The Human: Perception, Cognition and Memory

You can't design a good interface without knowing who's on the other side of it — and humans come with fixed, measurable limits. Our eyes miss things, our attention is a spotlight, our short-term memory is tiny, and our hands take a predictable amount of time to reach a target. The wonderful news for designers is that these limits are so reliable you can put numbers on them and design around them. This page is about the human as a system with known specs.

Perception, cognition, memory

Perception — how we take the world in (mostly sight, for screens). We notice contrast, motion and grouping before anything else, and we're easily fooled: we skim rather than read, miss things we aren't looking for ("inattentional blindness"), and about 1 in 12 men can't rely on red-vs-green. Design implication: make important things perceptually pop, and never carry meaning by colour alone.
Cognition — how we think, decide and process. Thinking is effortful and slow compared with recognising, so we lean on habits and shortcuts. Every extra decision or step costs mental effort ("cognitive load"). Design implication: reduce choices, chunk information, and prefer recognition over recall.
Memory — short-term (working) memory holds only a handful of items for seconds; long-term memory is vast but slow to write. The famous figure is "7 ± 2" chunks (Miller) — which is why phone numbers are grouped. Design implication: don't ask people to hold things in their head; show it on screen.

Two of these limits are so precise they became laws with equations — and both are HCI classics you should know.

Fitts's law — the time to hit a target

In 1954 Paul Fitts measured how long it takes to move a pointer (a finger, a mouse) to a target. The result: the time depends on how far you have to move (distance $D$) and how big the target is (width $W$).

MT = a + b \cdot \log_2\!\left(\frac{2D}{W}\right)

In words: bigger and closer targets are faster to hit; smaller and farther ones are slower — and the relationship is logarithmic, so doubling the distance adds only a fixed amount of time, not double. Drag the slider to shrink the target and watch every point on the curve rise:

This is why good interfaces put frequent buttons big and near where your cursor already is, why the corners and edges of a screen are prime real estate (you can't overshoot an edge, so it's effectively infinitely wide), and why a tiny "×" close button in a far corner is a usability sin.

The four corners and edges of a screen are the fastest targets of all, even though they're small, because you can't overshoot them — the cursor slams into the edge and stops. In Fitts's terms the effective width $W$ becomes enormous, so the movement time drops. That's why the classic menu bar sits flush at the very top edge, and why a corner "hot spot" is such a powerful place to put an important control.

Hick's law — the time to choose

Fitts covers moving; Hick's law covers deciding. The more options you present, the longer it takes someone to choose — again logarithmically:

RT = a + b \cdot \log_2\!\left(n + 1\right)

where $n$ is the number of equally-likely choices. Adding options slows people down, but with diminishing effect — going from 2 to 4 choices hurts more than going from 20 to 22. Drag the slider to change how quickly a person processes each option:

The design lesson: fewer, well-organised choices are faster. A menu of 40 flat items is slow and daunting; grouping them into 5 categories of 8 lets the user make two quick decisions instead of one huge one. This is why menus nest, why "less is more," and why a checkout with one clear button beats one with ten.

Two traps with these laws:

Both laws are logarithmic, not linear. Movement/choice time does not double when distance or the number of options doubles — it grows by a smaller, shrinking amount. Saying "twice as many menu items = twice as slow" is wrong.
Don't blindly minimise choices. Hick's law says fewer options are faster to choose from, but hiding everything behind endless sub-menus adds steps and harms discoverability. The goal is good organisation (chunking), not just a smaller number. And Hick's law assumes options are equally likely and the user must consider each — it doesn't hold when one option obviously stands out.

Putting the specs to work

Add it all together and the human "spec sheet" writes your design rules for you: keep working memory light (show, don't make them remember), make targets big and reachable (Fitts), keep choices few and grouped (Hick), make the important things perceptually obvious, and never depend on a single sense or colour. Design for the human you actually have, not an imaginary one with perfect eyes, infinite attention and unlimited memory.