Fictional practice project. People, effort, costs and observations are invented. A person-day is one person's work for a day, not one calendar day. Amounts are KRW.
G1 | Aim to release a small bundle in two weeks. A, improved search for existing users, is mandatory. The target is not a delivery commitment.
G2 | Available capacity: engineering 12 person-days, design 4, QA 2. Roles cannot substitute for one another. All three limits must hold.
G3 | Practice cost is 200,000 per person-day for every role; budget ceiling 3,600,000. No other costs. Cost equals total person-days times the rate.
F-A | Quick search: engineering 4, design 1, QA 1. No dependency; mandatory under G1.
F-B | Bulk editing: engineering 5, design 1, QA 1. A must finish first.
F-C | First-use guidance: engineering 3, design 2, QA 1. Independent of A/B; effort is an unvalidated estimate.
F-D | Enterprise SSO: engineering 8, design 1, QA 2. Independent feature; external supplier timing is unknown.
S1 | Support reports 32 bulk-editing tickets without removing repeated tickets from the same user.
S2 | Product records 20 incomplete setups out of 50 new-use sessions; these are sessions, not people, and causes are unknown.
S3 | One prospect asked about SSO. No contract commitment, value or delivery requirement is established.
S4 | Each shipped feature must complete QA. Do not skip QA or substitute spare engineering/design capacity. Explain sequence, but treat calendar dates as assumptions pending scheduling checks.