Preservation audit#
The replacement manual separates current behavior, contributor procedures, research leads, and dated evidence. The original book remains available during review. The file inventory and supporting-file inventory record the starting material; a file-level destination alone is not a completeness check.
For each original section, the audit records an explanation or reference destination, a retained historical record, an open investigation, or a reason the claim was superseded. Source links preserve precise evidence; they do not replace an explanation of current behavior.
Subject audits#
Complete inventories#
Section inventory: 1,150 headings across all 91 original Markdown pages, with source ranges, current destinations and retained records.Supporting-file inventory: 126 entries, including shared examples, assets, bibliography and build inputs.Compatibility mapping: 1,306 original HTML page and fragment targets. The assembled site retains these targets and their archived counterparts; the assembler verifies them at every build.External-link audit: the full check’s unresolved access failures and the verified replacement for the refused SOI paper host. These are separate from the passing local/Git-link checks.
The section inventory distinguishes current explanations (preserved), material
split between the manual and its retained record (split), dated history
(history), primary-source notes (literature), and investigations (research).
The archive catalogue links each original page at the
audit snapshot. The record role pins that snapshot separately from the
implementation revision used by source links.
Completion checks#
Each subject needs its current contracts, ownership, limitations, worked examples, paper mappings, and relevant test obligations accounted for. Experiments retain their original revision, configuration, run identity, outcome, and qualification limits. Rejected attempts remain discoverable. User acceptance is recorded separately from source inspection and build checks.
These inventories establish destinations, not independent certification of every historical sentence or experiment artifact. The subject maps identify corrected claims; evidence work names unresolved provenance and checking limits. The replacement is already the main local book; reader acceptance and website publication remain separate steps.