Origins: 2026

Six and a half months, February to August 2026, dated from filenames and commit history. It is published for the same reason everything else here is: a claim is easier to judge when you can see how it was arrived at, including the parts that were tried and dropped.

1. 5 February 2026 · phase 0: pre-history The three foundational documents are written inside a different project entirely, Issues-FS, about issue tracking. They will not reach this corpus for four months. Everything at altitude 1 comes from one of them.
2. 21 February – 24 March · phase 1: graphs as infrastructure The first graph thinking is cryptographic, not semantic: trust as a key graph. It produces the earliest graph-native sentence in the corpus, “revocation is the absence of trust, not the presence of a revocation entry”, which is still one of the best. Paragraph-as-file appears on 23 February and then lies dormant for five months.
3. 25 March – 2 May · phase 2: graphs as the model of the system The first inflection. A graph is proposed as the source of truth about the product itself, which produces the reframe: “a bug is where reality diverges from the model.” An Ontologist role is created.
4. 18 – 31 May · phase 3: documents become graphs Compliance as a living graph; rules as a fractal graph; then the universal document-to-graph pipeline. Graphs stop being infrastructure and start being the product.
5. 1 – 5 June · phase 4: the concept explosion Four days produce skills-as-graph, skill-as-projection, semantic knowledge graphs of identity, trust-through-connectivity. The first “meaning through connectivity” in the founder's own voice in this corpus. And every one of these briefs assumes a philosophy that is not in the repository.
6. 10 – 11 June · phase 5: the import, half-executed An agent notices the assumption gap. Ten Issues-FS documents are imported and library/concepts/ is created (phase 1 of the import memo). Phase 2, adding the cross-reference so agents can find them, was specified and never executed. That is why this book exists; see below.
7. 10 June · phase 6: visualisation discipline The blob anti-pattern, verb edges, the subgraph flip: most of altitude 2, in one day. The founder asks for his own LinkedIn series on graphs as prerequisite reading. It is not provided. It still has not been.
8. 16 – 30 June · phase 7: the formalisation Peak density. Confidence through evidence, then a six-brief burst between 26 and 28 June: paths that read as language, directed edges and node explosion, twins, node type formulas, ontologies of ontologies, the grounding ladder. This is where the philosophy becomes testable.
9. 12 – 24 July · phase 8: the architecture A graph at every boundary: six properties falling out of one decision. Registers of registers; messages as graph transformations; sovereignty computed rather than claimed.
10. 28 July – 2 August · phase 9: the regulation build Doctrine meets a real artefact: every paragraph is a graph, applied to the EU AI Act. Paragraph-as-file is resolved, closing a loop opened on 23 February. Appendix A of the 28 July brief is the first and only place the corpus writes its own concepts down in one list, the direct precursor of this book.
11. 6 – 9 August · phase 10: meaning itself Concepts, not words. Decompilation, not compilation. The author as the only oracle, and “that is not what I meant” reframed as success. And the diagnosis that produced this book: the people who will use this do not know what an ontology is.

The routing failure, which is the reason this book exists

It is worth stating exactly, because it is mechanical and it is fixable, and because the diagnosis generalises.

The generalisable version: a document that nothing links to is, for a reader who follows links, a document that does not exist. That is the same finding an agent reported when it tried to consume the parent project's documentation and could not follow the links, and it is why the companion site's llms.txt (its index file for AI agents) carries facts rather than topic labels.

Paths not taken

The provenance discipline this book argues for, applied to itself. These were real proposals by capable people and they did not happen; recording that is cheaper than letting someone rediscover them.