7a. Paper design v1 — the document under walkthrough

kerfmaster pin list · UI / workflow decisions

rulingA decision by Aristide or Jordan. True because it was decided; it can be superseded, but it cannot be stale.

Content last changed 2026-08-17 — computed from the item itself, not typed.

This contract states a decision, not an implementation.

takes —

makes —

fails if Any interface code written before both partners sign off on the paper; any stage in the app that skips a gate the paper lists.

Contract last changed 2026-08-17 — computed, not typed. Dated separately from the text above, so neither date can speak for the other.

Commissioned by Jordan (definition phase closed enough, 2026-08-17), started on Aristide's go-ahead the same day. Source of truth: design/paper-design-2026-08-17.md, rendered to /design/ by render_research.py (data/generator split). This pin points at the document; the document, once signed off, is the spec the interface is coded from — it is the reverse-engineering step pin 6b promised, done as one readable page.

What the paper fixes

Spine: Intake → Heal → QC → Part review → Nest → Cut plan → Post → Deliver+Archive; per stage: what runs automatically, what the operator may hand-edit, which gates block advance, what the audit trail records.
Aristide's order ruling folded in: tabs/leads/start points are chosen AFTER the nest, in the sheet frame (paper sec 7).
Roles: owner/operator full; sales scope proposed, pending Q1. Shop floor = file recipients, not users.
Tables as owner screens with provenance (measured/decided/assumed); multi-machine profiles carrying the setup/verify variable.
Not in v1: estimator, engrave layers, rapid routing. (Bridges WERE on this list until Jordan’s Q2 answer pulled them in.)

The four questions — answered by Jordan same day (Aristide present)

Q1 sales is for QUOTING — gets the nest to check quoted material usage; quote nest is distinct from the production nest, approval/post gates unchanged. Q2 “automate it, flag it, make it easy for operator to double check and edit” — automatic bridges move INTO v1 (supersedes ‘bridges later’; a scope change, recorded as such). Q3 multi-workstation shop, shared workstations, macOS-style switching — three options written into the paper (§4); DECIDED: Option C (Aristide, 2026-08-17) — PIN to switch, password/biometric on owner-only actions. Q4 n+1 strategy: app proposes next free O-number per machine scheme, operator override always, audited.

Next step

Paper walkthrough with -61 nest, novi sign, 33215, a dirty .ai, plus the eight edge cases listed in §8 — then sign-off, then code.

Paper walkthrough — run 2026-08-17 (report at /design/walkthrough/)

Three jobs walked screen-by-screen with numbers measured from the files that day; .ai leg DEFERRED by Aristide until that leg is ready. The spine held — no stage missing, none out of order; tabbing-last passed its hardest real test (the -61 traffic rule: tab choice and cut order are coupled, decidable only in the sheet frame). Five findings, fixes folded into the paper same commit:

W1 a customer file can BE a finished nest (-61 was): intake classifies part / parts / finished nest; cut-as-drawn makes the nest screen verify-only.
W2 an incomplete material row (14 ga, no beam comp) warns at INTAKE, not first at post.
W3 cutter comp is DECLARED at cut plan (G41 norm, G40 for art/text), never inferred — the novi sign needs it.
W4 release screen lists every ASSUMED value the post consumed for owner confirm — live case: 1/2" beam comp .020 read from 33215.NC, still unconfirmed.
W5 machine swap invalidation rule: heal/QC survive, nest re-validates, cut plan + post re-run, diff shown.

Also measured on the way: -61's header lies twice (claims unitless/English + 123.6×95.8" extents; geometry is mm, 72.2×40.2") — gates judge measured geometry, never headers.

Findings approved (Jordan, 2026-08-17): W1/W2/W3/W5 “Yes”; W4: “We will have to field verify the numbers, but if you pull them from the table as discussed we can easily just change it in the table” — assumed values ride from the table, release screen shows them, field verification lands as a table edit (provenance → measured). Remaining before code: formal sign-off from both partners; .ai walkthrough when that leg is ready.

SIGN-OFF (2026-08-17): Jordan approved all five walkthrough findings; Aristide: "yes." The paper design is the agreed spec for v1, minus the .ai leg (own walkthrough when ready). Pin type flips intent → ruling: the document is now a decision, not a proposal. Next: interface design options (in discussion), then code.