells. A contact is recorded on entry to a listed sensor cell; one contact is an observation event, not an automatic mission failure.

**Preserved baseline — B-ROUTE-01:** start 20; goal 3; route 20 → 21 → 22 → 23 → 18 → 13 → 8 → 3; sensors 6, 7, 12. Prediction: zero contacts. Authored trace: 20 (0), 21 (0), 22 (0), 23 (0), 18 (0), 13 (0), 8 (0), 3 (0), where parentheses give cumulative contacts. This original record remains unchanged.

**Published storm route notice — H-ROUTE-02:** east-gallery cells 23, 18, 13 and 8 are closed. The available west path is 20 → 15 → 10 → 5 → 0 → 1 → 2 → 3. Its current published sensor pattern is 5, 6, 12. Every move is adjacent, no closed cell is entered, and the only sensor entered is cell 5. Lena predicts one contact before each execution.

| West-path entry | Sensor contact on entry | Rehearsal 2 cumulative contacts | Final-run cumulative contacts |
| --- | --- | --- | --- |
| 20 (start) | No | 0 | 0 |
| 15 | No | 0 | 0 |
| 10 | No | 0 | 0 |
| 5 | Yes | 1 | 1 |
| 0 | No | 1 | 1 |
| 1 | No | 1 | 1 |
| 2 | No | 1 | 1 |
| 3 (goal) | No | 1 | 1 |

**R2-ROUTE-02, authored 07:17:** observed sequence equals the published west path; one contact at 5 matches the prior prediction. **F-ROUTE-03, authored 07:10:** the final run rechecks H-ROUTE-02, then records the same sequence and one contact independently. The earlier zero-contact result is never reused as the revised prediction.

## Custody and receipt register

The original is **O-HAL-01**, the mass-nine star-atlas plate case initially held by **Dr Edda Quill**, observatory custodian, at **Record dome, stable cabinet RD-2**. The declared stable state is the closed case resting on the cabinet support with the position marker unchanged. This records preservation within the model; it does not establish the unseen condition of every plate.

The receiving archivist is **Dr Tomas Venn** at **Dispatch balcony, intake DB-1**. Jonah coordinates each handoff; Priya checks the content reference; Quill exercises the custodian approval. The digital carrier is **DC-H8**, containing the thirty-section Cinder package matching H-8. It travels along the west path; O-HAL-01 remains at RD-2. Carrier mass is 0.2 units, within the remaining capacity-seven handling limit.

Each rehearsal and the final run is a separate authored simulation starting with O-HAL-01 at RD-2 under Quill's custody. The baseline's simulated movement does not remove it from the next run's initial state. All times below are authored scenario event times, not browser-measured performance.

| Record | Authority and item | Transfer, location and acknowledgement |
| --- | --- | --- |
| B-CUST-01 / B-RCPT-01, 07:12 | Initial H-3 authority; O-HAL-01 with accompanying Aster 30/H-3 record | Quill releases the supported original to Venn at DB-1 after the capacity-eleven cradle and baseline route checks. Venn acknowledges: “Received O-HAL-01 and Aster, thirty sections, H-3, at DB-1.” Baseline end custodian: Venn; end location: DB-1. |
| R2-CUST-02 / R2-RCPT-02, 07:26 | Signed H-8 amendment; carrier DC-H8 with Cinder 30/H-8; route R2-ROUTE-02 | Quill authorises the copy and retains O-HAL-01 at RD-2. Venn checks both section count and H-8, then acknowledges: “Received DC-H8, Cinder, thirty sections, H-8, at DB-1. Original O-HAL-01 remains with Dr Edda Quill at RD-2.” Jonah records the acknowledgement and Quill confirms the retained position marker. |
| F-CUST-03 / F-RCPT-03, 07:19 | Current H-8 authority rechecked; carrier DC-H8 with Cinder 30/H-8; route F-ROUTE-03 | Priya repeats the count/reference check; Venn independently acknowledges the same complete copy and retained-original details under receipt F-RCPT-03. Quill confirms O-HAL-01 is still closed, supported and unmoved at RD-2. Final copy recipient: Venn at DB-1; original custodian: Quill at RD-2. |

The copy receipts do not claim transfer of the physical original. A later physical handover requires a new supported handling plan and a new custody record.

## Preserved Version 1

Inspect arrival → restore apparatus → verify H-3 atlas → verify cradle → cross east gallery → confirm custody.

Supporting conditions: H-3 reference current; east gallery available; cradle verified; supported capacity 11 for mass 9. Original objective and preferred physical method are recorded separately.

## Rehearsal 1 — Baseline

| Checkpoint | Predicted / required | Authored result |
| --- | --- | --- |
| Scene and kit | Record observations and retrieve four items | Middle-item recall substitution corrected and retained |
| Connector / lift | West/south at level 1; 4 opposite output turns | Both requirements met |
| Circuit | Localise supply break, isolate, verify | Open cable repaired; 9 V and light confirmed |
| Atlas / authority | 30/H-3; authorised custodian approval | Aster selected; positive and negative permission checks agree |
| Handling / route | Capacity 11 ≥ mass 9; east gallery available; 0 contacts | Supported path meets conditions |
| Custody | Quill releases O-HAL-01 to Venn at DB-1 under H-3 | B-CUST-01 and B-RCPT-01 record the physical item, Aster 30/H-3, recipient, location and 07:12 acknowledgement |

**Result:** supported physical recovery under baseline conditions. This result does not authorise movement after those conditions change.

## Rehearsal 2 — Published storm changes

Notice: the east gallery closes and the cradle is unavailable. The remaining carrier capacity is 7 for a mass-9 original. A signed amendment requires the 30-section annotated H-8 research copy.

| Checkpoint | Proposed or expected | Authored result and diagnosis |
| --- | --- | --- |
| Reuse physical plan | Continue the successful baseline | Stopped before movement: missing cradle invalidates support |
| Scope the change | Preserve valid evidence | Connector/circuit results retained within scope; current route and reference rechecked |
| Current atlas | Match amended 30/H-8 authority | Cinder selected; Aster's historical H-3 result preserved |
| New agreement | Research access with original preserved | Quill authorises DC-H8/Cinder and retains O-HAL-01, closed and supported at RD-2; R2-CUST-02 |
| Revised route | H-ROUTE-02: 20 → 15 → 10 → 5 → 0 → 1 → 2 → 3; sensors 5, 6, 12; predicted 1 contact | R2-ROUTE-02 records the same sequence and one contact at 5; closed east cells are avoided |
| Receipt | Venn confirms DC-H8/Cinder 30/H-8 at DB-1 and Quill's retained original at RD-2 | R2-RCPT-02 records the full acknowledgement at 07:26 |

**Result:** verified digital recovery after a documented revision. No unsupported original movement occurs.

## Version 2 and revision rationale

Preserve original → verify H-8 copy → agree digital custody → check west route → transfer copy → confirm receipt.

The revision changes the affected transport and content conditions. It does not pretend the original baseline was unreasonable, erase its evidence or silently redefine physical recovery as digital recovery. The research-access objective remains; the preferred method changes. The original stays stable with the observatory custodian. Future physical-provenance work requires a supported handover.

## Final run record

Current checks repeat rather than inheriting permanent approval from rehearsal. Priya verifies Cinder has 30 sections/H-8 under the current signed authority. Lena reads H-ROUTE-02, predicts one contact on 20 → 15 → 10 → 5 → 0 → 1 → 2 → 3, and records one contact at 5 in F-ROUTE-03. Venn confirms receipt of DC-H8/Cinder at DB-1 in F-RCPT-03 at 07:19. Quill confirms original O-HAL-01 remains closed, supported and unmoved at RD-2 in F-CUST-03. The precise trace and acknowledgements appear in the registers above.

**Resolution:** verified digital recovery with the original stabilised in place.

**Consequence:** research access is achieved while physical-original transport remains outstanding. Matching a reference does not establish calibration or scientific correctness.

## Team account (984 words)

Operation Lamplight asked us to recover usable, verified star-atlas research from Halcyon Observatory while preserving the original instrument records. We defined success as delivering the currently authorised research copy or original with a documented custody trail. Physical removal was our initial preference, not an unconditional requirement. We agreed to stop any movement whose support conditions were unverified. This distinction mattered when the storm changed both the route and the equipment available to us.

Our four roles were observer/navigator, systems specialist, investigator and coordinator. The observer owned the arrival record and route predictions. The systems specialist owned the connector, lift and support-circuit checks. The investigator owned manifest comparison and permission expectations. The coordinator owned handoffs, custody and changes to the shared plan. We required each result to reach another role: a private successful test could not silently become permission for the whole team to act. Each member also reviewed one dependency outside their main responsibility.

At the arrival desk, the observer recorded a 06:45 display, blue beacon and open inspection hatch. A note claimed all atlas plates were intact. We preserved that statement as an account rather than an observation of every plate. The cracked outer case was evidence of damage to the case, not proof that every record inside was lost. This separation prevented the atmosphere of the scene from deciding the investigation before the manifest and handling checks had taken place.

The route-kit exercise used Wedge, Lens, Spool and Beacon at Gate, Basin, Workshop and Dome. The observer initially swapped the middle two items, then strengthened the Lens-at-Basin image and retrieved the sequence with the list covered. We retained that correction. The mnemonic supported carrying a short instruction set between stations, but the written kit remained available for consequential verification. We did not turn successful recall into a claim of unlimited memory or make memory a substitute for a required check.

The workshop connector began north/west at level zero. A 270-degree clockwise rotation mapped its ports to west/south; moving it to level one changed height without changing orientation. The lift then required four output turns from six input turns using a 24-tooth driver. The systems specialist selected a 36-tooth follower, released the interlock, set the 270-degree cam and attached the spring. Six input turns produced four opposite output turns. The model trace explains both the ratio and the readiness conditions; merely showing movement would have omitted half the diagnosis.

The support circuit provided another independent check. Powered readings were nine volts at the source and fuse output, then zero after the cable and at the lamp supply. That transition supported an open cable. The specialist isolated power, replaced the cable and restored power. Nine volts at the lamp together with visible light verified the complete repair. We treated these as authored model observations, not hardware measurements, and kept the apparatus state beside each reading so isolated zeros could not be mistaken for diagnostic evidence.

Under the initial signed reference, the required atlas contained thirty sections with digest H-3. Aster met both conditions, Cinder contained thirty with H-8 after annotation, and Dawn contained twenty-nine with H-1. The investigator selected Aster and stated the limit: a digest match establishes identity against that reference, not the scientific truth of every stellar observation. The local charter allowed all roles to read, maintainers to service and custodians to approve release. We checked useful operations as well as denied ones so a blanket lockout could not masquerade as a repair.

The coordinator arranged a delivery of three environmental packs to West Store with code AZURE. The receiver read back all fields before confirmation. At the council, the original atlas had mass nine while the unsupported carrier capacity was seven. A verified cradle raised supported capacity to eleven. The custodian accepted either a documented verified-copy transfer or supported original movement under the initial research-access mandate. We selected physical recovery for the baseline rehearsal only after the cradle check, rather than treating a confident equipment claim as permission.

Rehearsal one used the baseline conditions. The instrumented route was 20, 21, 22, 23, 18, 13, 8, 3 with published sensor cells 6, 7 and 12. We predicted zero contacts and the recorded model trace agreed. With the cradle verified, Aster selected and custody confirmed, physical recovery met the stated conditions. We saved this as Version 1. Its success demonstrated the plan under those conditions; it did not prove that the same sequence would remain possible after the storm.

The second rehearsal introduced a published east-gallery closure and made the cradle unavailable. The unsupported carrier still had capacity seven, so moving the nine-unit original could not be justified. A signed amendment also required the annotated H-8 research copy. We stopped the original transport proposal before movement, preserved the baseline result and revised the plan. Cinder became the correct current copy; the historical Aster comparison remained valid under H-3. The coordinator negotiated verified digital recovery while leaving the original stable with the custodian.

Version 2 retained the successful connector and circuit findings within their stated scope, added a current-reference check, used the available west route and required a new receipt. The second rehearsal completed that revised chain. Our final run repeated the current checks rather than treating the rehearsal as permanent authorisation. Cinder matched thirty sections and H-8, the recipient confirmed receipt, and the original's location and custodian were recorded. The resolution was verified digital recovery with stabilisation of the original, not physical extraction.

The change preserved research access while accepting a consequence: the physical original remained at Halcyon and would require a later supported handover if provenance work demanded it. Calibration and scientific accuracy also remained separate questions. Our strongest decision was recognising that the original objective allowed a different resolution; our weakest early assumption was treating cradle availability as a stable background fact. The completed record therefore includes both successful and interrupted plans, their authorities and the checks that justified changing course.

## Ten-minute presentation package

The 984-word team account above is the complete prepared spoken script. The schedule budgets speech at **130 words per minute** and rounds each spoken segment up to the next second: **458 seconds of speech**, **120 seconds for six demonstrations**, and **22 seconds for handoffs and visual transitions** total **600 seconds**. The demonstrations are silent pauses; pointing to records while speaking does not add an unbudgeted demonstration.

| Segment | Speaker | Script allocation | Speech / demonstration / transition budget |
| --- | --- | --- | --- |
| 0:00–1:15 | Jonah | Paragraphs 1–2; 154 words. Contract, roles and stop conditions. Show objective beside preferred method while speaking. | 72 / 0 / 3 seconds |
| 1:15–3:11 | Lena | Paragraphs 3–4; 158 words. Observation and mnemonic correction. Pause twenty seconds on the arrival record and twenty on covered-list retrieval. | 73 / 40 / 3 seconds |
| 3:11–5:14 | Omar | Paragraphs 5–6; 173 words. Connector, lift and circuit. Pause twenty seconds at the blocked/readied mechanism and twenty on the powered trace. | 80 / 40 / 3 seconds |
| 5:14–6:19 | Priya | Paragraph 7; 90 words. Reference identity and permissions. Pause twenty seconds on the comparison and matrix. | 42 / 20 / 3 seconds |
| 6:19–7:02 | Jonah | Paragraph 8; 86 words. Handoff and supported capacity. Show B-CUST-01 and its actual recipient while speaking. | 40 / 0 / 3 seconds |
| 7:02–7:42 | Lena | Paragraph 9; 79 words. Baseline route and preserved trace. Point to B-ROUTE-01 while speaking. | 37 / 0 / 3 seconds |
| 7:42–8:44 | Priya | Paragraph 10; 83 words. Closure, missing cradle and changed authority. Pause twenty seconds on the paired H-3/H-8 records. | 39 / 20 / 3 seconds |
| 8:44–10:00 | Jonah | Paragraphs 11–12; 161 words. Revision, final result and limits. Show the west trace and F-RCPT-03 beside both plans while speaking. | 75 / 0 / 1 second |

At a slower delivery rate the team must rehearse and shorten the spoken script before presenting; it must not silently borrow time from another segment. The full records remain available for questions outside the ten-minute presentation. This is a complete written script and timed delivery plan, not a claim that it has been performed.

Role notes: Lena must distinguish a camera viewpoint from the fixed grid and the baseline zero from the revised one-contact trace; Omar must distinguish ratio from readiness and supply from output; Priya must distinguish reference identity from scientific truth; Jonah must identify Venn's copy receipt and Quill's continuing custody of O-HAL-01 at RD-2.

## Individual defences

### Lena — Observer / navigator (300 words)

My contribution was to make the arrival and route records specific for other roles to use. I recorded the 06:45 display, blue beacon and open hatch separately from the note claiming all plates were intact. The damaged outer case did not establish the state of every plate. I wanted the investigator to receive observations and questions, not a story that already decided the answer.

I also used the four-location mnemonic for the route kit. My first retrieval exchanged Lens and Spool. Recording that substitution mattered because the second attempt could otherwise look effortless and reliable than it was. I made the Lens-at-Basin association more distinctive and retrieved again with the list covered. The written kit remained available for verification; I did not claim that memory should replace a record when a record can be carried.

For navigation I preserved the exact route and the published sensor cells before executing it. The baseline path produced zero contacts under its stated pattern. That result was not a general property of the corridor or a real surveillance claim. When the gallery closed, route availability changed independently of the old sensor prediction. I had to update the map used by the coordinator rather than merely repeat the earlier zero.

My cross-role dependency was the systems specialist's support result. A route that reached dispatch could still be unusable for an unsupported load. I therefore asked whether the cradle condition held before recommending original transport. An alternative was to focus only on the shortest path, but that would have optimised a route for a movement the equipment could not justify. The final digital transfer preserved research access, while leaving physical provenance work for a later supported handover. I can defend that choice within our charter, but I cannot use this demonstration to claim physical field proficiency.

### Omar — Systems specialist (300 words)

I owned the connector, lift and support-circuit checks. For the connector I predicted the labelled mapping before moving it: a 270-degree clockwise turn took north/west to west/south, and the later move to level one did not change that orientation. Keeping those two requirements separate helped the observer use my result without relying on a similar-looking silhouette.

For the lift I calculated the follower rather than selecting a familiar part. Twenty-four driver teeth multiplied by six input turns, divided by four desired output turns, gave thirty-six follower teeth. The driven shaft turned oppositely. That calculation did not establish readiness: the interlock, cam and spring had their own conditions. I recorded their states before applying input so a later reader could distinguish a blocked mechanism from a wrong ratio.

The circuit trace showed nine volts before the cable and zero after it. I used that transition to localise the supply break, isolated power for replacement and then verified both supply voltage and light. The observations belong to this authored model. I did not measure a constructed apparatus and should not present the ideal values as physical test evidence.

My most consequential action was refusing to treat the unavailable cradle as a minor inconvenience. The original had mass nine and the unsupported carrier capacity was seven. A successful lift test did not change that capacity limit. The investigator depended on this distinction when comparing physical and digital recovery options.

A possible alternative was to wait for replacement support and preserve the original objective's physical preference. Our charter prioritised verified research access, so the coordinator could instead agree a digital transfer with stable custody. I supported that revision while recording that later physical movement still needed equipment verification. I would recheck the apparatus before any such handover rather than reuse this rehearsal's result indefinitely.

### Priya — Investigator (300 words)

I owned the relationship between a selected atlas and the authority that made it the required deliverable. Under the first signed reference, Aster's thirty sections and H-3 digest satisfied both conditions. Cinder's later annotation and H-8 digest made it different, not automatically better or malicious. Dawn's twenty-nine sections were incomplete against the request. I recorded the comparison before recommending a copy.

The storm rehearsal included a signed amendment requiring H-8. I changed the recommendation to Cinder while preserving Aster's earlier match. This was a change in the authorised target, not evidence that the original calculation had been wrong. Explicitly putting both references next to their selections made the revision understandable without pretending we had known the amendment earlier.

I also checked the local permission charter. All roles needed reading, the maintainer needed service and the custodian needed release approval. Negative tests mattered because a reader should not approve release. Positive tests mattered equally because denying everything would stop legitimate recovery. The charter remained separate from the archive amendment: changing the required content did not silently change every role's authority.

My cross-role dependency was the coordinator's mission contract. If the task had required immediate physical provenance, a verified digital copy alone would not have completed it. The stated charter instead sought research access while preserving the original. I therefore supplied a bounded verification result and let the coordinator negotiate the resolution rather than claim technical certainty answered every value question.

I would have challenged an unsupported statement that matching H-8 proved the scientific truth of the atlas. Identity against a signed reference is narrower than calibration or research validity. A useful next investigation would inspect the amendment's rationale and the original observation records. The final account records those limits so a successful copy transfer does not hide the questions still open.

### Jonah — Coordinator (300 words)

My responsibility was to turn technical results into a justified plan. I wrote the contract as verified research access with preservation of the original. Physical recovery was our baseline method, but it was not an requirement. That distinction gave the team room to revise the method without quietly changing what we claimed to have achieved.

I assigned each check an owner and identified who needed its result. For the environmental-pack handoff, the sender specified West Store, three units and AZURE. The receiver repeated all three fields before I accepted confirmation. A general statement that the team understood would not establish which instruction had been acknowledged. I treated delivery instructions as separate agreements rather than inheriting that confirmation.

The council made two constraints visible: the custodian retained responsibility for the original, and the unsupported carrier could not carry its nine-unit mass. The verified cradle made baseline physical recovery possible. When the cradle became unavailable and the east gallery closed, I stopped the transport proposal instead of asking the team to make the old plan happen regardless.

I negotiated verified digital recovery after the investigator applied the signed H-8 amendment. My cross-role dependency was explicit: I needed the investigator's current-reference result and the systems specialist's statement about support. Neither a correct digest nor a working lift alone authorised every recovery method. The revised plan therefore retained the original with a named custodian and required a recipient's new confirmation of the copy.

An alternative was to delay all delivery until replacement support arrived. Under the research-access charter, that would postpone usable information without improving the immediate digital check. We chose timely verified access while documenting the physical handover. I cannot claim that this resolved provenance research or calibration. My defence is that the chosen resolution met the stated contract within its recorded limits.

## Attribution

All characters, records, rules, narrative and submission examples are authored teaching material. The 3D scene and takeover controls represent those same fictional facts. They do not certify real-world security work, physical construction or a student's personal experience. A learner should transfer the method to the assigned case and submit their own recorded reasoning.

## Method to carry forward

Change the method when its dependencies fail while keeping the objective and unresolved obligations explicit.

Lamplight first tests physical recovery, then uses a portable digital carrier. Last Light offers terminal transmission, physical carriage or supported handover; select only the checks your chosen approach actually needs.

The complete example explains one defensible approach. In your own challenge, select observations and tests from the supplied rules, keep your prediction separate from the result, and justify both the checks you perform and those you omit.