10-foot UI

The name is a measurement. A phone is held at one foot, a laptop at two, and a television is watched from about ten, which means a 55 inch screen across a living room subtends roughly the same angle as a laptop on a desk. Everything follows from that. Type that would be generous on a desktop is unreadable at ten feet, so body text starts where desktop headings stop; a card that would be a thumbnail becomes the size of a playing card held at arm’s length; and the number of things that can be on screen at once collapses, which is why a TV home screen is a handful of horizontal rails rather than a grid of forty items.

The input changes more than the distance does. There is no pointer: there is a directional pad with four arrows and a select button, and possibly a voice key. Every navigable element must therefore be reachable by up, down, left, and right from wherever the highlight currently sits, which turns layout into a graph problem. The highlight is not a decoration but the entire location system, so it is drawn far heavier than a focus ring on the web, usually as a scale, a glow, and a colour change at once, and it is never allowed to be absent: a screen that arrives with nothing highlighted has stranded the viewer. Because focus and selection travel together here, moving the highlight is expected to preview or load, which is a case of selection follows focus that would be an antipattern in a desktop list. Hover does not exist, so nothing may hide behind it, and every state a design wants to show has to be reachable by arrows alone.

The horizontal rail is the register’s signature layout, and it is a carousel with one crucial difference: it is scrolled by the focus rather than by a gesture or a pair of arrow buttons. Pressing right moves the highlight to the next card and the rail slides only when the highlight would otherwise leave the screen, which is why a rail can be forty items long without a scrollbar, a pagination control, or any indication of its length. The cost is depth. Getting to the fortieth item is thirty-nine presses, so real systems lean on rows as categories, on a “continue watching” rail pinned first, and on search by voice for anything the viewer already knows the name of.

The last constraint is physical. Consumer televisions still overscan, cropping a few percent off every edge of the picture, and a panel that is exactly the reported resolution may still lose its outermost rows to a bezel or to display scaling. The remedy is the title-safe margin, a border of five percent or so that carries no text and no interactive element, which is why TV interfaces look as though they float in the middle of the screen. Two further living-room facts complete the register: viewers sit in a dark room, so backgrounds run dark and pure white is avoided as glare; and remote input arrives with real latency over infrared or Bluetooth, so the highlight must answer instantly even when the content behind it has not loaded, or the viewer presses right three more times and overshoots by three.