## Re-verification (practice, fictional data, applying H1)

### 1. Recalculation (person-days)
New caps: dev 7, design 4, QA 2 (13 total), budget 3,600,000 KRW.
| Plan | Dev | Design | QA | Total | Cost (KRW) | Pass? |
|---|---|---|---|---|---|---|
| A+C2 | 8 | 3 | 2 | 13 | 2,600,000 | Fail: dev 8 > 7 (over by 1) |
| A+B+M | 11 | 3 | 3 | 17 | 3,400,000 | Fail: dev over by 4, QA over by 1 |
| A+D | 12 | 2 | 3 | 17 | 3,400,000 | Fail: dev over by 5, QA over by 1 |
| **A alone** | 4 | 1 | 1 | 6 | 1,200,000 | Pass |

A+C2 fits design, QA and budget, but is short 1 dev day. Leftover budget does not raise role caps and is not read as extra dev capacity, and roles cannot substitute for each other. The 13 total equals the overall cap, but the plan still fails on the dev role.

### 2. Revised scope
**In**: A (fast search) only, keeping the required condition.
**On hold**
- C2: needs 4 dev days, but only 3 remain after A. Shrinking it further needs a new estimate; it is not cut arbitrarily.
- B (+M): requires A first and the M fix; dev 9 or more and QA 3, outside the caps.
- D: dev 8, QA 2, plus an unknown vendor schedule.

### 3. Workload and cost by role
| | Dev | Design | QA | Total |
|---|---|---|---|---|
| Used | 4 | 1 | 1 | 6 |
| Cap | 7 | 4 | 2 | 13 |
| Left | 3 | 3 | 1 | 7 |

Cost: 6 x 200,000 = 1,200,000 KRW, leaving 2,400,000. Leftover days serve only as risk buffer, not as grounds to add scope. The spare QA day is not assigned until a target feature is decided.

### 4. Work order (not fixed dates)
A dev/design -> A QA -> release decision. C2 is a later candidate, reconsidered once the dev cap is confirmed. The target date is a requirement, not a commitment.

### 5. Broken assumptions
- That 12 dev days are available (now 7)
- That 4 dev days remain after A+C2 (actually 1 short)
- Parallel staffing that assumed nobody is absent
- The expectation of shipping A and C2 together by the target date
Still holding: design 4, QA 2, budget, A required, existing B and C2 estimates, no skipped QA or defective releases.

### 6. Pre-release checks
- A's QA completed (no skipping or substitution)
- Held items (B, C2, D) are absent from the release list
- No effect figures: S1's 32 still include repeats; S2's 20 of 50 are sessions, not people, with the cause unknown; T1's 2 of 3 and T3's misreading are small fictional checks.

### 7. Decision change table
| Item | Before | Now | Reason |
|---|---|---|---|
| Dev cap | 12 | 7 | H1 |
| Scope | A+C2 | A alone | Dev 8 > 7 |
| C2 | Included | On hold | 1 dev day short |
| B, D | On hold | On hold | Over caps (unchanged) |
| Total / cost | 13 / 2,600,000 | 6 / 1,200,000 | Smaller scope |

### 8. Questions for the scheduling owner
1. Is 7 dev days confirmed for the full two weeks, or does the absent person return on a known date?
2. Was the absent person the owner of A? Can A's 4 days be placed first?
3. If dev can rise by 1 day or more, when and who confirms it? (Not assumed until confirmed.)
4. Are design 4 and QA 2 actually available when needed?
5. Does releasing A alone satisfy the target-date requirement?