Depth, parallax and occlusion are becoming primitives. What happens to a component library when the canvas gains a third axis.
PUBLISHED JUL 06, 2026 • 7 MIN READ
A button has had two states and two dimensions for thirty years. Give the canvas a z-axis and almost every assumption baked into component libraries needs re-examining — starting with the idea that stacking order is a styling concern rather than a spatial one.
In flat design, importance is communicated by size and colour. In spatial layouts, proximity to the viewer does the same job more directly and with less visual noise. A modal does not need a scrim if it is genuinely closer to the user than everything behind it.
The third axis is not decoration. It is the cheapest hierarchy signal you will ever add to an interface.
Clipping content at a container edge is a two-dimensional compromise. When surfaces have real depth, content can pass behind another plane and remain understood as continuous — which is why spatial lists feel less abrupt than their scrolling equivalents.
None of this is free. Depth cues that read as elegant to one user read as motion sickness to another, and screen readers have no concept of a z-axis. Every spatial affordance needs a flat fallback that carries the same meaning, honoured through prefers-reduced-motion.