# Chapter 4 · Twelve techniques, and what admits them · brief · 13 September 2026

The fourth chapter of the _Creation_ track, and its second merge.
`living-scenes-technique-catalog` (25 slides) and `living-scenes-compatibility`
(19) become one fifty-two-slide chapter under the new id
`living-scenes-technique`, per the Phase D rule that merged chapters take a new
id and keep every slide id. Nothing merges away: all forty-four inherited slides
are rewritten in place and eight are new.

## The job

A boundary between two segments raises two different questions, answered by
different code. **May these two sit adjacent?** is the compatibility scorer —
four policies and seven equally weighted dimensions over the segments' own
descriptors. **May this technique play here?** is the tone gate — a template
allowlist, a crisis frame, a Metis lock and an operator grant. Between them sits
the catalog: twelve techniques of twelve fields each, composing nine validated
primitives, with five continuity thresholds apiece.

**Running example.** One boundary — the sky segment handing over to the lesson —
named, composed, measured, scored, and admitted or refused. Executed on this
host against the pinned modules: the catalog and its freeze depth, the
transition validator at both edges of all nine ranges, the continuity gate, the
scorecard adapter, the per-technique gate, the compatibility scorer over fifteen
pairs and the running pair at five threshold settings, the tone helper over all
sixty technique-and-band cells, the admission gate over nine candidate filters
and five refusals, and the override constructor over eight inputs.

## Three learning promises (the cover)

1. **A technique is a record** — twelve fields, including five thresholds and a
   provenance tag.
2. **One metric is refused, not defaulted** — no in-repo FVD model exists, so
   the adapter demands a value or an explicit sentinel.
3. **Policy answers before arithmetic** — four gates refuse a pair whatever the
   seven dimensions average to.

## Three things to remember (the close)

1. **Ask who supplied the number.** Each technique carries five thresholds; the
   compatibility scorer ships none, and at the pin only its own test sets one.
2. **Ask which gate refused.** Four policies, four gates and a malformed
   threshold pair all produce refusals, and the candidate filter throws every
   code away.
3. **Ask what the call left behind.** Void, a missing model, and a freeze one
   level deep.

**Bridge.** The next chapter takes up compose assist and catalog governance —
who may suggest a technique, and who may change the twelve.

## Sections

Six sections, plus the cover and the close.

| Section                         | Slides |
| ------------------------------- | ------ |
| What a technique is             | 5      |
| The twelve, and what each costs | 6      |
| The primitives underneath       | 5      |
| Continuity, and what survives   | 9      |
| What a pair is asked            | 9      |
| Verdicts, templates and gates   | 10     |

**A deviation, recorded.** The `section-divider` layout admits two to six
sections, and six sections of seven would cover only forty-two of the forty-four
content slides. The validator's cap wins: the last three sections run to nine
and ten slides rather than the usual five to seven.

## Treatment decisions

- **Three legacy graphs retire.** `scene-tech-layers`, `scene-compat-pipeline`
  and `scene-compat-gating-order` become `graph-diagram` compositions. Counted
  from the builder's own output afterwards, **nine** slides still render as
  legacy graphs library-wide. One of them is in a Living Scenes guide —
  `scenes-keep-share` in `living-scenes-governance`, which Creation chapter five
  reaches — and the other eight sit in privacy, Telegram, payments, native,
  Veritas, Metis, operations and V2 guides that Phase D has not rewritten.
  (Chapter three's receipt said three remained in Living Scenes guides and
  twelve library-wide; it missed `scenes-keep-share`, so the figure was four and
  twelve. Three of those four close here.)
- **Both source manuscripts retire.**
  `content/48-living-scenes-technique-catalog.json` and
  `content/49-living-scenes-compatibility.json` held nothing but their own
  guide's slides. The decks of both guides are removed.
- **Seventeen inherited tables become nine.** The ceiling is one table in four
  slides. The two launch-catalog threshold tables and the two template matrices
  stay tables because a reader needs those values side by side; the rest become
  stat panels, compare panels, card grids and graph diagrams.
- **Eight slides are new**: a cover, six section dividers and a close.
- **One real capture, and it answers the chapter's own objection.** Every module
  here is pure and none has a non-test caller, so it is fair to ask whether any
  product surface applies these rules. One does: `ComposeClient` builds its
  boundary-technique picker by running the Compose Assist twin of the admission
  filter over all twelve ids. A new harness,
  `tools/presentations/tests/technique-admission-capture.mjs`, renders the
  shipped component from the pin and reads its options back off the DOM — seven
  offered, and `smash-cut`, `whip-pan`, `jump-cut`, `cross-cut` and `montage`
  absent from the document entirely. `scene-compat-enforcement-filter` becomes a
  `capture-callouts` slide over the alternatives panel.
- **The bespoke review harness retires.** `living-scenes-techniques-review.mjs`
  named four files that no longer exist, was never in the Node suite's glob, and
  resolves `axe-core` from a package that does not carry it on this host. The
  merged chapter is swept by `eve-edition-review.mjs` at four viewports and by
  the library-wide `presentation-center.mjs` and `eve-design.mjs`, which run the
  same readability audits. Its review config is repointed at the merged guide.

## Two inherited claims corrected

- **"Sixty agreements" between the two allowlist tables.** The catalog and
  Compose Assist hold the same policy in opposite orientations, and executed
  over all sixty template-and-technique combinations the two **decisions** agree
  60 of 60, with and without the operator grant. The **arrays** do not: the
  catalog lists all twelve for Arete while Compose Assist gates three of them in
  code. The agreement is between the gates, not the tables, and three constants
  — the per-template allowlists, the crisis trio and the operator-curated Arete
  set — are kept in step by hand.
- **The override record's limits.** Executed, `recordOverride` accepts a blank
  artifact id, both segment ids naming the same segment, and a NaN timestamp,
  and `Object.freeze`s only the outer record — reassigning the nested verdict's
  `kind` succeeds. A durable pattern already exists in this repository, in
  `review-persistence`'s decision lifecycle, which writes through a repository;
  the compatibility override is a constructor that resembles one and is not.

## What this chapter does not establish

No frame was rendered, nothing was measured from media, and no boundary was
played. Every module executed here is pure: it returns a decision and performs
nothing. The continuity numbers are **inputs** to the gate, and the scorecard
adapter says so in its own comment — _"fvd ← a DECLARED-ABSENT model-bound seam:
no in-repo FVD model exists, so callers must inject a real value or an explicit
policy sentinel — this adapter never fabricates one."_

Neither the blend kernel's compatibility scorer nor its `assertTechniqueAllowed`
has a non-test caller at the pin. The one product surface in the chapter reaches
the same policy through Compose Assist's own copy of it, not through the kernel
functions the rest of the chapter executes — so the capture shows that a member
is offered seven techniques, and does not show that the kernel decided it. The
chapter states that on its own slides rather than only in this brief.
