Datasets:

---

maxwellinked

/

time-lapse-artifacts

Like 3

The current timing documentation omits 51 indexed recordings from its population summary, and the photograph guide directs readers to a field now retained in the prior projection manifest. This change reconciles the timing prose and restores the documented route through existing evidence.

At reviewed main revision 82789c0208741e1581b085c23c67684ce10a02f3, the existing timing validator reports 1,384 indexed files: 841 in the current view (607 filename-derived starts and 234 unknown starts), plus 543 outside it. The outside population comprises 492 historical records and the 51-record October 23 ingest. The separate wall-clock ledger records those 51 as unresolved active candidates; no start values are inferred or promoted.

Changes:

• Correct both current timing summaries and clarify the scope of not_in_timing_snapshot.
• Add a targeted regression check comparing both prose summaries with the validated projection counts.
• Follow the current photograph manifest's hash-bound prior_manifest_url for the retained 431-work selection evidence, requiring matching source/display hashes.
• Distinguish the paired projection's 355 populated source URLs from the separate 431-photograph download index.
• Improve the existing documentation index with separate Standard/Pre-Standard index links and routes to identity, photograph, timing and scoped October 23 validation evidence.
• Record the maintenance changes in the changelog.

Validation:

• 46 targeted timing, frozen-timing and archive-semantics tests passed.
• The new check fails when the stale outside count of 492 is restored in either edited summary.
• All 84 relative file links in the edited documentation resolve against the reviewed repository tree.
• All 431 selected source/display pairs match the scalar projection, download index and retained evidence entries; the prior manifest checksum matches the current manifest.
• All 431 download paths, sizes and LFS SHA-256 values match both the indexed download revision and reviewed main; all 76 new-JPEG source/output bindings match.
• Four source JPEGs downloaded, decoded and matched their indexed SHA-256 values. Four corresponding Viewer rows and generated image assets returned HTTP 200 at the reviewed main revision. Generated image delivery was checked separately; byte identity with embedded display JPEGs is not asserted.

Scope: six documentation files and one regression-test file. No media, metadata indexes, generated views, work identities, photograph selections, frozen evidence, tags or DOI snapshots change. No full media decode, full repository test run or archive-wide hosted-image delivery claim is made.