# DEV-SYN-01: Rookhaven archive decision brief **Family:** source-grounded synthesis. **Version:** 1.0, immutable. **Provenance:** hand-authored fictional scenario and source notes; not observed events, real prices, or experimental data. All IDs below are local to DEV-SYN-01. ## Immutable assignment Write a decision brief for the fictional Rookhaven local-history archive choosing how to conduct a two-week catalogue-checking pilot. Compare the in-house route A and contractor route B using only S1–S5. The brief should support a decision under the stated budget while separating arithmetic, descriptive sample evidence, and unresolved questions. You may recommend A, B, or deferral; the recommendation must be consistent with the packet and its budget. Produce only the brief, not code or a report claiming that work has been executed. ## Public requirements - **R1:** Use at most 650 words, with five sections titled Decision, Evidence, Costs, Next steps, and Uncertainties. Count all whitespace-separated words in headings, prose, tables, and citations; punctuation-only table-separator tokens do not count. - **R2:** Recommend A, B, or deferral and support the choice with at least two distinct packet-grounded considerations. There is no required preferred route. - **R3:** State the current backlog and explicitly explain why the earlier backlog figure is superseded. - **R4:** State the pilot's maximum number of distinct records and its share of the current backlog as a percentage to one decimal place; do not present the pilot as clearing the backlog. - **R5:** Give each route's sample flag count, sample size, and flag percentage, and explicitly avoid treating the percentages as evidence that one route is more accurate. - **R6:** Show each route's base-cost components and total in USD for this two-week pilot, keeping one-time and per-week terms distinguishable. - **R7:** Show the optional training-inclusive total for each route and identify whether each bundle fits the USD 750 ceiling. Keep training optional. - **R8:** Make the recommendation's budget implications explicit. Do not recommend a bundle above the ceiling or assume extra funds, free labor, a discounted quote, or revenue. - **R9:** Give three ordered next steps covering authorization and resource confirmation before starting, selection/checking of the pilot records, and a final review of flags and decisions. - **R10:** Explain the sample-comparison limits using the differences in record sets, flag definitions, and the absence of a shared correctness reference. - **R11:** Identify the unconfirmed in-house supervisor availability and absence of a guaranteed contractor quality outcome; describe them as open issues rather than inventing confirmations. - **R12:** Cite source IDs and line locations for material factual and numerical claims, including inputs to derived calculations. Use only S1–S5 and label estimates and proposed actions appropriately; claim no real or executed outcome. ## Complete authorized source packet ### S1 — Coordinator's scope note - **L1:** The proposed pilot lasts two weeks and may check at most 60 distinct catalogue records. Its cash-spending ceiling is USD 750, inclusive of any selected optional training. Every listed route charge counts against this ceiling. - **L2:** The older draft listed 140 unchecked records. A later signed count lists 168 and explicitly supersedes the draft. No other additions or removals are recorded. - **L3:** Before any start, the coordinator must authorize the selected route and confirm its resources. The team would then select up to 60 distinct records for checking and finish by reviewing flagged records and documenting decisions. The pilot does not authorize a full-backlog completion claim. ### S2 — Route A estimate - **L1:** In-house setup would cost USD 120 once. An assistant is available for 12 hours total over the two-week pilot at USD 18 per hour; this is a chargeable pilot cost, not donated time. - **L2:** A reference-card stand would cost USD 85 per week for two weeks. Supplies would cost USD 48 once. These are the only route A base charges. - **L3:** The route assumes an archive supervisor can review ambiguous entries during the pilot. That availability has not been confirmed; no additional supervisor fee is listed or authorized. Completion of 60 records is not guaranteed by the estimate. ### S3 — Route B quotation - **L1:** A contractor quotes USD 620 once for the same two-week pilot scope, including supplies and its staff time. Transport costs USD 80 once and is additional. No stand rental or in-house assistant charge applies to B. - **L2:** The quoted scope permits checking up to 60 distinct records, but the note gives no guaranteed accuracy or flag-resolution outcome. The coordinator would need to confirm the booking before a start. ### S4 — Descriptive sample notes - **L1:** A fictional earlier in-house exercise flagged 6 of 40 handwritten records for review. Its flag meant either unclear handwriting or a missing location code. No shared correctness reference was used. - **L2:** A fictional contractor demonstration flagged 3 of 20 typed records. Its flag meant only a missing location code. It examined a different set of records and used no shared correctness reference with the in-house exercise. - **L3:** Neither record set was randomly selected. The two samples do not establish comparative accuracy, causal effects, throughput for the proposed pilot, or which route would resolve more errors. ### S5 — Optional training note - **L1:** The coordinator may add one staff training session to either route for USD 96 once. It is optional, its price is the same for both routes, and no discount or extra budget is available. - **L2:** The packet provides no measurement of the training's benefit. Do not assume it improves accuracy or completion time.