Fitts's law
If you pack controls tight or shrink hit targets to win a few pixels back, you are not just trading aesthetics for ergonomics. You are asking people to spend more time and attention on each move, and you will see it in errors as well as stopwatch time.
The time to acquire a target increases with distance to the target and decreases with the target’s width along the axis of motion.
Paul Fitts framed that relationship quantitatively in 1954 work on rapid aimed movement. Later HCI practice (including ISO 9241 work on pointing tasks) often writes movement time as a constant plus a slope times an index of difficulty built from distance D and target width W along the approach axis, with a logarithmic term so halving W or doubling D hurts more when the control was already small or far away.
In dense layouts, icons in a tight grid shrink W and raise D between neighbors, so the same screen real estate buys slower, fussier aiming. Spreading things out without growing targets can ease crowding errors but still raises travel time, so the fix is not always “more whitespace” alone.
On touch, fingers are blunt pointers. Effective W is whatever the user can reliably land inside, not the box your CSS drew. Below a certain size, people start aiming twice or missing and correcting, which is another way of saying the law won.
Palettes, tabs, and inline actions compete for the same motor budget as the canvas. If the most common next click is far from where the hand already is, you pay that distance on every repetition. Edges and corners can behave like larger targets in one dimension because the pointer cannot overshoot past the bezel, which is why classic menu bars and full-screen hot corners feel easier than the same pixel width floating mid-field.
Hick’s law names the decision-time side when the choices are cognitive peers. Fitts’s law is a model for pointing. It still gives you a shared vocabulary when someone asks for “just one more icon in the toolbar” or a row of 16-pixel affordances on mobile. Bigger along the motion axis, closer to the last focus of attention, or both. Everything else is negotiation.