Corners are bounded in two dimensions and have the lowest cost
Aliases: corner target · two-dimensional constraint
What it is
A screen corner is bounded by two display edges. For a mouse pointer moving diagonally from inside the screen, overshoot is stopped on both axes, making the corner easier to acquire than a one-edge target. It is often described as a two-dimensional infinite target.
Why it happens
An edge removes stopping error on one axis while overshoot remains possible along the edge. A corner removes both. A user can make a coarse rapid movement toward it and let the boundaries complete the convergence. This reduces the cost of locating the corner target, not the cost of discriminating several functions placed near it.
Studying it
Compare corner, edge, and interior targets at matched distances, recording time and misses at several approach angles. Hit areas must actually touch one or two screen boundaries; an icon drawn in a corner with an inset hit box does not manipulate the two-dimensional constraint. Check whether the operating system intercepts the corner gesture.
Where it stops holding
The gain is strongest for mouse or trackpad cursor movement. A finger is not physically stopped by a touch boundary in the same way, and full-screen corners can compete with system gestures. Multiple functions sharing one corner create semantic and accidental-activation conflicts; speed cannot justify unpredictability.
Applying it
- On desktop, reserve a stable corner for one very frequent, unambiguous action, with a hot zone reaching both real display edges.
- Avoid putting a fine-grained cluster in the same corner or changing its command with context.
- Test mouse, trackpad, and touch variants separately, and verify that system corner behaviour does not take the application action.