Knowledge Weaver: how dmitrii.dk is published

A record of the content collections, public release log and editorial checks used for this site.

Filed under
Case study · Knowledge Weaver
Published
Language
EN
Reading time
7 min
Version
v0.9.3

Knowledge Weaver is the publishing structure behind dmitrii.dk. It connects project cases, journal entries, media archives, reading trails and the public release log.

Public structure

The project shelf records status, subject and whether a full case is available. The journal stores a title, language, theme, reading time and publication record for each entry. Media pages keep publication details and archive links together. Reading trails connect existing material around a question.

Long articles share a reading shell with publication facts, related links and an AI involvement note. Index pages use their own layouts because a project ledger, a chronological journal and a media archive carry different information.

Content records

Published writing lives in MDX files. Frontmatter supplies the metadata used by indexes and article pages. The build validates required fields before a release can be produced.

This gives the editor a reviewable record of copy and metadata changes. It also keeps the public archive independent of a third-party writing interface.

Release record

The AI changelog lists each site version with its date, toolchain, AI involvement level and human editor note. The homepage reads the current version from that public record.

The log names Codex for v0.9.3 because it handled the implementation and copy pass under Dmitrii's direction. Earlier entries retain the tools recorded when those releases were made.

Technical foundation

The public site uses TanStack Start and React. MDX holds the articles, TypeScript defines the content model, and CSS supplies the visual system. Production runs on Cloudflare Workers.

Release checks include content metadata validation, redirect validation and a production route smoke test. They can find broken structure and unavailable pages. They do not judge whether an argument is good or a product claim is supported, so editorial review remains separate.

Current limits

There is no public site search or multi-author editing workflow. Readers move through indexes, journal themes, related links and trails. Some project entries remain short records until there is enough verified material for a case study.

The public release log is the authoritative history of changes to this site.