⚖️ Canon audit
The canon was contradicting itself
Updated 2026-07-12 18:00 WEDT · · Designee
I audited all 11,000 lines of the lexicon. 9 real contradictions — the parked note said 5.
Four would have made a session build the wrong thing. I resolved 5 by the later-date rule.
Your 17:50 economy dump resolved two of them (tessera = a completed token). Two still need you: credit pricing + draw odds. Builds held on those.
⚠️ Needs your ruling
1. Credits — what does money actually buy?
Jul 10 €5 → 3 credits
(~€1.66 each)
Jul 11 "canon" €5/mo → 8 credits
(~€0.63 each)
Jul 12 $5 → 21 credits
(~$0.24 each)
The value of a credit swings 7× across three sections, each written as settled.
And a new one: Jul 12 says "free account = 1 credit = full participation" — but Jul 11 says
basic membership is the €5. Is there a free tier at all?
My rec: answer the A/B ladder on the credit-pricing page first, then I'll
make everything else conform to it. Nobody is building a pricing surface until you do.
2. "Chip" — ✅ your 17:50 dump answered this
You settled the object model, so the dead word resolves itself:
tile (bespoke atom) → coin (a template, not a collectible) → token (coin + tiles
on its key = the currency) → tessera (a completed token = the real collectible).
So: the unit is a TILE → "shadow chip" = shadow tile. The
MushMe pledge chip is staked and spent → that's a coin. Two objects, one dead word —
as I guessed, now confirmed by your own model. One nod and I unblock the mush_chip_ledger
rename.
3. Draw odds — your 3× cap vs the φ-boost
Jul 10 (locked) odds ∝ √tokens · HARD CAP 3× · 30% of slots reserved for newcomers
Jul 12 odds ×φ per rejection (1 → 1.62 → 2.62 → …) → "near-guaranteed spot"
Three rejections puts you at φ³ = 4.24× — straight through your 3× cap. And
"near-guaranteed" eats the newcomer reserve, which was the thing protecting first-timers.
I took my first answer to the council and they took it apart. I'd said "the 3×
cap is global, φ works inside it." They called that a patch, and they're right: Rule A measures
merit; the φ-boost compensates for bad luck. Capping a bad-luck compensator with a merit ceiling
is a category error — and it makes the promise ("keep trying and you'll get in") mathematically
false, because the cap silently prevents exactly what φ was added to deliver. It creates a
ceiling the loyal person never sees.
New rec (5 of 9 voices converged on this independently): stop using a
multiplier for persistence. Three axes, one mechanism each —
• Merit: odds ∝ √tokens, cap 3×. Unchanged.
• Newcomers: 30% of seats, drawn first, first-timers only. Untouchable.
• Persistence: a guaranteed seat after 3 rejections, served from its own seat
(≈1 per event) — not from the merit pool. Counter resets when you attend. No φ.
It's honest ("you're 2 rejections from a guaranteed seat" is true and legible; "your odds are
2.618×" is neither), it can't eat the newcomer reserve, and there are no competing multipliers
left to reconcile.
Nice coherence: for Friday's 6-person den, a 30% reserve = 2 seats — exactly the
2 newie slots already built. The reserve and the newie slots are the same mechanism.
And φ keeps its home: the geometry and the coin ladder (0.62 → 1 → 1.62 → 2.62 → 4.24),
where it's structural. In a fairness mechanism it becomes a promise we can't keep.
4. Are the five core EEs free, or do they cost credits?
Jul 11 "the EEE family is FREE and always-energized the instant you sign up — no credit needed"
Jul 12 "each credit activates a realm/tool/companion" + "the 5 that come with basic membership"
My rec: free. If a new person's first act is spending a scarce credit to wake
the guide who's supposed to welcome them, the welcome costs something — which is the wrong first
impression. Credits gate realms, not the helpers who greet you.
✅ Already resolved (no action needed)
⭐ Tessera = a COMPLETED token. Your 17:50 dump finally answered the question
we'd been circling for weeks — what the word actually means. An access token is an
incomplete token; fill every edge and it becomes a tessera. That one definition collapsed two of
the four conflicts on this page. And your "flip" is a phason flip — we already have the engine.
I told the Codys to wire the existing one rather than build a new mechanic.
Tiles trade — your dump superseded my ruling, and yours is better. I'd said
"same-type swap only, no value transfer." You explained the real reason: every tile is bespoke,
so holding someone else's tile of the right type still proves you made one yourself — and that you
interfaced with a person. The trade engine is upgrading: a layer-2 token makes its layer-1
prerequisite surplus, so you trade the outgrown one away. You release what you've outgrown.
designeee.app is not a deploy rung — design workspace only (it already leaked
science labs once). Rungs stay lab → test → social.
The role-ladder table was stale — still listed rung 2 as "Tui" (it's Oldie), and
Team actually comes before full Steward. Sessions were about to build from it.
"Wide/narrow" now mean corners, not tiles — but the old role-glyph spec still
says "3 wide tiles at a vertex." A builder following that literally would place the wrong rhombi.
Flagged before anyone did.
Penrose 6/7 has no coherent mapping (1–4 are corners; a vertex is 10) — so I've
told everyone not to hard-code those constants until you settle it.