The Correlation Network isn't a static diagram — it's rebuilt from the site's own live content every time something is added. This page is the changelog: every dated entry below is a real diff between two snapshots of the actual graph, computed from this project's own history, not a summary written after the fact.
Every entry above is a real diff between two point-in-time snapshots of data/constellation.json — the exact same file that draws the Correlation Network graph. A node is a case, key figure, location, theory, ancient record, or book; an edge is a real cross-reference between two of them, pulled directly from this site's own links.
The first nine entries were reconstructed from this project's own git history, going back to the day the Correlation Network itself first launched. Every entry after that is appended automatically as part of the site's standard publishing routine, the same moment the graph itself is regenerated — nothing here is written or estimated by hand.
This log tracks how big the graph has become. Patterns in the Archive asks what it now says: which figures and places are genuinely load-bearing across independently-researched cases, whether reported craft shapes drift by decade, and whether anything observable actually predicts a case's Data Integrity Score. Those are sample statistics, recomputed from scratch every time an entry below is added — so each growth step here is also a step toward the subgroups over there being large enough to mean something.