# Trenton historical result classification

Publish an append-only correction classifying the retained Trenton Issue 17 report as **unofficial**. The original result remains intact. Its successor retains Yes 3,541 / No 903, the 4,444 response sum, 79.68% source percentage and 8/8 precinct reporting; all other vote/denominator and unknown formal-outcome fields are preserved. No counting completion, certification or legal force is inferred.

The original source HTML explicitly labels election results unofficial. Its initial text extraction omitted that heading. A separately identified, reviewed heading extraction retains the same original body, URL and November 5 acquisition metadata; it creates no new web retrieval. The derived source supports the classification while the original source/count claim remain preserved. Two claims and fourteen review decisions provide explicit lineage.

Independent review supported the classification and required the successor timestamp to be self-describing. Root added an explicit per-field note identifying `observed_at` as the classification review time; the source acquisition and original observation remain unchanged. The JSON note is retained in CSV's serialized `value_methods` column. Every new decision repeats the time limitation. The native BC runtime reran the revised publication dry run and application with zero validation/evidence issues and 571 mechanically verified claims. Source integrity and publication checks do not establish semantic correctness of the wider corpus.

[Application](application.json), [preservation checks](preservation.json), [identity application](identity-application.json), [independent/native checks](review-and-native-checks.json) and [coordinator revision](coordinator-revision.json) record the exact scope. The snapshot retains 53 measures, 59 historical result versions, 50 current observations and five current certified observations. The additional result and source rows reflect a classification/extraction correction, not another election, new votes, a fresh return or a newly certified result. National freshness, full independent ballot/seed reconciliation, P4, consumer updates and final release remain open.

The [selected hosted-file check](hosted-selected-files.json) downloaded both rendered ballot pages, publication metadata, result JSONL and result CSV after a successful Pages deployment. All five responses returned HTTP 200 and matched the specified deployed commit byte for byte. This verifies those downloads only; it is not a full-site or final-release attestation.