The Design of Everyday Things

Norman's move is to relocate blame. When a door is pushed when it should be pulled, the ordinary response is embarrassment; his response is that the door failed to communicate. The book is a systematic case that usability problems are design problems with identifiable causes.

The reframing is not merely generous. It is productive: a user's confusion becomes evidence about the artefact, which makes it something that can be investigated and fixed rather than tolerated.

If people keep getting it wrong, the object is telling you something about itself, not about them.

Devmark’s reading — not a quotation from the book

An affordance is a relationship between an object's properties and the capabilities of whoever is using it — what it makes possible. A signifier is the perceivable signal that communicates where the action should happen.

Norman added the second term in later editions precisely because the first had been widely misused in design writing to mean "a visual hint". The distinction now does real work: a flat panel affords pushing; a plate positioned where the hand goes signifies where.

Three further principles carry most of the practical weight. Natural mapping arranges controls so their spatial relationship matches what they control — the reason a badly mapped hob produces a lifetime of wrong burners. Feedback confirms that an action registered, and must be immediate and informative.

Constraints — physical, cultural, semantic and logical — narrow the space of possible actions so that the wrong one is difficult or impossible. Norman's preference is always for making an error hard rather than warning against it.

Norman models interaction as two gaps. The gulf of execution is the distance between what you want to do and the means the system offers. The gulf of evaluation is the distance between what the system did and your ability to tell what it did.

Most usability failures live in one of the two, and naming which one is failing is often enough to suggest the fix. A confusing interface with clear feedback is a different problem from a clear interface with none.

People build a mental model of how a thing works from whatever the thing shows them, and then act on it — which is why a thermostat is so often turned up high in the belief it will heat faster.

Norman's point is that the designer cannot install a model directly. The only channel is the system image: what the artefact, its behaviour and its documentation actually convey. When the system image is incoherent, users will build a model anyway, and it will be wrong.

You cannot give someone a mental model. You can only build an object from which they will construct one.

Devmark’s reading — not a quotation from the book

The chapters on error are the book's most consequential contribution outside design. Norman separates slips — the right intention, wrongly executed — from mistakes, where the intention itself was wrong, and argues they need different remedies.

His broader argument is that "human error" is usually the last link in a chain of design decisions, and that treating it as an explanation stops the investigation exactly where it should begin. This thinking runs directly into modern safety-critical practice.

Norman distinguishes knowledge a person must carry from knowledge the environment supplies. Precise behaviour does not require precise knowledge if enough of it is available in the world at the moment of action.

The design consequence is a preference for recognition over recall — for making the necessary information present rather than requiring it to be remembered. Almost every good interface decision of the last forty years is a variation on this.

Later editions add a process argument: solve the right problem before solving the problem right. Norman's version separates a divergent phase that widens the problem space from a convergent one that narrows it, then repeats the pair for the solution.

He is candid that the first diamond is the one organisations skip, and that most design failures are failures to question the brief rather than failures of execution.

The examples are physical and dated — doors, taps, telephones, slide projectors — and readers wanting screen-interface guidance will have to do the translation themselves. The later editions add digital material but the book's intuitions remain object-shaped.

It also says comparatively little about the contested parts of modern product work: designing against a user's interest, attention capture, and the ethics of persuasive systems. Norman has written on these elsewhere; the book largely assumes the designer is trying to help.

It is the book that gave the field its vocabulary. Affordance, signifier, mapping, feedback, conceptual model, forcing function, slip and mistake are all in common use, and most practitioners use them without knowing where they came from.

Its deeper contribution is a stance rather than a method: that confusion is data, that blame is a dead end, and that the object is answerable to the person rather than the reverse.