DOM
Semantic, indexable, responsive
Real HTML as a canvas texture—still selectable, searchable, accessible, and interactive. Explore the API through original, progressive-enhancement demos.
const canvas = document.querySelector("canvas");
const ctx = canvas.getContext("2d");
Checking browser capability…
Canvas gives us low-level visual control. The DOM gives us language, forms, accessibility, selection, and browser-native behavior. The experimental API creates a bridge: draw an element, then tell the browser where its interactive surface landed.
Semantic, indexable, responsive
Programmable, spatial, GPU-ready
One surface, both capability sets
Each study isolates one idea: native rendering, spatial transformation, and graceful fallback.
A real form is painted into Canvas 2D. In a supporting browser, the input remains focusable and editable. Everywhere else, the same interface stays usable as ordinary HTML.
Move through a field of live pixels. This compatible study previews the kind of optical treatment a DOM texture can enter.
Tilt a semantic card stack in 3D. It models the transform handoff needed to keep hit targets aligned with rendered pixels.
The core loop is compact. The architectural consequences are the interesting part.
1<canvas id="scene" layoutsubtree>
2 <form id="controls">
3 <label>Intensity</label>
4 <input type="range" />
5 </form>
6</canvas>
layoutsubtree tells the browser to lay out
descendants of the canvas and expose them to accessibility
systems.
Select it, copy it, translate it, and find it in-page.
Forms retain focus, editing behavior, and browser integrations.
Send the result through 2D Canvas, WebGL, or WebGPU.
Assistive tech, crawlers, and agents can inspect the DOM.