# The naming question — the change-control record

**Status** ⏸ **WAITING ON THE FOUNDER at stage 5 of 7 — one of the three questions is now
answered, two remain.**
**Opened by** brief 40: *"one of the first interesting questions to ask is, is the name of
the book correct? And I don't think it is."*
**Subject** *Creating a Book Using Fractal Semantic Graphs*, **v0.1.0** — the book's own
version, not the site's

This is the first run of the seven-stage workflow brief 39 asked for, and the first time
the team defined in brief 40 has been used. It is deliberately a small question, because
the point of a first run is to find out whether the machinery works.

> **This pack was renamed at v0.6.4.** It shipped as `v0.6.3__the-naming-question`, stamped
> with the *site's* version, for work reviewing a book at **v0.1.0** — a version that book
> has never had. The founder caught it: *"the url that you used for that name review is
> then confusing, what book is it about, and we didn't have v1 of the making book."* Three
> version streams run in this repository and a bare number does not say which. Book work is
> now stamped `<book-slug>__vX.Y.Z__<slug>`, and a gate enforces it.

---

## 1. Plan — which roles have a voice, and which are only asked

Brief 40 names this as part of the exercise: *"a good example of creating a plan and then
defining which agents should have a voice on this, and which agents we should ask for an
opinion."* The distinction was drawn like this:

| Role | Voice or opinion | Why |
|---|---|---|
| **researcher** | voice | The map is the input everything else argues from |
| **editor** | voice | A title is a promise to a reader, which is this role's centre of gravity |
| **publisher** | voice | A title is a store listing and a version move |
| **writer** | opinion | Must live with it, does not choose it |
| **librarian** | opinion | The old name must survive as a superseded name |
| **QA** | opinion | Asked only whether the change can be verified |
| **developer** | not asked | Nothing here is a code question |

The founder decides. No role approves.

## 2. Map — what the book actually is

[The researcher's map](../../team/researcher/debriefs/v0.6.3__map-the-making-of-book.md).
Its headline is computed, not argued:

> The title's two content words are the two rarest things in the book. *Fractal* appears
> **9 times in 31,221 words** and *semantic graph* **6** — against *release* 227, *graph*
> 193, *book* 182, *agent* 173. Six of the nine name the **other** book. Exactly one use is
> substantive, and even that one is about the other book's claim.

And the book had already said so itself, in its front matter: the title *"was set in the
commission"*, unlike everything else in it.

**The map also disagreed with the founder**, which is what a map is for. Brief 40 judges
the book leans explorer; the body contains **one code block and no shell commands** across
twelve chapters, with the technical material quarantined in an appendix that is 21% of the
words. As written it leans villager. That is either a finding about the title or a finding
about the body, and only the founder can say which.

## 3. Define — the candidates

[The editor's proposal](../../team/editor/debriefs/v0.6.3__the-title-proposal.md), built
only from words the book actually uses:

| | Title | Serves | Costs |
|---|---|---|---|
| **A** | *Writing a Book with an Agent* | villager first | a rename only |
| **B** | *The Loop* | either | a cover that says nothing alone |
| **C** | *Gates Buy Speed* | explorer | a real content pass first |
| **D** | *Ship, Then Improve the Ship* | either | ambiguous in print |

Deliberately excluded: *workflow* (8 uses), *fractal* (9). And the two words carrying the
founder's own thesis in brief 40 — *scaffolding* and *feedback loop* — appear **zero**
times in the book. They are the right ideas with the wrong evidence.

**Editor's recommendation: A, unless the founder wants a content pass, in which case C.**

## 4. Review — the opinions

- **[Publisher](../../team/publisher/debriefs/v0.6.3__opinion-on-the-title.md):** supports
  A. The Leanpub upload has not happened, so this is renaming a **draft** and the window is
  free until it is uploaded. The title lives inside the front matter and colophon, so the
  rename **is** a content change: the book's version moves v0.1.0 → v0.2.0. **Do not
  rename the folder slug.**
- **[Writer](../../team/writer/debriefs/v0.6.3__opinion-on-the-title.md):** supports A now,
  wants the C brief next. C means writing the scaffolding argument into the body with the
  estate's evidence: two to three thousand words. Separately: the colophon says *"the second
  book is not written"*, which was true at v0.5.11 and is false now.
- **[Librarian](../../team/librarian/debriefs/v0.6.3__opinion-on-the-title.md):** no
  preference. One requirement: **supersede, never delete** — the old title must be recorded
  as a former title, and the frozen narration in the archived version tables must not be
  edited.
- **[QA](../../team/qa/debriefs/v0.6.3__opinion-on-the-title.md):** supports any candidate,
  asked for two gates first.

## 5. Approve — the founder's, and open

### Answered by brief 42: it is a revision

The editor's first question was *"A or C — is the next edition a rename or a revision?"*
Brief 42 answers **revision**, and specifies its shape rather than leaving it open:

> The first part of the making the book should almost be a feature[s section]… literally
> describe, screenshot features, and basically say: here is what writing a book in 2026
> looks like… The current book is really cool at providing how we got there, but for a lot
> of people we need to show what THERE looks like. We need to show them what is the art of
> the possible. Then we show how we got there, because then they're motivated.

**Art of the possible first, history second.** That is a bigger change than either candidate
assumed: not a new opening chapter but a new opening *part*, and a reordering of what the
book is for. The writer's estimate of two to three thousand words was for putting the
scaffolding argument into the body; this is more.

It also moves the editor's recommendation. **C — *Gates Buy Speed*** was ruled second
because it promised an argument the body only half made. With part one written, the body
makes it. The map's finding stands though, and now cuts the other way: the book measured
**one code block across twelve chapters**, and a feature tour is where code and screenshots
belong. A revision that adds part one moves the book toward the explorer audience the
founder said it already served.

Nothing has been written. [The capability register](../v0.6.7__the-feature-tour/00__the-capability-register.md)
counts what part one could truthfully show, and names two things it cannot be written
without: **screenshots that show the workflow** (the 45 existing figures show the reader and
the engine, not the process) and **the founder's Leanpub failure story**, which is the
book's most persuasive evidence and cannot come from this repository.

### Still open

1. **The title itself.** Brief 41 offers ***Making a Book with Agents*** as *"a temporary
   hold"*. It is close to candidate A and says **agents**, not Claude. Is the hold the
   decision, or a placeholder while part one is written?
2. **Does the cover name Claude?** Brief 40 said *"writing a book with Claude"*; brief 41
   said *"making a book with agents"*. The book names Claude four times in 31,221 words.
   The two memos point different ways and the founder has not chosen.
3. ~~**Retitle to match what it is, or revise to match what it should be?**~~ **Answered:
   revise.** Brief 42 settles it.

## 6-7. Implement, and approve the implementation

Not started. Blocked on stage 5.

---

## What this release did do

QA's first requested gate is **built and shipped**, because it protects the rename itself:
a book is now checked to be called the same thing in `gen_bookmeta.REGISTER`, in its
`build.py`, and in its own front-matter heading. Nothing checked that before, which made a
rename the change most likely to leave one of the three behind.

It found a disagreement on its first run — the Universe volume heads its front matter with
the series name and puts its own name below. That is normal for a volume and not a mistake,
so it is now **declared in the register** as `front_matter_title` rather than hidden as a
special case inside a test. The gate was then run red against a half-finished rename before
being trusted.

**QA's second gate is not built**: file paths named in book prose must exist. It would have
caught chapters 4 and 15 naming `admin/tests/universe.test.mjs` the day it was deleted at
v0.5.20. It is the first tractable step into the estate's known hole — *prose has no
freshness gate* — and it is the recommended next piece of work whichever title wins.