How Everything works
A small interpreter over wf-store, with explicit ownership and repairable content state.
Everything is an application layer over ordinary wf-store APIs. It adds wiki interpretation: logical pages, routes, publication, templates, safe public discovery and owner tools. The storage service does not need to become a wiki database.
Architecture overview follows a request through the edge, application and storage. The tag model explains where identity and state live. Repair over prevention describes the tradeoff behind ordinary tag corrections.
The wf-store section introduces immutable assets, renditions, readiness and independent retrieval families. Security and ownership explains why anonymous reading does not imply anonymous editing or raw-storage access.
The central constraint is simple: one corpus contains all wiki asset types. There is no parallel content registry, per-type database or static HTML publishing system to keep synchronized.
In this section
- Architecture overview — Follow public requests and owner writes through the current system.
- The tag model — Tags define wiki meaning; content JSON carries the writing.
- Repair over prevention — Prefer observable state and ordinary correction over elaborate consistency machinery.
- wf-store — The asset and retrieval service underneath Everything.
- Security and ownership — Anonymous readers get public wiki operations; only the configured owner gets writes.
The responsive reading path
The application returns a navigation shell and skeleton, loads the article separately, then loads related results near the viewport. Its warm caches reuse immutable page bytes, template selections/definitions and SDK setup. Page selection remains live; template staleness is deliberately tolerated.
- Caching and freshness — Understand one-call warm fetches, source-byte expiry and instance-lifetime templates.