Touch target

A touch target is a region, and the mark inside it is a separate thing. That separation is the whole word. A sixteen pixel glyph centred in a forty-four pixel button is a forty-four pixel target: press anywhere in the padding, well clear of the drawing, and the control fires. Take the padding away and the same glyph looks identical in a screenshot while the region that answers has shrunk to the size of the ink. Every argument about targets is really an argument about which of those two boxes someone measured.

The numbers are worth carrying, and they do not agree because they answer different questions. Apple asks for at least forty-four by forty-four points, Android for forty-eight by forty-eight density independent pixels, and WCAG 2.2 sets twenty-four by twenty-four CSS pixels as its AA floor, with exemptions for inline links and for small targets that have enough clear space around them. The floor is a floor. Anything used one handed, in motion, or repeatedly deserves the platform figure rather than the minimum that passes. The size question has an entry of its own on this site, under touch target size; this one is about the region the size is a property of.

Spacing belongs to the same idea, because the failure people actually hit is the neighbouring control rather than empty space. Two targets that touch each other are worse than two slightly smaller ones with a gap, and WCAG says so explicitly by letting spacing substitute for size. When targets are packed into a row, the honest fix is fewer of them, not a tighter grid.

Growing the region without growing the paint is the everyday craft. Padding on the control itself is the cleanest version, since it keeps one box for layout and hit testing. A pseudo-element stretched beyond the visible bounds works where the layout cannot spare the room, as long as nothing overlaps the extension. Making the whole row carry the press on behalf of the small icon inside it is often the best answer of all. And the word is not only about fingers: a trackpad, a head switch, eye tracking, and an unsteady hand all have the same aiming problem, which is why the criterion is written about targets in general rather than about touch.