Code Brain | The Creator's Aesthetic

These days we rarely hear the word "taste." People use other terms, but it's still a fundamental concept we can't do without.

Whether in the past or present, scientists, engineers, musicians, architects, designers, writers, and painters all use the same word to describe good work — beauty. Is this merely coincidence? Or do they share some common meaning? And if there is a common thread, can we take the patterns of "beauty" discovered in one field and apply them to another?

For those of us who design products, this is more than a theoretical question. If beauty truly exists in the world, we need to be able to recognize it. We need good taste to create good things. Rather than treating beauty as some ethereal abstraction, or navel-gazing around such ethereal concepts, let's consider a practical question: how do you make beautiful things?

When it comes to "taste," many people say "taste is subjective," and they genuinely believe this. They don't know why they like something. Maybe because it's beautiful, or because their mother owned one, or because they saw a movie star using it in a magazine, or because they know it's expensive. Their thinking is often unclear.

Most of us have been taught, consciously or unconsciously since childhood, not to sort through our tangled thoughts. When you tease your younger brother for coloring people green in his coloring book, your mother will likely say something like "you each have your own aesthetic."

At this moment, your mother isn't teaching you the essentials of aesthetics — she's just trying to stop your argument.

Like much of what adults tell us, this is contradictory and ambiguous. After teaching you that taste is merely personal preference, they take you to a museum and tell you: look closely, because da Vinci was a great artist.

What effect does this have on children? How will they understand "great artist"? After years of being told "everyone's taste is just their own way of doing things," they're unlikely to grasp directly that a "great artist" means someone's work is better than others'. Instead, they're more likely to see "great artist" as simply matching their personal preference — like someone saying in a book "broccoli is good for you."

Calling taste a matter of personal preference is a fine way to avoid arguments. The problem is, it's not true. You sense this the moment you start designing things.

Whatever someone's job is, they naturally want to do it well. Soccer players want to win games; CEOs want to increase profits. Doing your job well brings genuine pride and pleasure. But if your job is designing products, and you don't acknowledge that beauty exists, there's no way to do your job better. If taste is just personal preference, then everyone is perfect: you simply do what you like, and that's enough.

Like any work, when you keep at design, you get better. Your taste changes. And, like others improving at their work, you know you're getting better too. If this is true, your past taste wasn't merely different — it was worse. The notion that "there's no such thing as good or bad taste" collapses under its own weight.

"Relativism" is fashionable now, and even after you grow up, it can still hinder your thinking about taste. But if you step outside your narrow self and at least admit to yourself that there is such a thing as good design and bad design, then you can begin studying good design carefully. How does your taste change? What mistakes do you make? What have others learned from design?

Once you start thinking about these questions, you'll discover that beauty is strikingly similar across different fields. The principles of good design appear again and again.

1

Good design is simple

You hear this in mathematics and painting alike. In math, it means shorter proofs are often better proofs. For mathematical axioms, "less is more" applies especially well. In programming, this largely holds true. For architects and designers, it means beauty should depend on a few carefully chosen structural elements rather than surface ornament. (Ornament itself isn't bad — it only becomes bad when used to disguise weak structure.) Similarly in painting, still lifes done from careful observation and solid models tend to be more interesting than flashy, repetitive work. In writing, this means: express what you mean in brief sentences.

This emphasis on simplicity may seem odd. You might think things are naturally plain, and decoration means more work. But when people attempt creative work, something shifts. Beginning writers adopt grandiose tones that sound nothing like how they actually speak. Designers use flourishes and waves to show their artistic sensibility.

2

Good design is timeless

Every mathematical proof, if correct, is timeless. As the mathematician Hardy said: "There is no permanent place in the world for ugly mathematics." He agreed with aircraft designer Kelly Johnson — the best solution won't be ugly, and eventually someone will find a better one. Aiming for timelessness is a way to help yourself find the best answer: if you don't want someone else's answer replacing yours, you must provide the best answer yourself. Some masters' work is so outstanding that it leaves almost no room for later improvement. Sculptors after Dürer all had to live in his shadow.

Aiming for timelessness is also a way to avoid trends. By definition, trends change over time. So if you make something that looks good in the distant future, its appeal must come from its own qualities rather than from being "trendy."

Strangely, if you want your work to appeal to future people, one method is to make it appeal to past people too. We can barely guess what the future will be like, but we can be certain that future people will care as little about current trends as past people did. So if your work appeals to both people today and people from 1500 years ago, it will very likely appeal to people living in 2500 as well.

3

Good design solves problems

A kitchen stove has four burners arranged in a square. Each burner is controlled by a knob. How do you arrange the knobs? The simplest placement is a row of four. But this isn't optimal. These knobs are for human use. If you arrange them in a row, people must pause each time to figure out which knob controls which burner. A better approach: arrange the knobs in a square, mirroring the burners themselves.

Much bad design works hard but in the wrong direction. In the mid-20th century, sans-serif fonts were fashionable. These fonts were purer, closer to letterforms themselves, but this wasn't the most important problem for printed text. For legibility, what matters more is being able to distinguish letters from each other. Though it looks somewhat Victorian, the lowercase g in Times Roman clearly distinguishes itself from lowercase y.

Solutions can keep improving, and so can problems. In software, a thorny problem can often be transformed into an easier equivalent. Physics made far more progress when it stopped using scripture to justify itself and instead framed its problems as predicting observable behavior.

4

Good design is suggestive

Jane Austen's novels hardly describe what anything looks like, yet she tells stories so vividly that readers imagine the scenes themselves. Similarly, suggestive paintings like the Mona Lisa tend to be more compelling than descriptive ones. Everyone has their own interpretation of the Mona Lisa.

This principle in architecture and design means a building or object should allow users to employ it as they wish. A good building, for example, fades into the background of whatever life people want to lead, rather than making them execute a program of the architect's predetermined life.

In software, this means providing users with basic elements they can combine freely, like Lego bricks. In mathematics, it means a proof that becomes the foundation for many new results is preferable to one that is difficult but leads nowhere. In science, citation count can generally serve as a rough indicator of value.

5

Good design is often funny

This principle may not hold in all cases. But Dürer's sculptures, Saarinen's Womb Chair, the Pantheon, and the original Porsche 911 design all seem funny to me. Gödel's incompleteness theorem is amusing like a joke.

I think this is because humor reflects power to some degree. Having a sense of humor is something strong people do; maintaining humor means laughing off misfortune, while losing it means being wounded by it. Thus one mark of strength is not taking things too seriously. Confident people are like swallows, viewing everything lightly. Hitchcock's films, Bruegel's paintings, even Shakespeare — all excellent examples.

Good design doesn't have to be funny, but I find it hard to imagine a humorless design being good design.

6

Good design is hard to do

If you observe people who do great work, one thing they seem to share is that they work very hard. If you're not working hard, you're probably wasting your time.

Solving difficult problems requires enormous effort. In mathematics, difficult proofs demand elegant solutions, which are often interesting. Engineering is the same.

When climbing a mountain, you must discard everything unnecessary from your pack. With difficult sites or insufficient budgets, architects can only make spare designs. When the challenge becomes solving the problem itself, the frills are cast aside.

Not all difficulty is beneficial. There's useful pain and useless pain. What you need is the pain of gritting your teeth and sprinting forward, not the pain of stepping on a nail. Solving hard problems may be good for designers, but fickle clients or unreliable materials are another matter.

In art, portraiture traditionally holds the highest status. This isn't merely because faces move us more than other images, but because we're so good at observing faces that portrait painters must work doubly hard to satisfy us. If you paint a tree and angle a branch 5 degrees off, no one notices. But angle a person's eye 5 degrees, and everyone sees it.

When Bauhaus designers adopted Sullivan's "form follows function," their actual principle was "form should follow function." If merely making the function work is hard enough, then form must be entirely determined by function — there's no room for error. Wild animals are beautiful precisely because they live hard lives.

7

Good design looks easy

Great athletes make winning look effortless, and great designers do too. But this is usually an illusion. Excellent writing that reads easily has often been revised repeatedly.

Some of the most important discoveries in science and engineering look so simple that you say to yourself, I could have thought of that. But then, why didn't you?

Some of da Vinci's portraits are just a few lines. You look at them and think: if I put these ten or so lines in the right places, I could draw such a beautiful portrait too. Yes — but you need to find where those right places actually are. The tiniest error makes the whole piece collapse.

In fact, line drawing may be the most difficult because it demands near perfection. In mathematical terms, lines are closed-form solutions; less skilled artists can only approximate the same problem through iteration. Many children abandon drawing around age ten, partly because they decide to start drawing like adults, and the first thing they try is drawing faces with lines.

In most fields, "looking easy" is achieved through practice. Practice trains what's called "muscle memory." Sometimes you train your body. Piano masters play notes faster than their brains can send signals to their hands. Artists too, after sufficient practice, see with their eyes and it flows automatically from their hands, like tapping your foot to rhythm.

When people say they're "in the zone," I understand this to mean they have complete control of themselves. This makes you less hesitant, freer to think when facing difficulties.

8

Good design is symmetrical

Symmetry may simply be one way to achieve simplicity, but it's so important it deserves separate mention. The abundance of symmetry in nature shows its significance.

There are two kinds of symmetry: repetition and recursion. Recursive symmetry means repetition of sub-elements, like the veins in a leaf.

Historically, symmetry was once so overused that it's become unfashionable in some fields. Since Victorian times, architects have consciously built asymmetrical buildings. By the 1920s, "asymmetry" became an explicit premise of modernist architecture. Even so, these buildings are often only asymmetrical on the main axis, with details still heavily using symmetry.

In writing, you'll find symmetry everywhere, from phrases in sentences to plots in novels. You'll find the same in music and art. By composing entire works from identical elements, mosaic art (and some of Cézanne's work) achieves tremendous visual impact. Symmetrical composition produces some of the most memorable paintings, especially works where two halves echo each other, like The Creation of Adam and American Gothic.

In mathematics and engineering, recursion is especially useful. Inductive proofs are both concise and beautiful. In software, problems solvable by recursion are usually the best solutions. The Eiffel Tower is striking partly because its form is recursive — big towers with smaller towers on them.

The danger of symmetry is that it can replace thinking, especially when used in heavy repetition.

9

Good design resembles nature

It's less that resembling nature is good in itself, and more that nature has solved many problems through long evolution. So if your design resembles nature, that's a good sign.

Imitation differs from copying. Few object to a story being written like real life. In painting, subjects taken from life are often misunderstood, yet they're also extremely valuable tools.

Resembling nature works in engineering too. From early on, ships were built like animal thoraxes, with spines and ribs. Though sometimes people must wait for better technology: early aircraft designers shaped planes like birds, but this was wrong because they lacked sufficiently light materials or power (the Wright brothers' engine weighed 152 pounds but produced only 12 pounds of thrust), or the kind of complex mechanical control systems birds have. But I can imagine small unmanned reconnaissance drones in fifty years flying like birds.

Computers are now powerful enough that we can simulate not just nature's environments but its evolutionary results. Genetic algorithms may let us create things too complex to design through conventional means.

10

Good design is redesign

Few people get things right on the first try. Professionals discard early work; revision is part of the plan.

You need some confidence to convince yourself to discard work. You need the ability to think, because you need to understand why it became what it is. For example, people beginning to paint are usually reluctant to redraw mistakes. They feel it's already good enough, and if they redraw parts, it might get worse. On the contrary — they believe their drawing isn't that bad, perhaps because they originally thought so.

This is dangerous. You should cultivate dissatisfaction with yourself. Da Vinci often drew a line five or six times to get it right. The Porsche 911 prototype was unbalanced; only after redesign did its distinctive rear emerge. Wright's early design for the Guggenheim had the right half somewhat like a pyramid temple; he inverted it to create what we see today.

Mistakes are normal. Don't treat them as disasters — admit them boldly and correct them. Da Vinci more or less reinvented sketching as an art form, making drawing a way to explore more possibilities. Open source software has fewer bugs because they admit the possibility of bugs.

The right tools make revision easier. Oil paint replaced tempera in the 15th century, letting painters more easily tackle difficult subjects like the human body, because oil can be mixed and glazed while tempera cannot.

11

Good design can copy

Our attitude toward copying often swings back and forth. Beginners unconsciously imitate others; as they gain skill, their work consciously strives for originality. Eventually, they realize getting things right matters more than being original.

Unconscious imitation almost inevitably leads to bad design. If you don't know where your ideas come from, you may be imitating another imitator. The Raphaelite school dominated mid-19th century painting so thoroughly that nearly everyone learning to paint imitated him (often through multiple layers of imitation). It was concern about this very phenomenon that led to the Pre-Raphaelite Brotherhood.

The ambitious don't stop at imitation. When your taste advances to the next stage, you consciously begin attempting originality.

I think the greatest masters eventually reach a state beyond ego. They simply want to find the right answer, and if someone else has found it, there's no reason not to use it. They're confident enough to use others' results without worrying about losing their individuality.

12

Good design is strange

Some outstanding works are simply inexplicable: Euler's formula, Bruegel's Hunters in the Snow, the SR-71 Blackbird, the programming language Lisp. They're not just beautiful but strangely beautiful.

Perhaps I'm not clever enough to be certain why. A dog seeing a can opener probably thinks it miraculous. If I were a genius, I might find e^(i*pi) = -1 perfectly natural. After all, it is.

Most characteristics I've mentioned earlier can be cultivated, but I don't think "strangeness" can be. At best, you can avoid killing it when it begins to appear. Einstein didn't want relativity to be strange; he sought truth, and truth itself proved strange.

I studied at an art school where students' ideal was developing their own style. But everyone walks differently to begin with. Michelangelo didn't try to create the Michelangelo style; he simply tried to make good work, and his style formed inevitably.

Whatever style you develop emerges naturally. This is especially true of strangeness — there are no shortcuts. Mannerists, Romantics, generations of American high school students have all sought this "Northwest Passage," which seems not to exist. The only way to get there is to pursue making good work, then look back afterward.

13

Good design comes in clusters

The great artists living in 15th century Florence included Brunelleschi, Ghiberti, Donatello, Masaccio, Fra Filippo Lippi, Fra Angelico, Verrocchio, Botticelli, da Vinci, and Michelangelo. Milan was as large a city as Florence then. How many 15th century Milanese artists can you name?

Florence had unique conditions then that couldn't be sustained — today's Florence isn't the same. We must assume that whatever genius da Vinci and Michelangelo possessed, someone in Milan possessed it too. So why didn't Milan produce its own da Vinci?

Today, the US population is roughly a thousand times that of 15th century Florence. In theory, a thousand da Vincis and a thousand Michelangelos exist among us; we should witness artistic miracles daily. But this doesn't happen, because creating a da Vinci requires not just genius but 1450s Florence.

Nothing is more powerful than brilliant people gathering to solve a difficult problem together. Compared to this, talent is trivial: da Vinci's genetic advantage couldn't overcome the gap between being born in Milan versus Florence. Today people are far more mobile, yet great achievements still disproportionately emerge in hotspots: the Bauhaus, the Manhattan Project, The New Yorker magazine, Lockheed's Skunk Works, Xerox PARC.

In any historical period, there are hotspot fields and groups producing great work. If you're distant from them, achieving greatness on your own is nearly impossible. You can influence trends at most, but you can't escape them. (Perhaps you can, but Milan's da Vinci clearly didn't.)

14

Good design is bold

In any era, people take certain absurd things as correct with such conviction that questioning them risks ostracism or violence.

If our era were different, this would be wonderful. But as far as I know, it isn't.

This problem afflicts not just every era but more or less every field. Much Renaissance art was considered too secular at the time: according to Vasari, Botticelli confessed to the church and gave up painting, while Fra Bartolomeo and Lorenzo di Credi burned some of their own work. Einstein's relativity offended many contemporary physicists and wasn't fully accepted for years — France didn't accept relativity until the 1950s.

Today's experimental error is tomorrow's new theory. If you want to make great new work, you shouldn't ignore where common sense contradicts truth — you should pay special attention to these points.

Actually, I think finding ugliness is easier than imagining beauty. People who make beautiful work seem to do so by fixing what they find ugly. Great work often emerges when someone sees something and thinks: I can do better than this. For centuries, everyone was satisfied with traditional Byzantine Madonnas painted from a template, until Giotto looked at them and found them stiff and unnatural. Copernicus was troubled that his contemporaries could tolerate geocentrism, while he felt there must be a better explanation.

Merely being unable to tolerate ugliness isn't enough — only when you're deeply familiar with a field can you identify what needs improvement. You need constant practice. When you become an expert in a field, you'll hear that faint inner voice: "This is terrible! There must be a better way." Don't ignore this voice; let it take root. The trick to doing excellent work: very demanding taste, plus the ability to satisfy it.

Original link: https://paulgraham.com/taste.html This essay is collected in Paul Graham's 2004 collection Hackers & Painters: Big Ideas from the Computer Age