# Read the Web: HTML, CSS, the DOM & Accessibility

**Track:** Foundations of AI-Native Development — AI-Native Full-Stack Development — complete (30)
**Framework / surface:** web dev (web platform)
**Level:** Beginner
**Prerequisites:** Verification-Driven Development

**In one line:** Learn the browser’s durable substrate well enough to inspect and correct generated interfaces.

## Theory, aesthetics & inspiration

Frameworks come and go, but every one of them compiles down to the substrate Tim Berners-Lee proposed at CERN in 1989 and shipped as HTML: a document of elements, styled by CSS, represented at runtime as the DOM tree, reacting to events. AI-native builders do not need every property memorized; they need enough web literacy to inspect what generated code ultimately becomes. Semantic HTML is the load-bearing part — a button should be a `button`, form fields should carry labels — because native elements bring keyboard behavior, accessibility semantics, and predictable tooling with them, which is what the W3C's WCAG guidelines and the browser's accessibility tree measure. Coding agents make this literacy more important, not less: a generated interface can be visually plausible while being unnavigable by keyboard and incoherent to assistive technology. Open the inspector; read the tree; know when the agent has produced the wrong thing under a right-looking surface.

**Builder question:** If the visual styling disappeared, would the underlying document still communicate its structure and controls correctly?