book.surface.v1 — a static WRL profile for the book's scenes, proposed (read-only inventory)

Status: PROPOSAL. Nothing here is added to `WRL/relation-v2.js`, no identity is minted, no runtime is claimed. Written 2026-09-12 for the WRL lane, in the form of graphonomous/v2/handoff/research/R17_WRL_SOURCE_INVENTORY.md, to be adjudicated beside the pending graphonomous.semantic.v3 obligation. Measured at WRL 32160fec77a1 and opensentience.org 0a7717c76f44. Every count below is derived by the script in the appendix from opensentience.org/_patterns/scenes/*.json; none is typed.

The question. Travis asked (2026-09-12) whether the book's Structure — on the surface animations — 32 scenes over six stencils — can be "coded in WRL". This document answers it the way the tree answers vocabulary questions: by measuring what the scenes state, checking what an existing profile can carry, and stating the minimal obligation for a new one, so the WRL lane can judge the row against what it would have to hold.

0. What a scene is, in its own words, and why that bounds the inventory

A scene is data: a stencil that lays out a compute surface (carriers, slots, stations, wires, meters, rings) and a list of steps, each with a caption, an optional takeaway, and generic actions that mutate object attributes or move loci between named positions (_patterns/surface/surface.mjs; plan §6). The engine executes it. It is labelled an illustration on every page; the witness beside it is the evidence.

Two facts bound what WRL can do with it. (1) A static V2 profile "implies no runtime and derives exactly { rulepack_id }" and is "refused by the text surface" (WRL README, the profile table). So a scene under a static profile gets identity — a sem- id for its topology — and nothing executes it but the engine. (2) The only lowered profile is forge.world.core.v1 (five roles, two edge kinds). Ten of the 32 scenes already have an honest world in it and a Film the forge reduced (plan §7 P10); the other 22 do not, for reasons each page states. Making the forge execute a locus-on-carrier scene would need a lowered profile with runtime semantics in TRVM — a new runtime, not a row — and is out of scope here.

1. The registries, with their enumeration points

Enumerated by the appendix script over scenes/*.json (the only authority for what a scene contains):

Per stencil, the ops it uses: migrate state 34, move 10, tag 6, count 3, label 1, spawn 1; join move 30, state 27, join 22, note 17, spawn 12, tag 4; multiplex move 44, state 16, tag 4; wire state 39, note 38, move 26, wirenote 7, spawn 5, remove 3, tag 1; progress meter 14, place 6, state 5, move 1; nest state 25, note 5, move 4.

2. Not authority (excluded from identity)

• Anchors and coordinates. The 36 anchor names are layout — where a stencil draws a slot — exactly as

layout is not meaning in WRL. They would not enter a world.

• Captions, takeaways, intervals, colours. Prose bound to frames and presentation. A scenario document may carry

them; the world's id must not.

• The engine. surface.mjs is the executor and is versioned by its bytes in artifact.json; it is not part of

any scene's identity.

3. What forge.world.core.v1 can and cannot represent of this family

Can: any scene whose semantics are signals over wires into gates, rotors and poses — the ten records that carry a wrl.world today (join → Door with one input port; observable → Spinner rotor; projection → Spinner→Orb; identity/meaning/replay/sizes → the starter world). Their Films are on the pages, reduced by the forge.

Cannot: a locus that moves between carriers (there is no carrier role — the runtime is the carrier); a mailbox in source (Mailbox has no surface spelling, Core 0.1.2 §14b); meters (progress is a property of a Film, not a topology); rings (boundaries are not routes); a join whose refusal names a field (a Door refuses at a port). That is 22 of 32 scenes, each page carrying its one-line reason.

4. The minimal book.surface.v1 obligation, measured

Exactly what the registries in §1 state as topology, and nothing more. Attributes that steps mutate (state, note, tag, count, value) are role attributes, as v0's roles carry ["node"]; they are not relations. Movement (move, place, spawn) is a change of one relation, PLACED_AT. The row, in the shape of GRAPHONOMOUS_V0_ROW:

{
 "book.surface.v1": {
  "derivation": "static",
  "rulepack_id": "book.surface.rules.v1",
  "policies": [
   "book.surface.rules.v1"
  ],
  "domain": "surface",
  "signature": {
   "orientation": "directed",
   "texture": "solid",
   "arity": 2,
   "endpoint_roles": [
    "source",
    "target"
   ]
  },
  "roles": {
   "LOCUS": [
    "node",
    "state",
    "tag"
   ],
   "CARRIER": [
    "node",
    "state",
    "label"
   ],
   "SLOT": [
    "node",
    "state"
   ],
   "STATION": [
    "node",
    "state",
    "note"
   ],
   "WIRE": [
    "node",
    "state",
    "note"
   ],
   "METER": [
    "node",
    "value",
    "max"
   ],
   "RING": [
    "node",
    "state",
    "note"
   ],
   "MAILBOX": [
    "node",
    "state",
    "count"
   ]
  },
  "endpoints": {
   "PLACED_AT": [
    [
     "LOCUS",
     "CARRIER"
    ],
    [
     "LOCUS",
     "SLOT"
    ],
    [
     "LOCUS",
     "STATION"
    ],
    [
     "LOCUS",
     "RING"
    ],
    [
     "LOCUS",
     "WIRE"
    ]
   ],
   "JOINS": [
    [
     "LOCUS",
     "STATION"
    ]
   ],
   "COMPOSES_WITH": [
    [
     "LOCUS",
     "LOCUS"
    ]
   ],
   "VERIFIED_BY": [
    [
     "WIRE",
     "STATION"
    ]
   ],
   "MEASURES": [
    [
     "METER",
     "CARRIER"
    ]
   ],
   "CONTAINS": [
    [
     "RING",
     "RING"
    ],
    [
     "RING",
     "LOCUS"
    ]
   ]
  }
 }
}

Sizes it would have to hold at 0a7717c76f44: 8 roles, 6 kinds, 11 endpoint pairs; 32 worlds (175 objects across them); a scenario document per world carrying its 156-step film-shaped sequence of relation and attribute changes (94 of them with prose). None of these numbers is a promise; they are the sizes of the scene registry as measured.

What the row derives: { rulepack_id } and nothing else — no runtime, no Film. What it refuses: the text surface; a scene world would be V2 relation IR (JSON), converted from the scene file by a book-side adapter, sealed by relation-v2.js, and its id printed beside the animation. Adding a row "changes no existing identity: selection is a hasOwnProperty on the key" (relation-v2.js).

5. What the book does if the row is accepted, and if it is refused

Accepted: an adapter scene → V2 world + scenario; the build seals every scene's world through relation-v2.js (P16: a scene whose world does not seal is refused), prints the sem- id and a scenario digest beside the animation, and the edition's artifact.json carries the 32 ids. The engine remains the executor and the page keeps saying so. Cost: one adapter, one build step; identity is cheap at any size.

Refused: the scenes stay JSON, hashed into artifact.json as they are today — deferred with their coordinates preserved and no edge, the D-073 pattern — and the deferral is reported on the conclusion page.

Either way, the strong path for Part III is not WRL. The Progress scenes are about loci on carriers, and the executor of that is ComputeDriven's floors, whose runs exist as receipts with instruction counts (computedriven/receipts/R7-EXECUTED.md). A stage-B analogue for Carrier Multiplexing is a replay of a ComputeDriven transcript, and it is listed in the plan as such.

6. Not done here

No row added; no sem- id minted under a profile that does not exist; no claim that WRL executes a scene; no change to any existing identity. The graphonomous.semantic.v3 obligation is older and pending; this proposal does not jump it.

Questions for adjudication

1. Is a static row the right instrument for a vocabulary that has a runtime elsewhere (the book's engine), or

does WRL-P0 intend static rows only for vocabularies with none?

2. Are attributes-on-roles (state, note, tag, count, value) admissible in a static row, or must every

mutable thing be a relation?

3. Should a scenario (the step sequence) have WRL identity at all, or is it a run input in the D3 sense — not part

of the world's identity — digested by the consumer?

4. Naming: book.surface.v1 vs opensentience.surface.v1; and whether MAILBOX here may share a name with the

unwritable core role.

Appendix — the census script (run it; the table above is its output)

import json,glob,collections
stencils=collections.Counter(); ops=collections.Counter(); states=collections.Counter(); objkinds=collections.Counter(); anchors=collections.Counter(); steps=takeaways=0
for f in sorted(glob.glob('opensentience.org/_patterns/scenes/*.json')):
    s=json.load(open(f)); stencils[s['stencil']]+=1; p=s.get('params',{})
    loci=p.get('loci',[]); objkinds['LOCUS']+= loci if isinstance(loci,int) else len(loci)
    for k,key in (('RING','rings'),('CARRIER','carriers')):
        if key in p: objkinds[k]+=len(p[key])
    if 'slots' in p: objkinds['SLOT']+=p['slots']
    if s['stencil']=='join': objkinds['STATION']+=1
    if s['stencil']=='wire': objkinds['STATION']+=3; objkinds['WIRE']+=1
    if s['stencil']=='progress': objkinds['METER']+=4
    if s['stencil']=='migrate' and p.get('mailbox'): objkinds['MAILBOX']+=1
    for st in s['steps']:
        steps+=1; takeaways+=bool(st.get('takeaway'))
        for a in st.get('actions',[]):
            ops[a['op']]+=1
            if a['op']=='state': states[a['state']]+=1
            if a['op']=='spawn': objkinds['LOCUS']+=1
            if a['op'] in ('move','place'): anchors[str(a.get('to',a.get('at')))]+=1
print(dict(stencils), dict(ops), dict(states), dict(objkinds), len(anchors), steps, takeaways)