Programmatic logo generation

The idea

Instead of relying on AI image generation (which introduces randomness), the CW graph letterform logo could be generated programmatically. The logo IS a graph — nodes at computed positions, edges as lines. Code gives pixel-perfect control.

Possible approaches

• SVG math — compute node positions along a C-curve and W-zigzag, output SVG directly. Simplest, most portable.
• D3.js — force-directed or manually positioned graph, export as SVG
• Three.js / WebGL — for 3D/isometric variations, export as PNG or render live on the docs site
• p5.js / Canvas — quick prototyping of variations
• Figma plugin — generate variations inside Figma for final cleanup

Why this is interesting

• Every variation in LOGO_PROMPTS.md (proportion, weight, alignment, connection style, color) becomes a parameter
• Could generate a grid of 100 variations instantly by sweeping parameters
• The logo could be a live interactive element on the docs homepage (WebGL)
• SVG output means true vector at any size — no AI upscaling needed
• Version-controllable — the logo IS code

When to explore

Low priority — AI generation with the sketch-to-digital prompts is faster for now. Revisit if:

• AI outputs aren’t precise enough
• We want an animated/interactive logo on the site
• We need hundreds of size/color/weight variations for a brand kit