Interaction Design Principles

UI/UX Design Principles

Chapter 4 · Interaction Design Principles

Chapter 1 mentioned affordances and feedback only in passing, as Don Norman's own real contribution to the field. This chapter goes deep on both, plus two genuinely useful, real predictive laws (Fitts's and Hick's) and the cognitive-science concept underneath all of it: cognitive load.

Affordances: A Real, Two-Step Correction

"Affordance" wasn't Norman's own term originally. Psychologist James J. Gibson introduced it in his 1966 book, developing it further in his 1979 work The Ecological Approach to Visual Perception — an affordance, to Gibson, is what an object objectively offers or provides, whether or not anyone notices it. A chair affords sitting whether a person has spotted it or not.

Norman's 1988 Adaptation — Narrower, on Purpose

In The Design of Everyday Things (1988), Norman adapted Gibson's concept for design, deliberately narrowing it: a perceived affordance is an action possibility a person can actually, visibly recognize — not just an objective fact about the object. This shift made the concept genuinely useful for designers (you can design for what's perceivable), at the cost of no longer matching Gibson's own original, broader meaning.

A Real, Honest Self-Correction

Norman later acknowledged that this narrowing was unintended, and introduced a separate term, "signifier," specifically for the visual cue that communicates a possible action (a button's own drop shadow, an underline on a link) — distinct from the affordance itself. A flat, un-styled clickable element can still technically afford clicking; without a real signifier, most users will never discover that it does.

Feedback

Where a signifier tells a user what's possible before an action, feedback confirms what actually happened after one — a button's own visible press state, a loading spinner, a confirmation message. Missing feedback is one of the most common real sources of user uncertainty: did that click actually register?

Fitts's Law: Predicting How Long a Click Takes

Psychologist Paul Fitts published a real, formal model in 1954 for how long it takes a person to point at a target, based on the target's distance and size:

ID = log₂(2D / W)          (Index of Difficulty)
MT = a + b·ID              (Movement Time)

D = distance to the target's center
W = the target's own width/tolerance

A larger, closer target is measurably faster to reach than a small, distant one. This real formula has direct, verifiable UI consequences: bigger buttons for frequent actions, related controls placed close together, and the real, named "infinite edges" effect — an element pinned to a screen edge has effectively infinite size in that one direction, since the cursor can't overshoot past the edge. "Magic corners" (where two edges meet) are effectively infinite in both directions — the real reason macOS anchors its menu bar to the very top of the screen, and why older Windows versions placed the Start button in a screen corner.

Hick's Law: Predicting Decision Time

William Hick (1952) and Ray Hyman found that decision time increases with the number of choices — but logarithmically, not linearly, so doubling the choices doesn't double the time:

T = b · log₂(n + 1)

T = decision time
n = number of choices

A Real Caveat — Don't Apply It Blindly

Hick's Law genuinely holds for unfamiliar or randomly-ordered choices. But real research found the relationship changes with familiarity and structure: searching a randomly ordered list scales linearly, not logarithmically, while an alphabetically ordered menu lets a user who already knows the item's name find it in genuinely logarithmic time. The same underlying "don't misapply a cognitive-science number without checking the real conditions it was measured under" lesson from Chapter 3's own Miller correction applies here too.

Cognitive Load: The Resource Every Design Decision Spends

Educational psychologist John Sweller developed Cognitive Load Theory in the late 1980s, describing the real, limited mental effort available in working memory. It splits into three real, distinct types:

Intrinsic Load

The task's own inherent difficulty — fixed, not something a designer controls.

Extraneous Load

Unnecessary burden created by how information is presented — genuinely controlled by the designer.

Germane Load

Effort spent actually building understanding — the "good" load, freed up when extraneous load is low.

The Real Job of UI Design

A designer can't reduce a task's own inherent difficulty (intrinsic load) — booking a genuinely complex multi-leg flight is inherently harder than checking a single balance. What a designer can control is extraneous load: confusing layout, unclear labels, inconsistent patterns, all real, avoidable tax on a user's own limited mental effort. Reducing extraneous load is, in a real sense, most of what good interaction design actually is.

Three Real Principles, Compared

Principle	What it predicts	Direct UI takeaway
Fitts's Law	Time to reach and click a target	Bigger, closer targets for frequent actions; use screen edges/corners
Hick's Law	Time to decide among choices	Fewer, well-organized choices — but check familiarity/ordering first
Cognitive Load Theory	Total mental effort a task consumes	Minimize extraneous load — the only one of the three a designer controls

Hands-On Exercises

Exercise 1

A teammate says "affordance and signifier mean the same thing — Norman uses them interchangeably." Explain, in your own words, why this is inaccurate, using the real distinction the chapter describes.

📄 View solution

Exercise 2

A designer wants to place a "Save" button that will be clicked constantly throughout a long editing session. Using Fitts's Law, explain, in your own words, two real, concrete placement/sizing decisions that would make this button genuinely faster to use.

📄 View solution

Exercise 3

A team cuts a genuinely necessary step out of a multi-step tax-filing flow, arguing it reduces the user's cognitive load. Explain, in your own words, which specific type of cognitive load this actually would (or wouldn't) reduce, and why removing a genuinely necessary step is a different move than reducing extraneous load.

📄 View solution

Chapter 4 Quick Reference

Affordance — coined by James Gibson (1966/1979); Norman narrowed it to "perceived" in 1988, then introduced "signifier" separately once he recognized the narrowing was unintended
Fitts's Law (1954): MT = a + b·log₂(2D/W) — bigger, closer targets are faster; screen edges/corners are effectively infinite-sized
Hick's Law (1952): T = b·log₂(n+1) — fewer choices decide faster, logarithmically, but familiarity/ordering changes the real relationship
Cognitive Load Theory (Sweller) — intrinsic (fixed), extraneous (designer-controlled), germane (useful) load; good UI design minimizes extraneous load specifically