One corpus, many content types

A corpus is wf-store's top-level collection of assets, tags and retrieval state. Everything deliberately uses one corpus for its wiki content, including articles, page templates, media and related assets.

It does not require one corpus for pages and another for templates. Nor does it maintain a second page registry. That would introduce coordination work without improving the basic reading experience.

Instead, tags describe the kind of asset, its logical identity, its route, its publication state and its collection membership. Page revisions retain immutable source while tags govern which revision appears to a public reader.

This lets the wiki contain many topics and many different kinds of content without fragmenting the underlying storage model. A topic is a way of browsing, not a separate storage boundary.

wf-store's internal generations, chunk locations and retrieval structures are implementation details. They do not become durable public page IDs or URLs. See The tag model and Assets, components and renditions for the details.