Every interface convention that feels obvious today — a button you press, a list you scroll, a modal that blocks everything behind it — was invented to solve a problem specific to flat rectangular screens. Spatial computing removes the rectangle, and a surprising number of those conventions simply stop making sense once there's no frame to define what's in the interface and what isn't.
Depth is not the same as stacking order
On a flat screen, layering is a stacking order — one thing sits in front of another. In a spatial interface, depth is physical distance, and that has real consequences: an element placed too close feels aggressive and uncomfortable to look at for long, while something too far away becomes hard to read or select accurately. Good spatial design treats depth as a deliberate design decision with real ergonomic limits, not as a free visual trick.
Gaze and gesture need forgiveness built in
Pointer-based interfaces can rely on pixel-precise clicks. Gaze and hand-tracking can't — human eyes drift, hands aren't perfectly steady, and the sensors themselves have margins of error. Interfaces that ignore this and place small, densely packed targets end up frustrating in a way that a mouse-driven UI never would, because every miss feels like the system failed you rather than the design failed the input method.
The designers doing the best work in this space right now are the ones borrowing least from flat-screen habits — building around confirmation gestures, generous hit areas, and audio or haptic feedback that replaces the visual affordances a screen used to provide for free.
