8.5 — Photography, Design and the Everyday Beautiful

Most of the visual decisions that reach people are not made in galleries. They are made about typefaces, packaging, signage and photographs, by people whose work is invisible when it succeeds.

Photography

What do the three camera settings actually do?

Everything in photography is a trade between three controls, and each one has a side effect. That is the whole subject.

Aperture is how wide the lens opening is, written as an f-number. A smaller f-number means a wider opening — f/1.8 is wide, f/16 is narrow — which is confusing until you know that the number is the focal length divided by the opening's diameter, so it is a fraction.

Its side effect is depth of field: how much of the scene in front of and behind your subject is sharp. A wide aperture gives a shallow depth of field and the blurred background that makes portraits look expensive. A narrow one keeps everything sharp, which is what a landscape needs.

Shutter speed is how long the sensor is exposed. Its side effect is motion: a fast shutter freezes a bird's wing, a slow one turns a waterfall into smoke and makes a handheld shot blurred.

ISO is how much the signal is amplified. Its side effect is noise: raising ISO lets you shoot in the dark and adds grain.

The trade is unavoidable. In dim light you must open the aperture (losing depth of field), slow the shutter (risking blur), or raise ISO (adding noise). Choosing which cost to pay is the decision a photographer is making, and every automatic mode on a camera is a rule for making it on your behalf.

Why do phone photos look good with a tiny lens?

Because most of the work is being done after the shutter closes.

A phone's sensor and lens are small, which physically limits how much light they gather and how shallow the depth of field can be. Computational photography compensates by taking several exposures and combining them.

Press the button and the phone typically has already been capturing frames. It aligns them, averages them to cancel noise, takes the well-exposed parts of dark frames and bright frames to build a single image with detail in both shadows and highlights, sharpens selectively, and applies a colour rendering tuned by the manufacturer.

Portrait mode is a depth estimate, not optics. The phone builds a map of how far away each part of the scene is — from two lenses, or a dedicated sensor, or a machine-learning model — and blurs by that map. It is why it fails on hair, glasses and the gaps in a wicker chair: the depth map has an edge problem, not a lens problem.

The consequence worth knowing is that phone photographs are interpretations rather than records. Two phones photographing the same scene produce visibly different colours, contrast and detail, because their manufacturers have different opinions about what a good photograph looks like.

What makes a photograph good?

The same things that make any picture good, plus one that is specific to the medium: the moment.

Composition rules are worth knowing as defaults rather than laws. The rule of thirds places the subject a third of the way across rather than centred, because a centred subject in a rectangle is static. Leading lines carry the eye. Framing uses something in the foreground to enclose the subject. A photograph with a clear foreground, middle ground and background reads as deep.

Light does more than composition. Photographers talk about the golden hour after sunrise and before sunset, when the sun is low, so the light is warm, comes from the side, and produces long soft shadows that show form. Midday sun is overhead, which flattens faces and produces harsh shadows under the eyes.

And the moment. Henri Cartier-Bresson's phrase, the decisive moment, means the instant at which the elements in a scene arrange themselves into a coherent image — a fraction of a second earlier or later and it is nothing. That is the thing photography can do that no other picture-making can, and it is why the great documentary photographers are judged on timing.

Which photographs changed something?

"Migrant Mother" (1936), Dorothea Lange's photograph of Florence Owens Thompson during the American Depression, which drove federal food aid to the camp within days. Thompson later said she felt exploited by it and received nothing.

"The Blue Marble" (1972), taken by the Apollo 17 crew — the first photograph of the whole illuminated Earth — and "Earthrise" (1968) from Apollo 8. Both are routinely credited with contributing to the environmental movement, on the grounds that seeing the planet as a single finite object changes how you argue about it.

"Pale Blue Dot" (1990), taken by Voyager 1 from six billion kilometres at Carl Sagan's request, in which Earth is less than a pixel.

Nick Ut's photograph from Trảng Bàng (1972) and Eddie Adams's execution photograph (1968) from Vietnam, both of which shifted American opinion about the war and both of which raise the question of what a photographer owes the person in the frame. Adams spent the rest of his life uncomfortable with his, arguing that it told the truth about one second and lied about everything around it.

Design

Why do typefaces matter?

Because reading is a physical process and letterforms measurably affect it, and because a typeface carries associations before a single word is read.

The basic split: serif typefaces have small strokes at the ends of letters, sans-serif do not. The traditional claim that serifs are easier to read in long text is weakly supported at best, and the real determinants are size, spacing, line length and contrast.

What is well established is legibility at a glance, which is why road signage typefaces are designed specifically: open letterforms, distinguishable characters, and generous spacing so a word can be read as a shape at speed and at an angle.

A typeface also carries a period. Helvetica reads as neutral corporate modernity because it was adopted by so many corporations in the 1960s and 70s, not because there is anything neutral about it. Times New Roman reads as officialdom because it was the default in word processors for two decades.

Comic Sans is worth one line. It was designed for speech bubbles in a children's software product, is used everywhere it should not be, and is genuinely useful for readers with dyslexia because its irregular letterforms are harder to confuse with one another — which is why the mockery of it is not entirely fair.

What is good design, in a sentence that is actually usable?

A thing is well designed when you can tell what it does and how to use it without being told.

Donald Norman's term for this is affordance: the visible property of an object that suggests how it is used. A flat plate on a door affords pushing. A vertical bar affords pulling. A door with a handle that must be pushed is a design failure, and it is such a common one that it is called a Norman door.

The second concept is feedback: after you act, the object must tell you something happened. A switch that clicks, a button that lights, a progress bar. Systems without feedback produce the behaviour of pressing the button repeatedly.

The third is constraint: making the wrong action physically impossible. A SIM card tray that only fits one way, a plug that cannot be inserted upside down. A constraint is better than an instruction, because it does not rely on anybody reading.

Dieter Rams's ten principles for good design are the standard longer statement, and the last one is the one people remember: good design is as little design as possible.

Why are some everyday objects so hard to improve?

Because they have already converged on a shape that solves several constraints at once, and any change makes one of them worse.

The paper clip, the safety pin and the bicycle are the standard examples. The bicycle's diamond frame arrived around 1890 and has not changed in geometry because it is a very efficient use of tubing — every member is in tension or compression along its length, with almost nothing in bending.

The reason a "better" version usually fails is that the existing one is a local optimum in a space with many constraints: cost, manufacturability, familiarity, repairability, and the fact that everybody already knows how it works. A redesign that improves one constraint by 20 per cent and breaks familiarity is a net loss.

This is worth carrying beyond objects. The most common failure of a redesign is optimising for the constraint you noticed and ignoring the three you did not.

The visual language you already read

What is the grammar of a poster or a page?

Four things, and once you can name them you can see them being used on you.

Hierarchy: what you are meant to read first, second and third, established by size, weight, colour and position. A poster with three elements of equal weight has no hierarchy and is exhausting.

Contrast: differences big enough to register. A slightly larger heading reads as a mistake; a substantially larger one reads as a heading.

Alignment: things lining up on an invisible grid, which is why professional layouts feel calm and improvised ones feel restless. Almost every printed page you have ever seen sits on a grid.

Whitespace: the empty area, which is not wasted. It groups related things by putting space around them, and it gives the eye somewhere to rest. Cheap design is usually identifiable by the absence of it, because whoever paid for the space wanted it filled.

Why do warning signs look like that?

Because the shape and colour carry the message before the symbol is read, and the system is standardised internationally.

Red is prohibition and danger, yellow is warning, blue is mandatory instruction, green is safe condition or escape. Shape reinforces it: a circle for a command, a triangle for a warning, a square or rectangle for information.

The redundancy is deliberate. Around one in twelve men has some form of colour vision deficiency, so a system relying on colour alone fails for a substantial minority — which is why the shape and the symbol repeat the message, and why traffic lights are arranged by position as covered in 1.4.

The pictograms themselves were largely developed from the 1930s onward, and the most influential system is Isotype, devised by Otto Neurath and Gerd Arntz in Vienna as a way of communicating statistics to people regardless of language and literacy. The airport and public-signage symbols in use worldwide descend from that work and from the 1970s AIGA symbol set, which was placed in the public domain.

Why does everything in a supermarket look the way it does?

Because packaging is doing several jobs at once and most of them are not aesthetic.

Shelf position and eye level are bought, not earned. Colour blocking lets a brand be identified at three metres and at a glance. The shape carries category information — you can identify a shampoo bottle, a milk carton and a medicine box in silhouette.

Two specific patterns are worth noticing. Premium products use more whitespace, smaller type and fewer colours, because a crowded package signals cheapness; the message is that the product does not need to shout. And health claims cluster on the front while the ingredient list, which is legally required and actually informative, is in the smallest permitted type on the back.

The ingredient list has one genuinely useful rule: ingredients are in descending order by weight. If sugar is second, the product is mostly sugar, whatever the front says.

What comes next

The next Part turns from what you look at to what you read — the books everyone names, where words come from, and the writing systems of the world.