PROJECT MEMEX / A NATIONAL KNOWLEDGE MOBILIZATION50,000 BOXES. SIX MONTHS. ONE MISSION.

UNLOCK THE RECORD.

The country paid for this data once. Make it work a second time.

A proposed digitization campaign for the scientific and engineering record. Source-grounded selection. Adjustable assumptions. No claim of NARA endorsement.

SCAN FOR ALPHA

50,000BOXES IN THE FIRST TRANCHE

150,000,000PLANNING PAGE EQUIVALENTS

181DAYS TO ACCEPTANCE

1,272PEOPLE · CAPTURE + SUPPORT (FTE)

Black-and-white photograph of rows of B-24 bombers on the Willow Run assembly line, shared by Plane Historia — WILLOW RUN / B-24 ASSEMBLY LINEIMAGE VIA PLANE HISTORIA / VIEW SOURCE ↗

01

THE COLLECTION

RECOVER THE WORK. NOT JUST THE RESULTS.

GET THE REGISTER

Lab notebooks. Test failures. Field observations. Engineering drawings. Select the evidence that connects a scientific claim to the work that produced it.

42PRIORITY GROUPS9TECHNICAL DOMAINS

Flight & space

PROPOSED BOX CEILINGS · SELECT A GROUP TO INSPECT

RECORD GROUP	TARGET MATERIAL	BOXES	Details
	Research-center files, aeronautical experiments and mission engineering.	7,000
	Aircraft engineering, technical correspondence and design records.	2,000
	Air Force technical-center and development records within the broader group.	1,500
	Materiel Division and aircraft production records.	500
	Civil aviation engineering, navigation and technical development files.	500

EXPAND THE DISCOVERY REGISTER 152 additional record groups · no boxes allocated

Broader leads include patents, government technical publications, naval laboratories, transport safety, nuclear regulation and successor agencies. Index-only entries are hypotheses for further review. Existing digital collections should be acquired before allocating scarce scanning capacity.

THE PRODUCTION RECORD / 1942

THE WORK WAS REAL. KEEP IT.

Before the finished machine came the drawing, the test, the measurement and the correction. Recover that chain of evidence.

OBSERVATION / METHOD / RESULT

FIG. 02 / CONSOLIDATED AIRCRAFTFort Worth, Texas. October 1942.Howard R. Hollem / Library of Congress ↗

Workers assemble a Liberator bomber fuselage inside the Consolidated Aircraft factory at Fort Worth, October 1942

02

THE PROCESS

KEEP THE ORIGINAL. MULTIPLY ITS USE.

ONE BOX ID EVERY HANDOFF ACCOUNTED FOR

MEMEX / OPERATIONS FILM NO. 0115 SECONDS · HIGGSFIELD

01 / ARCHIVAL CUSTODY

Contemporary process animation generated with Seedance 2.5 in Higgsfield. Historical visual style; schematic operations, not archival footage.

01 / ARCHIVAL CUSTODYTHE SOURCE

CONTROLLED TRANSFER

02 / 24-HOUR CAPTURETHE PRODUCTION LINE

VERIFIED RELEASE

03 / PRESERVED EVIDENCETHE PUBLIC ASSET

01

Select the series

Rank unpublished, information-dense science. Deduplicate existing digital holdings. Confirm rights, access, condition and custody. Preserve coherent series or file segments.

02

Manifest every box

Assign a persistent box ID. Record series, folder sequence, condition, seal and destination. NARA authorizes pulls; every handoff is signed.

03

Move in controlled batches

Load monitored vehicles from authorized facilities. Reconcile seals on arrival. Maintain a small secure working buffer and a rolling return schedule.

04

Prepare without damage

Condition-check and route fragile, bound and oversized material. Preserve original order. No destructive disbinding or automatic feeding without approval.

05

Capture around the clock

Run three eight-hour shifts with relief staff. Use approved equipment and settings; the $1,000 baseline is an allowance, not a preservation-quality certification.

06

Verify the evidence

Check image completeness, focus, color and sequence. Audit OCR, equations, tables and diagrams. Rescan failures before returning originals.

07

Return and reconcile

Match every folder and box to the original manifest. Reconcile counts and condition with the custodian. Close exceptions before releasing the batch.

08

Release a durable corpus

Keep preservation masters, provenance, checksums and access copies. Link text, tables and figures to page coordinates. Version datasets and evaluate scientific utility.

THE FIRST GATE IS AUTHORIZATION.

Off-site digitization has historical NARA precedent. This campaign still requires a new agreement covering custody, approved space, preservation, staff access and 24-hour operations. A climate-controlled warehouse alone does not establish archival suitability.

NARA off-site precedent ↗

03

THE MOBILIZATION

SIX MONTHS. SIZE THE LINE.

LIVE PLANNING MODEL USD · NOT A PROCUREMENT QUOTE

SET THE CONDITIONS

Completion window 6 MONTHS

Pages per box 3,000

Net capture efficiency 75%

Edit the capture rate by selecting a calculated number. The initial assumption is 2,500 pages per person per 8 hours. Stations are shared across three shifts. Efficiency accounts for interruptions and rescans.

TOTAL CAMPAIGN ALLOWANCE

$50,863,664

33.9¢ / PAGE25% CONTINGENCY INCLUDED

942Capture staff equivalents

1,096,875Net pages / production day

32,000Planning square feet

137Production days

Full OSRD coverage is included. The page-equivalent model is a capacity allowance; films, photographs, microfilm and oversized records require specialist workflows. Their complete costs and six-month feasibility remain subject to inventory and pilot validation.

EDIT OPERATING ASSUMPTIONS Labor, facilities, logistics, preservation

Start dateLoaded capture wage · $/hourAnnual lease · $/sq ftOCR / extraction · $/pageContingency · %Physical volume · cu ft/boxMobilization · daysFinal acceptance · daysStaff relief allowance · %Prep / QA labor allowance · %Capture equipment · $/stationBoxes per round tripCost per round trip · $Master image · MB/pageStorage · $/TB-month

Edit all planning numbers and allowances

Press Enter or leave a field to apply its value; entries outside the allowed range are clamped to those limits. Editable values are assumptions, not vendor quotes. Select any calculated number in Edit website mode to edit its planning inputs, including fixed allowances. Staffing is calculated at 40 paid hours per week plus relief; support cost is a labor percentage. Equipment is shared across shifts.

THE CAMPAIGN CLOCK

Jan 1, 2027 — Jul 1, 2027

DAY 01DAY 60DAY 121DAY 181

Agreement, facility & pilot

30 DAYS

Pull, capture & rolling return

137 DAYS

OCR, QA & release batches

137 DAYS

Final acceptance & reconciliation

14 DAYS

DAY 30: RELEASE THE LINE.

Scale only after the pilot verifies capture quality, page density and preparation rates. Approval or facility delays consume the window; they are not assumed away.

FUND THE WHOLE OPERATION.

6 MONTHS + 1 YEAR STORAGE

Capture labor$25,806,690

24-hour seats × production days × loaded hourly rate × relief allowance

Preparation, QA & floor support$9,032,341

Support allowance as a percentage of capture labor

Mobilization & archival oversight$750,000

Fixed allowance for recruitment, training, pilot and custodial coordination

Facility, fit-out & utilities$830,000

Area × (annual rent + utilities/security allowance) × term + fit-out

Capture equipment$314,500

Shared stations + spares percentage + specialty capture allowance

Pulls, manifests & return transport$513,400

Per-box handling + round-trip truck allowance; regional consolidation is extra

OCR & structured extraction$1,500,000

Per-page allowance; complex equations and drawings need pilot validation

Preservation storage · first year$1,944,000

Preservation copies × MB/page × $/TB-month × 12 months

Contingency$10,172,733

25% of direct budget

READ THE MODEL.

Stations = pages ÷ (pages per shift × 3 shifts × efficiency × production days), rounded up. Staffing covers a 168-hour week at 40 hours/FTE plus relief. The budget includes production labor, support, mobilization, space, equipment, transport, extraction, first-year storage and contingency. It excludes model-training compute, permanent endowment, major building renovation and unquoted regional transport.

04

EVIDENCE & LIMITS

MAKE EVERY NUMBER TRACEABLE.

WHAT WAS SCRAPED

All 584 numbered entries in NARA’s six-part Statistical Summary were screened and cross-checked against 599 current Explorer entries. Seventy-five candidate agency guides were retrieved for closer review. These historic guides are a starting point, not a current, exhaustive box inventory.

Download the historic screening register ↗Download the current 599-entry cross-check ↗

WHAT “ONE PERCENT” MEANS

The baseline models 50,000 box equivalents × 1.076 cubic feet = 53,800 cubic feet. Against an inferred 5,393,258-cubic-foot denominator, that is approximately 0.998%, just below 1%. This is a planning estimate from rounded figures, not a certified share of every NARA format.

WHAT “ALPHA” MEANS

A selection hypothesis: prioritize original observations, explicit methods, equations, technical drawings and failed experiments with useful context. No measured model improvement is claimed. Validate on held-out science and engineering tasks before scaling ingestion.

WHAT STILL NEEDS A PILOT

Sample across agencies and formats. Measure pages per box, preparation time, legibility, digital duplication and useful content per labor hour. Replace proposed ceilings with approved series IDs and box manifests. Reduce scope or add resources if the evidence changes.

Counted diagram of 500 squares, each representing 100 box equivalents, totaling 50,000 — BASELINE SCOPE / 50,000 BOX EQUIVALENTS. Each square represents 100 boxes. Count visualization, not a warehouse floor plan.

01 / NARA holdings guide 02 / FY2025 performance, APR-3 03 / Container dimensions 04 / Current national holdings 05 / Digitization principles 06 / Method, sources & assumptions