Skip to content

docs(skill): parameters must be dimensions, never absolute coordinates#71

Open
YLZha wants to merge 1 commit into
mainfrom
skill-rule-no-coordinate-params
Open

docs(skill): parameters must be dimensions, never absolute coordinates#71
YLZha wants to merge 1 commit into
mainfrom
skill-rule-no-coordinate-params

Conversation

@YLZha

@YLZha YLZha commented Jun 4, 2026

Copy link
Copy Markdown
Collaborator

What

Adds a Parameter Planning principle to the woodworking skill: a user parameter must be a dimension (length, width, thickness, depth, gap, radius, angle, count) — never an absolute coordinate (top_z, bottom_z, apron_top_z = 24.9 in).

Why

A coordinate parameter is brittle in two ways:

  1. It doesn't survive moving the model (it's tied to the world origin).
  2. It doesn't track other changes — bump table_h and a hardcoded sp_botz = 24.9 in stays put while everything around it rises, silently distorting or breaking the joint.

Positions should be derived in the build from dimensions and references (the spandrel bottom is apron_bottom - spandrel_depth, not an absolute sp_botz); the only thing measured from the world origin is the root (legs on the floor, per rule 10). The rule includes a litmus test: "if the customer moved this table across the room, or made it 2 inches taller, would this value still be correct as written?"

Prompted by a Ming-table build where literal sp_botz/slot_botz = 24.9 in distorted the spandrel and leg-slot joints when the table height changed.

🤖 Generated with Claude Code

Add a Parameter Planning principle: every user parameter must be a
measurable quantity (length, width, thickness, depth, gap, radius, angle,
count) — never an absolute position like top_z/bottom_z. Coordinate
parameters don't survive moving the model and silently break when a
related dimension (e.g. table_h) changes while the literal stays put.
Positions are derived in the build from dimensions + references; the only
world-origin reference is the root (rule 10). Includes a litmus test.

Prompted by a Ming-table build where literal sp_botz/slot_botz = 24.9 in
distorted the spandrel/leg-slot joints on a height change.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant