Breakdowns: Objects. Seventeen things you can hold, and what they teach about things you can't.

Warm-up: how to study objects

You have already met Rams's ten principles, Fukasawa's "without thought" and Kenya Hara's emptiness in the starter guide. Here you stop quoting them and start reading them in the objects themselves. Principles are compressed lessons; objects are the uncompressed source. When you read the source you see the trade-offs the principles hide.

Each product chapter has the same shape, so you can skim or study. A facts box. An annotated drawing with numbered pins. A list of numbered decisions that match the pins, each with why and what it cost. What it left out. Where it falls short. Then "Steal this for your software", a quiz and a short exercise. The drawings are schematic recreations, simplified on purpose: they show structure, not exact proportions. Look up photographs of the real thing as you read; the gap between the drawing and the photo is itself a lesson in what matters.

Study plan: one part per sitting. Keep a notebook (paper or a text file) titled "Steal list". Every time you hit a "Steal this" item that applies to your product, copy it in with today's date. By the end you'll have twenty to forty candidate changes, and the final chapter helps you pick three.

A breakdown is not a review. You are not judging whether you like the object; you are reverse-engineering the decisions that made it, and the costs those decisions paid, through seven fixed lenses.

Most people "study" good design by looking at it and feeling something. That feeling is real but it doesn't transfer. You can't open your editor and type "make it feel like a Braun". What transfers is a decision: this radio puts its only frequently-used control under the thumb, and hides the rest. Decisions are portable. Feelings are not.

The trouble is that decisions are invisible once they're made. A finished object looks inevitable, as if it could not have been any other way. Your job in a breakdown is to make it look chosen again. For every feature ask: what were the alternatives, why was this one picked, and what did it cost? If you can't name a cost, you haven't understood the decision yet. Every real decision costs something.

The seven lenses

Look through each lens in order. The order matters: it goes from why the thing exists, to how it looks, to how it's used, to what it's made of, to the small stuff, to its relatives, and finally to the absences, which are the hardest to see.

1 Purpose

What job does it do, for whom, in what situation? What is the one sentence that, if false, would make the object pointless? (The Kindle: "read books for hours without your eyes or the battery giving up.")

2 Form

Overall shape, proportions, hierarchy of parts. What is biggest? What do you see first? Does the form explain the function, or hide it?

3 Interaction

Every control: where it is, how it moves, what feedback it gives, and whether its current state is visible without touching it. Map the most common task, step by step.

4 Material and medium honesty

Is each material used for what it's good at? Does anything imitate something else (plastic pretending to be wood)? How does it age: does wear make it worse or more personal?

5 Details

Radii, gaps, the click of a switch, the label typeface, the colour of a single key. Details are where care becomes visible, and where cheap shortcuts show.

6 System and family

What is it part of? Siblings in the same line, modules that combine, accessories, refills, the shop it's sold in. Great objects are rarely alone.

7 What was left out

Features competitors had that this lacks. Controls you expected and didn't find. Labels, logos, colours, settings removed. The absences are where the courage is.

A worked example: the humble wall light switch

Take an object you touch daily and never think about. A good rocker light switch, the large flat kind common in European homes.

1. Purpose. Turn a room's light on or off, in the dark, with a full hand of shopping, without looking.
2. Form. A large flat rectangle. The whole face is the control, so the target is huge. (Fitts's law in plastic; see The Human Mind.)
3. Interaction. Press top or bottom. The rocker's tilt shows state at a glance, and the snap gives sound and touch feedback in under 50 ms.
4. Material honesty. Matte plastic that hides fingerprints. No fake brass.
5. Details. Rounded edges, no sharp seam to catch a fingernail. Sometimes a tiny glowing dot so you can find it at night.
6. System. Same frame for sockets, dimmers and blanks, so a wall of them reads as one family.
7. Left out. No label, no screen, no app, no "are you sure?". Costs: no scheduling, no remote control. Most people never miss them.

Now translate: the most-used action in your product should be a large target, show its state at rest, respond instantly, and ask nothing. That's one sentence of design direction, earned from a light switch.

The one-page breakdown template

Use this every time. One page, never more: the limit forces you to rank. Fill it in pencil or in a text file. Keep every one you write; after ten you'll have a personal pattern library no course can give you.

In the wild

• Rams's ten principlesDieter Rams, formulated late 1970s to 1980s
  Rams wrote his principles after decades of designing, as a summary of what his objects had already done. They are the output of breakdowns, not the input. Principles come from studying decisions; study the decisions and you'll write your own.
• The Design of Everyday ThingsDon Norman, 1988 (revised 2013)
  Norman's doors, stoves and taps are breakdowns of ordinary objects, each ending in a general idea (affordances, signifiers, mapping, feedback). Ordinary objects are as instructive as famous ones.
• Teardown sites like iFixitFounded 2003
  Hardware teardowns photograph every screw and score repairability. They're "lens 4 and 6" breakdowns: material and system. A narrow lens, applied consistently across many products, becomes a powerful comparison tool.

Physical products get their discipline for free, from cost, space, weight and physics. Software has none of those limits, so you must translate the reasons a physical decision was made, and then impose the discipline yourself.

A Braun radio couldn't have forty buttons: each one meant a hole in the case, a switch, wiring, assembly time, and a few more pfennigs on every unit sold. A pocket radio couldn't be bigger than a pocket. A Game Boy had to run on four AA batteries for a whole car journey. The designers often turned these limits into virtues, but the limits came first.

Your app can have forty buttons for free. It can be any size, any colour, any number of screens. Adding a setting takes ten minutes. That freedom is why so much software sprawls. Nothing pushes back.

So when you study an object, separate three things: the surface (colour, texture, shape), the constraint (what physics or cost forced), and the reason (the human need the decision served). Translate the reason. Recreate the constraint on purpose when it helped. Leave the surface behind.

Four kinds of translation

1. Direct. The idea works the same way on screen. The ET 66's colour-coded keys become a calculator app's coloured operator column. The Swiss clock's single red accent becomes one accent colour in your UI.
2. Metaphorical. The behaviour transfers, not the mechanism. MUJI's pull cord doesn't become a drawn cord on screen; it becomes "the main action happens with one obvious gesture and no menu".
3. Constraint transfer. A physical limit becomes a self-imposed budget. The Game Boy's battery life becomes a performance budget (load in under one second, run on a five-year-old phone). The 606's fixed rail spacing becomes a spacing scale.
4. Anti-translation. Some lessons tell you what not to bring across. A physical shutter dial is great because fingers find it without looking; a drawn dial on a touch screen loses that, and is often worse than a plain list.

The constraint budget

Because software won't push back, write your own constraints down and treat them as physics. Examples that work for a solo product: at most five user-facing settings; one accent colour; first screen usable with no account; page weight under 200 KB; cold start under one second on a mid-range Android phone; any action undoable for 10 seconds. These are your "price per button". Breaking one should feel as expensive as adding a part to a production run.

In the wild

• iOS 7Apple, 2013
  Removed the leather, felt and linen of earlier iOS apps and kept layered depth and motion to show structure. Drop the costume, keep the information the costume was carrying (hierarchy, state).
• Halide camera appLux, 2017
  Takes cues from manual cameras (exposure and focus as direct, visible controls) but uses swipe gestures fitted to a phone rather than drawn dials. Translate the behaviour of a camera, not its chrome.
• Teenage Engineering's OP-1 field and its software2022
  Even a hardware company's on-screen graphics are flat, playful diagrams, not drawings of the hardware. The screen speaks in its own language.

1. A clear acrylic lid instead of a wooden or metal one. You see the record, the arm and the dial with the lid closed: the machine explains itself. The usual account is that a metal lid was tried and rattled, and transparent acrylic solved the problem while showing the insides. Cost: acrylic scratches and shows dust, and the coffin jokes. Clear dust covers later became standard on turntables, which is the best evidence the decision was right.
2. A white-painted metal body, not a veneered cabinet. It reads as equipment, closer to a laboratory instrument than a sideboard. Cost: to many buyers in 1956 it looked cold and unfinished. Braun accepted a smaller market for a clearer identity.
3. Light wood only at the ends. The one warm material sits where hands touch and where the object meets the room. It isn't decoration; it's the structural end of the box. Cost: two materials to source and join, but the warmth made the white bearable at home.
4. Every control in one field on the top plate. Knobs and keys line up in a grid beside the turntable, where your eyes already are when you put on a record. Nothing on the sides, nothing hidden. Cost: you must open the lid to change station.
5. A slotted metal speaker front instead of cloth. The slots are the speaker's honest opening, cut into the same material as the body. Cost: less forgiving of dents than fabric; the pattern must be precise because it is the whole face.
6. Small, plain labels. Short words in a clean sans-serif, no logos competing with function. The type supports the controls instead of announcing the brand.

What it left out

Gloss veneer, brass trim, cloth grilles, curved "radio cabinet" shapes, a large logo, legs. Most importantly it left out disguise: the assumption that technology in the home should look like something else. That single absence is the reason the object still matters.

Limits and criticism

It was a combination unit with a built-in speaker, so sound quality was bounded by the box, and Braun's own later hi-fi went modular (separate amplifier, tuner, speakers). It was a product for a modernist minority; many households kept their walnut radiograms. And "honest" is not the same as "warm": the SK 4 needed those wooden ends to be livable.

In the wild (software with a clear lid)

• iA WriterInformation Architects, 2010
  The document is the interface; formatting is visible as plain Markdown characters rather than hidden behind a toolbar. Show the material, not a disguise of it.
• Stripe DashboardStripe
  Lets developers see the API requests and responses behind payments in its logs. Transparency is a feature for people who must trust the machine.
• Git clients like Fork or Tower2010s
  Show the diff, the branch graph and the staged state on one screen. One control field next to the material it changes.

1. The speaker grid is the face. Instead of hiding the speaker behind a decorative grille, a regular field of holes makes the sound's exit the main visual pattern. Cost: the hole pattern must be precise and evenly spaced, or the whole face looks wrong.
2. The tuning disc is the biggest control. Tuning was what you did most, so it got a large wheel under the thumb, readable at a glance. Cost: less room for anything else, which was the point.
3. The scale is on the control itself. Numbers on the disc rotate past a fixed mark, so the control and its display are one object. No separate pointer and window to align in your head (Norman would call this good mapping).
4. A pocket-shaped body with soft corners. Sized and rounded for a hand and a pocket rather than a shelf. Cost: a small speaker with limited sound, accepted as the price of portability.
5. Nothing else on the front. Secondary functions stay out of the face. The front shows only what the listener needs while listening and tuning.

What it left out

Chrome bezels, a large logo, multiple small knobs of equal size, decorative typography, a carrying handle that would add bulk. It also left out the idea that a radio must look like a miniature of a bigger radio.

Limits and criticism

A single-band pocket radio with a small speaker. The rotating disc is lovely for slow browsing but slow if you know exactly what frequency you want, the same tension the iPod's wheel would face with large libraries. A big, single control is only right when one task really dominates.

In the wild

• Apple iPodApple, 2001
  A screen above and one large round control below; Jony Ive later wrote the foreword to Sophie Lovell's 2011 book on Rams. The lineage is about hierarchy, not about circles.
• ShazamShazam, 2008 on iPhone
  For years, the main screen was dominated by one huge button to identify a song. When one job dominates, let one control dominate.
• OvercastMarco Arment, 2014
  The now-playing screen gives play/pause and skip the biggest targets; less frequent options sit smaller. Size by frequency.

1. A quiet display, set back. A plain digit window at the top, nothing around it but the dark body. The answer is the only thing to read. Cost: no second line of history, which software calculators later added.
2. Round, convex keys. Each key is a small dome. The shape says "press me" to the fingertip, centres the finger, and makes adjacent keys feel separate even without looking. Cost: more complex moulding than flat keys, and fewer keys fit.
3. Function keys in their own tone. Clear, percent and similar keys share a colour distinct from numbers, so the "dangerous" or unusual keys look different from the ones you press all the time.
4. Operations in a column, in their own colour. The arithmetic keys line up on the right edge in one tone. Your eye learns "verbs on the right, nouns in the middle". The grouping does the work a label would otherwise do.
5. One accent key. The equals key carries the only bright colour, marking the action that completes the task. A single accent is strong because it's alone. Cost: once used here, the colour can't be spent anywhere else on the object.
6. A strict grid. Equal spacing and alignment so the colours read as regions, not as a random pattern. Colour coding only works on a disciplined layout.

What it left out

Scientific functions, a second display line, a solar panel strip across the face, printed brand slogans, and the visual noise of every key competing equally. It also left out multiple accent colours: the restraint is what makes the one accent mean something.

Limits and criticism

Colour coding by tone alone is fragile: in poor light, or for people with low contrast sensitivity, a dark key on a dark body can disappear. The labels are small. And a homage can become a trap: by the 2010s, colour-coded round keys became a "Braun look" applied to apps where the grouping itself was not clear.

In the wild

• iPhone CalculatorApple, since 2007
  Round keys, a dark ground, numbers in one tone and the operators in a single orange column. Widely described as a homage to the ET 66. The homage works because the grouping logic came along, not only the look.
• GitHub's merge buttonGitHub
  The green merge action stands out as the completion step; destructive branch deletion uses a different, red treatment. Roles get colours, and the colours stay put.
• Things 3Cultured Code, 2017
  A mostly neutral interface where colour is used sparingly for meaning (for example the yellow star for Today, blue for the add button). Scarce colour carries meaning.

1. A wall rail as the foundation. The structure lives in slim aluminium tracks fixed to the wall, so the furniture has no heavy sides or legs. Everything else hangs from it. Cost: needs a wall you can drill, and careful installation.
2. A fixed hole pitch as the "grammar". Holes at a regular spacing let any shelf go at any height. The spacing is the interface between parts, and it hasn't been changed. Cost: you can't fine-tune to the millimetre; you live on the grid.
3. Few part types. Shelves, cabinets, desks and a few accessories. A small vocabulary keeps manufacture simple and keeps combinations predictable.
4. Backward compatibility as a promise. New parts fit old systems. Owners extend instead of replace, and the company sells to the same customer for decades. Cost: the company can never "redesign" the interface; innovation must happen within it.
5. Plain, neutral finishes. The system disappears behind the books and objects on it, and doesn't date as fashions change.
6. Designed to be moved and reconfigured. The shelves come down and go up again in a new home, with a new layout. Vitsœ has long offered planning help and buys back or resells used systems. The relationship continues after the sale.

What they left out

The 606 left out fixed sizes, matching "sets", and the idea of a finished piece of furniture. There's no "complete" state; it is always ready for the next change. Both objects left out planned obsolescence: nothing about them invites you to replace the whole thing.

Limits and criticism

Systems ask more of the buyer: you have to plan. The 606 is expensive compared with flat-pack shelving, and the fixed grid will occasionally fight an odd wall. Backward compatibility also locks the maker in; every new part inherits decisions from 1960. That's a feature for owners and a constraint for designers.

In the wild

• Plain text and Markdown filesMarkdown by John Gruber, 2004
  Apps like iA Writer and Obsidian store notes as plain files, so your notes outlive the app. An open, stable format is your wall rail.
• The web platform1990s to today
  Browsers still render pages written decades ago. "Don't break the web" is a backward-compatibility promise at planetary scale. Longevity is a design decision you keep making.
• Unix command-line tools1970s onward
  Small programs that each do one job and combine through text streams. A small vocabulary plus a shared interface generates endless configurations.

1. A simple geometric case. A square or rectangular body around a round face. The case stays matte and dark (or white), so the face is the only bright, readable area. Cost: almost no room for brand expression.
2. Numerals chosen for reading, not style. Clean sans-serif digits, sized so the face reads at arm's length, often with all twelve numbers rather than marks only. Cost: a busier face than a "minimal" four-marker dial, accepted because reading the time quickly is the whole job.
3. A hierarchy of hands. Hour hand shortest and thickest, minute hand longer and thinner. Weight tells you which is which before length does.
4. One accent colour, on the moving part. The second hand (yellow on many Braun designs) is the only colour. It shows the clock is running and gives the face life without adding noise. Cost: the accent can't be used for anything else, including the brand.
5. The alarm setting shown on the face. A distinct hand or marker shows when the alarm will ring, so you check it with a glance in bed, not by pressing a button.
6. Few, well-placed physical controls. A large snooze or alarm-off area that's easy to hit half-asleep; small, recessed setting wheels you only touch occasionally. Frequency sets the size, again.

What it left out

Date windows on most models, luminous decoration, multiple colours, a big logo on the face, textured dials, Roman numerals. It also left out any second accent: no red alarm button competing with the yellow hand.

Limits and criticism

Matte dark faces can be hard to read in the dark without lighting, and analogue faces in general are slower to read precisely than a digital display. Some models' setting wheels are small and fiddly. The style has been copied so much that a yellow second hand now reads as "Braun-like" decoration on products that lack the underlying hierarchy.

In the wild

• Apple Watch facesApple, 2015 onward
  Several faces use neutral dials with a single accent colour for the second hand or complications chosen by the user. One accent, user-chosen, on what moves.
• LinearLinear, 2019
  A largely monochrome interface where colour is mostly reserved for status and priority. Colour as state, not decoration.
• Google Calendar's current-time lineGoogle
  A single red line marks "now" across the day view. The accent marks the live moment.

1. Hang it on the wall like a fan. Mounting it on the wall takes it off the shelf and puts it at eye level, where a ventilation fan would be. Placement is part of the meaning. Cost: you need a wall spot near a socket, and it's not portable.
2. The disc spins in view. A clear cover shows the CD turning. You see that it's playing without any light or display, and the motion recalls fan blades. Cost: dust, and a visible mechanism has to look good while moving.
3. A pull cord as the main control. One pull to start, one to stop. No reading, no choosing, no learning. Your hand has done this before. Cost: a cord is imprecise for anything beyond on and off, so other functions need separate controls.
4. The speaker is the body. The square face is the speaker surround. No separate speakers, no wiring between boxes. One object, one shape.
5. Secondary controls kept small and out of the face. Volume and track buttons are small and sit at the edge. The face is reserved for the disc.
6. A visible power cable, accepted. The cord to the socket hangs down. Fukasawa didn't hide the honest fact that it needs power; the object stays simple enough that one cable doesn't spoil it.

What it left out

A display, a remote control, a stack of buttons on the face, a lid you open with a button, a brand name on the front, stereo separation. It also left out instructions. The point is that you don't need any.

Limits and criticism

CDs went away; the object's medium aged faster than its idea. Sound from a single small box is limited. And "without thought" depends on a shared memory: someone who never pulled a fan cord gets a neat gadget, not a moment of recognition. Borrowed behaviours are cultural, and not everyone has the same ones.

The best design dissolves into behaviour.Paraphrase of a recurring idea in Naoto Fukasawa's talks and his book with Jasper Morrison, "Super Normal" (2007)

In the wild

• Pull-to-refreshLoren Brichter, Tweetie 2, 2009
  Brichter's gesture borrowed the natural move of scrolling past the top to "ask for more". It spread across nearly every list-based app. The best gestures feel discovered, not taught.
• Spotlight and RaycastApple, 2005; Raycast, 2020
  One keystroke, then type what you want. Typing a name is behaviour everyone already has. Search is the software pull cord.
• AirPods caseApple, 2016
  Opening the case near an iPhone offers to connect. The natural action is the command. Turn an action people already take into the trigger.

Emptiness is not simplicity

Hara draws a distinction that matters for software. Simplicity, as European modernism practised it, is reduction toward a correct, rational form: remove until only the essential remains. Emptiness is something else: a form that is open enough to receive whatever the user brings. A simple object says "here is the right answer". An empty object says "here is room for your answer". A plain MUJI notebook isn't the correct notebook; it's a notebook that becomes yours as you fill it.

Hara's famous illustration is the MUJI "Horizon" advertising campaign of 2003: photographs of nearly nothing, a horizon line over the salt flat of Uyuni in Bolivia and over Mongolian grassland, with the small MUJI logo. The images carry no message. Viewers project their own. That is the brand's whole idea in one picture.

"This will do" versus "this is what I want"

Hara describes MUJI's target feeling with two Japanese phrases. Kore ga ii: "this is what I want", a strong, specific preference, the feeling most brands chase. Kore de ii: "this will do", which in MUJI's sense is not resignation but a calm, confident satisfaction. The product is good enough that you stop searching. It doesn't need to be your favourite; it needs to be right enough that you never think about it again.

Software chases kore ga ii relentlessly: delight, personality, endless customisation so each user gets exactly what they want. That produces settings pages with ninety toggles and products that need tuning before they work. A kore de ii product ships excellent defaults, and most people never change them.

The system's decisions

1. Materials shown as they are. Unbleached or off-white paper, kraft card, undyed fabrics where possible. Skipping bleaching and dyeing was originally a cost saving that became an aesthetic. Cost: a narrower colour range, and some people read it as "cheap".
2. Transparent and translucent plastics. Clear gel pens show their ink level; translucent polypropylene boxes show their contents. The material tells you what's inside.
3. Sizes that fit each other. Storage boxes, drawers and shelves are sized to stack and nest together across the range, so a new purchase fits an old one. A system, like the 606.
4. The label does the talking, and it's quiet. Product name, material, size and price on a simple label or tag; usually no logo on the product itself. Information replaces branding.
5. Standard formats and consistent stationery. Notebooks in standard paper sizes and simple rulings; pens in a few tip sizes and colours. The range is deep but not wide, so choosing is fast.
6. The shop as part of the product. Stores use natural materials and calm lighting with products shown plainly. The retail experience and the objects speak one language.

What it leaves out

Logos on products, seasonal colour trends, character licensing, aggressive packaging, sale-driven urgency, and the idea that each product should be someone's favourite. It deliberately avoids peak emotion in favour of steady trust.

Limits and criticism

"No brand" became a very strong brand; the absence of a logo is now instantly recognisable, which is a paradox Hara himself has discussed. Quality varies across a range this large, and "this will do" can slide into "this is mediocre" if the defaults aren't actually excellent. Emptiness only works when the empty thing is very well made.

In the wild

• iA WriterInformation Architects, 2010
  Famously shipped with very few settings and a carefully chosen typeface and measure; the team argued that good defaults were the feature. Kore de ii in software: decide, so the writer doesn't have to.
• Apple NotesApple
  An almost empty page with a few formatting options; used by a huge number of people precisely because it asks nothing. "This will do" can be the most-used product in its category.
• Plain-text notebooks: Obsidian, Bear2020, 2016
  Quiet canvases that become personal through use and structure the user adds. Empty vessels let users build their own meaning.

1. Design the place, not the product. Instead of adding a new object (a bucket, a rack), modify the environment where the behaviour already happens. Cost: it only works where you control the floor; it's not a product you can sell in a box.
2. Start from observed behaviour. The idea comes from watching people lean umbrellas against walls and set the tips in tile joints. Observation precedes form.
3. Use the existing support. The wall already holds the umbrella; the groove only stops it sliding. Do the smallest thing that completes what people are already doing.
4. No visible "design". Nothing to recognise as designed. The object succeeds when nobody notices it, which makes it impossible to market on looks.
5. Judge by use over time. A Super Normal object is evaluated after weeks of living with it, not in a showroom. The first-impression test and the long-use test are different tests, and Super Normal favours the second.

What it leaves out

Signature shapes, "statement" colours, surprise for its own sake, and the designer's ego. Super Normal leaves out the moment where someone says "oh, that's designed". ±0's products tend to be calm, softly geometric household items that look like idealised versions of what you already expected a humidifier or a toaster to look like.

Limits and criticism

Super Normal is hard to sell: it doesn't photograph well against loud competitors, and it can be mistaken for blandness. It also relies on having a well-observed "normal" to perfect; for genuinely new product categories there's no normal yet, and someone has to design the first one boldly.

In the wild

• Natural-language dates in Fantastical and TodoistFlexibits, 2011; Doist
  Typing "lunch with Asha next Friday 1pm" creates the event. It perfects the most normal input there is: writing a sentence. Super normal input beats novel widgets.
• macOS Finder's "Recents" and sensible defaultsApple
  Saving to a sensible default location and showing recent files reduces filing work without a new tool. Change the environment, not the user.
• Plain HTML forms done wellgov.uk design system, 2010s
  The UK Government Digital Service built a reputation on plain, extremely well-tested form patterns. Boring, perfected, and used by millions.

1. E-ink instead of an LCD. Reflective electronic paper reads like print in daylight, doesn't flicker, and uses power mainly when the page changes. That's what gives weeks of battery. Cost: slow refresh, ghosting, poor video and animation, and (until recently) no colour.
2. A keyboard, then none. The 2007 keyboard made searching and buying possible without a touchscreen. Once touch arrived, the keyboard became dead weight and went. The device got smaller, cheaper and more about reading. Lesson: a feature right for one era can be wrong for the next.
3. Page turns on the edges. The original had long page-turn bars on both sides so either hand could turn. The core action was physical and always under the thumb.
4. Battery measured in weeks. The headline spec was never speed or pixels. Battery anxiety ends; you stop thinking about charging, which is part of disappearing.
5. A front light, not a backlight. The Paperwhite lit the screen from the edges onto its surface, keeping the reflective, paper-like quality while allowing night reading. Later models added warm light to shift the tone toward amber in the evening. Cost: slight unevenness near the light source on some units.
6. Physical page buttons returned. The Oasis brought back buttons, because tapping glass with the hand that's holding the device is awkward. Touch is flexible; buttons are effortless. Both stayed.
7. An asymmetric grip. The Oasis moved the weight to a thicker edge so one hand holds it like a book's spine, with the screen rotating so either hand works. Form shaped around how people actually hold a book for an hour.

What it left out

Notifications from other apps, a general web browser worth using, email, social feeds, video, games, a glowing backlight. This is the decision that matters most. A tablet can show books too; it also shows everything else, all the time. The Kindle's absences are its product.

Limits and criticism

The device is tied to one store and ecosystem, a real cost to owners. Lock-screen advertising on cheaper models undercuts the calm. The software has often lagged the hardware: slow menus, cluttered home screens that push store recommendations. And e-ink refresh still makes navigation feel sluggish. "Disappearing" is a goal, not always an achievement.

In the wild

• Kobo e-readersRakuten Kobo
  A close competitor with physical page buttons on some models, warm light, and support for open formats like EPUB. The same "disappear" goal, plus a more open system.
• iA Writer's focus modeInformation Architects
  Dims everything but the current sentence or paragraph while you write. Define the "while" state and remove the rest.
• reMarkable tabletsreMarkable, 2017
  An e-paper tablet for writing that deliberately leaves out apps, browsing and notifications. Absence as the headline feature.

1. Hierarchical lists, all the way down. Artists, albums, songs, playlists: every screen is a list, and every list works the same way. Learn one screen and you know them all. Cost: deep libraries mean many levels.
2. A circular control for an endless list. A wheel has no end, so you can keep turning through a thousand items without lifting your thumb, unlike up/down buttons you must press a thousand times. The circle maps to "continuous".
3. Acceleration. Turn slowly and the selection moves one item per notch, precisely. Turn fast and it jumps many items per notch. One control serves both browsing and travelling. Cost: overshoot, which the slow mode then corrects.
4. Buttons on the wheel. The click wheel put Menu, Play/Pause, Back and Forward under the wheel itself, at the four compass points. Fewer separate parts, one area for the thumb. Cost: less tactile separation than distinct buttons.
5. A centre select button. Turn to choose, press the centre to go in, press Menu to go back. Three verbs cover the whole interface.
6. A click you can hear. The wheel produced a small click sound from the device as you scrolled, feedback that matched the speed of your movement. Cheap, and it made a touch surface feel mechanical.

What it left out

An FM radio, a recording function, an equaliser on the main screen, on-device file management, a removable battery, and, at launch, any way to load music other than syncing from a computer. Moving playlist editing and library management to iTunes on a big screen with a keyboard let the device do one thing with one control.

Limits and criticism

Early iPods were Mac-only and FireWire-only, and the non-replaceable battery drew criticism and a class action in the United States. Acceleration overshoots; with very large libraries, browsing alphabetically by wheel still beat search only because the device had no keyboard. When touch screens and search arrived (iPod touch, iPhone), the wheel's job largely moved to flicking with momentum, a different solution to the same problem.

In the wild

• iOS momentum scrolling and the alphabet indexApple, 2007 onward
  Flick to travel, touch to stop, slide the side index to jump to a letter in Contacts. The wheel's coarse/fine idea, rebuilt for glass.
• Apple Watch Digital CrownApple, 2015
  A small wheel for scrolling and zooming on a screen too small for fingers, with haptic detents. When the screen is tiny, a physical wheel returns.
• Video scrubbers in YouTube and iOS players2010s
  Scrubbing gives a coarse jump, with finer seeking available on many players. Any long timeline needs coarse and fine.

1. A small screen with playful, abstract graphics. Instead of numeric parameter lists, each sound engine and effect has a visual metaphor that moves as you play. You learn by watching what changes. Cost: less precise than numbers; some graphics are more charming than clear.
2. Four colour-coded encoders. Each knob has a colour, and whatever it controls on screen is drawn in that colour. You never read a label to know which knob does what on the current screen. The mapping changes per screen; the colours stay. It's the ET 66 lesson applied to dynamic controls.
3. Mode keys with icons, not menus. A row of buttons jumps straight to synth, drum, tape, mixer and other modes. Flat navigation, one press to anywhere.
4. A four-track tape as the recorder. The OP-1 simulates a four-track tape machine, with limited tracks and tape-like editing. Limits force decisions (bounce, commit, move on), which many musicians found freeing. Cost: power users hit the ceiling; that's the point and the frustration.
5. Transport controls borrowed from tape decks. Record, play, stop: familiar symbols from cassette machines. Borrowed behaviour again.
6. A personality throughout. Typography, sounds, tiny animations and even the packaging share a cheerful, engineered character. Personality makes people forgive limits and want to explore.

What it left out

A large screen, unlimited tracks, a full-size keyboard, deep menus, realistic skeuomorphic graphics of studio gear, and the assumption that a professional tool must look serious. The Pocket Operators left out the case.

Limits and criticism

Price: the OP-1, and even more the OP-1 field, are expensive for what they do, and the price rose over time. Some functions hide behind button combinations you must memorise. Playful graphics can obscure the actual value of a parameter. Constraints are freeing for sketching and frustrating for finishing.

In the wild

• Twitter's 140 characters2006 to 2017
  A technical limit (SMS length) turned into a writing style. Constraints shape what people make.
• Figma's multiplayer cursorsFigma, 2016
  Each collaborator's cursor and selection share a colour with their name. Colour as a live link between actor and effect.
• PlaydatePanic, 2022
  A small, yellow, one-bit-screen handheld with a crank, designed with Teenage Engineering. Personality plus constraint as a whole product idea.

1. A shutter speed dial with printed numbers. The dial's position is the setting. Read it from above at a glance, set it before raising the camera, even with the power off. Cost: discrete steps, and a dial takes space on the top plate.
2. Aperture and focus on the lens itself. The aperture ring and distance scale sit where those variables physically act. Mapping is direct: turn the lens to change what the lens does. Experienced photographers pre-focus by the scale without looking through the viewfinder.
3. A focus tab under the finger. Many M lenses have a small tab on the focus ring; its position tells you roughly the focus distance by touch alone. A state readable by feel.
4. An optical viewfinder you look through, not at. The rangefinder window shows the world directly, with frame lines, and you can see outside the frame (what is about to enter). Cost: parallax at close range and no exact preview of the image (the X100's hybrid finder adds an electronic overlay to address this).
5. Removing the brand mark. The "-P" models left off the red dot to make the camera less conspicuous for street photography. A product that takes its own logo off to serve the user's purpose.
6. Removing the screen (M-D). No rear display, no chimping (checking every shot). Just an ISO dial on the back where the screen would be. Extreme, controversial, and a clear statement: you're here to look at the world, not at the camera.

What they left out

Mode dials with twelve scene modes, deep menu trees for basic exposure, autofocus (on the M), zoom lenses (on the X100, a fixed lens), the red dot (M-P), and the screen (M-D). Each removal made the remaining controls more legible and the camera more about looking.

Limits and criticism

Manual focus rangefinders are slow for fast action and hard for close-ups. Leica prices are extreme; you pay a lot for less. Dials can be knocked accidentally (some have locks for this reason). And dial cameras still have menus; they just keep the essential variables out of them, which is the real lesson.

In the wild

• HalideLux, 2017
  Shows exposure and focus state on screen while shooting, with manual controls reachable by thumb. Camera-app state on the surface, not in settings.
• Gmail's search chipsGoogle
  Filters appear as visible chips under the search field, each removable. Active state as a readable, removable object.
• macOS menu bar mode indicatorsApple
  Focus modes, recording and microphone use show as icons in the menu bar while active. Important states stay visible while they're on.

1. A monochrome reflective screen. Mature, cheap LCD technology with no backlight. It drew little power and kept the price down. Cost: blur on fast motion, hard to see in poor light, no colour, criticised at launch as outdated.
2. The cross-shaped directional pad. Yokoi's team had already used the "plus" pad on Game & Watch handhelds. A flat, thin, durable control that the thumb operates without looking, far better than a small joystick in a pocket device.
3. Two action buttons, angled. A and B sit at a diagonal matching the natural arc of the right thumb. Few buttons kept games simple to learn.
4. A battery light. A small red LED by the screen dims as batteries run low. One honest indicator of the one resource that mattered.
5. Four AA batteries for many hours. Battery life was a headline feature: a long journey's worth of play, with batteries available in any shop. Rivals with colour screens needed more batteries and lasted a fraction as long.
6. Cartridges and a killer game. Tetris, bundled in many markets, was perfect for a small monochrome screen and short sessions. Hardware and software designed for each other.
7. A link cable for play together. Two Game Boys connected for two-player games. A cheap cable created a social feature, years before online play.

What it left out

Colour, a backlight, a TV tuner add-on as a core feature, extra buttons, and the race for specifications. The Wii, later, left out high-definition graphics (at launch) and the complicated many-button controller that intimidated non-players.

Limits and criticism

The screen really was poor in dim light, spawning an industry of clip-on lights and magnifiers. "Withered technology" can become an excuse for being behind; Nintendo later had periods where hardware lagged too far. The Wii's motion controls were imprecise for some games and novelty faded for many owners.

In the wild

• "Choose Boring Technology"Dan McKinley, essay, 2015
  Argues each team has a few "innovation tokens" to spend and should use proven tech elsewhere. Yokoi's idea, restated for software teams.
• Pieter Levels's productsNomad List, Photo AI and others
  A solo maker known for building profitable products on a very simple, mature stack. Withered tech lets one person ship and maintain many things.
• WhatsApp's early years2009 to 2014
  A small engineering team served hundreds of millions of users on a mature platform (Erlang), with a deliberately simple product. Proven technology plus focus scales further than novelty.

1. Bars, not numerals. Thick black bars for hours and thinner ones for minutes read faster at distance and from any language than small numbers. Everyone knows where three o'clock is. Cost: precise reading of a single minute takes a moment longer up close.
2. Black hands shaped as simple bars. No ornament, no tapering tips; maximum contrast and weight on white. The hour hand is wider and shorter, the minute hand longer.
3. A single red accent, shaped like a signal. The second hand is the only colour, and its disc tip borrows the guard's signalling paddle: a cultural sign of "departure" built into the clock. Cost: none worth naming; it's one of the best accent decisions in design history.
4. A white face with a black rim. Maximum contrast in daylight on open platforms. The rim frames the clock against busy backgrounds.
5. Stop-to-go: sweep, pause, release. Because every clock waits for the master pulse, small mechanical differences never accumulate. The pause is visible, and it makes the system's synchronisation something you can see. Passengers learn that the minute changes exactly when the red hand restarts.
6. One design everywhere. Every station, every platform, the same clock. You learn to read it once. The clock becomes part of the network's identity and trust.

What it left out

Numerals, brand names on the face, decorative hands, a date, and independent timekeeping. Each clock gave up being accurate on its own in exchange for being exactly as accurate as all the others.

Limits and criticism

Without numerals the face is slightly slower for children and for anyone who learned to read time digitally. The stop-to-go mechanism only makes sense inside a synchronised network; a standalone watch that pauses is charming, not functional. And the design is so recognisable that its use without permission became a legal matter.

In the wild

• iPad Clock appApple, iOS 6, 2012
  Used a face very close to Hilfiker's design, which led to a licensing agreement with SBB. Great public design is also intellectual property; borrow ideas, not faces.
• Status pages (for example, GitHub Status)2010s
  One public, consistent view of system state with timestamps and history. The station clock of a software service.
• Google Docs' "Saving… / All changes saved"Google
  A small visible pulse of system behaviour that builds trust in autosave. Show the network working, briefly and quietly.

1. A large-diameter handle. A wider handle spreads force over more of the hand and needs less grip strength to control. Cost: takes more drawer space and looked clumsy to some buyers at first.
2. An oval cross-section. A round handle rotates in a weak grip; an oval one stays aligned, so the blade angle stays where you put it.
3. Flexible fins where the fingers press. Soft ribbed fins give under the thumb and forefinger, providing grip even with wet hands. The handle meets the hand rather than the reverse.
4. A soft, non-slip material. Santoprene is comfortable, grippy when wet and dishwasher-safe. Cost: more expensive than plain hard plastic or metal.
5. A hanging hole. A large hole at the end lets you hang it, and it doubles as a visual end-stop. Small, practical, everyday.
6. A sharp swivelling blade. All the grip comfort is wasted if the blade is dull. Inclusive design still needs excellent core function.

What it left out

The assumption that "assistive" products must look medical. OXO tools are sold as ordinary, attractive kitchen tools; nothing marks them as being "for" anyone. It also left out thinness as a value: many classic utensils prized slim metal elegance over comfort.

Limits and criticism

Big soft handles take up more space and cost more to make. The look, widely copied, became a generic "ergonomic" style that sometimes adds rubber without the research. And for some severe disabilities even a good mainstream tool isn't enough; inclusive design complements, not replaces, specialist assistive products.

In the wild

• Captions and subtitlesAcross streaming and social video
  Built for deaf and hard-of-hearing viewers, now widely used by people watching in noisy places or quietly in public. The curb-cut effect in software.
• Dynamic Type on iOSApple, 2013
  System-wide text size that well-built apps respect; useful for low vision, tired eyes and older users. One setting for a need, honoured everywhere.
• Microsoft's Inclusive Design toolkitMicrosoft, 2016
  Frames disability as permanent, temporary and situational (one arm, an injured arm, a parent holding a baby). Extreme users stand in for many everyday situations.

1. A big wire clip. A bent spring-steel clip, large and visibly springy, grips thick fabric and notebook covers and doesn't snap off. It's also the pen's most recognisable feature. Cost: some find it oversized; it's a deliberate trade of elegance for robustness.
2. An ink window. A cut-out in the barrel shows the cartridge or converter, so you see ink running low before you run out mid-sentence. The pen shows its own resource.
3. A triangular grip section. Flattened faces guide fingers into the tripod grip teachers recommend. The shape quietly corrects technique without instructions. Cost: people with unusual grips (including many left-handed writers or those who rotate the pen) find it awkward; the guidance excludes some.
4. A steel nib that swaps in seconds. Nibs slide on and off, so you can change width or replace a damaged one cheaply. Repair and customisation at the price of a coffee.
5. Tough ABS plastic in bright colours. A pen for school bags, not display cases. It survives drops and loss is not a tragedy. Colours make it personal and easy to find.
6. A shape designed around the hand, not tradition. Classic fountain pens copied cigar-like forms. The Safari's form starts from how fingers hold a pen.

What it left out

Gold nibs, precious materials, a smooth round grip, a hidden ink reservoir, and the idea that a fountain pen is a luxury object. It made a beginner's tool without making it a toy.

Limits and criticism

The triangular grip is the main complaint: it helps the intended grip and fights others. It uses proprietary cartridges unless you buy a converter. Some find the plastic look too casual for professional use. Guidance through form is powerful, but whoever sets the guide decides who it fits.

In the wild

• Notion and Craft templates2010s to 2020s
  Pre-shaped pages that guide structure (a meeting note with headings ready) without a tutorial. Form that places the fingers.
• iCloud and Google storage metersApple, Google
  Visible usage bars in settings and on some main surfaces, with warnings before the limit. The ink window; best when shown where you work.
• Credit-card fields that format as you typeStripe Elements and others
  Spaces appear in groups of four, the card brand icon updates. The input's shape guides correct entry.

1. Map: only three angles. Beck restricted lines to horizontal, vertical and 45 degrees. The eye follows straight lines effortlessly; the diagram reads as ordered. Cost: real routes curve, and some distortions mislead walkers.
2. Map: even spacing and an enlarged centre. Stations are spaced for legibility, not distance. The dense centre gets room; the outskirts shrink. Cost: two stations that look far apart may be a short walk.
3. Map: interchanges made obvious. Distinct symbols where lines meet make changing trains the visual focus, because that's the decision travellers make.
4. Signs: a band and a fixed order. In the New York system, signs share a consistent structure (the band, then route symbols, then words, then direction), so you learn where to look once.
5. Signs: routes as coloured discs. Each route gets a symbol and a colour, used identically on maps, signs and trains. Colour is backed by number or letter, so it still works without colour.
6. Signs: one plain sans-serif. Consistent, legible lettering across the system (at first a typeface called Standard, later Helvetica). Uniformity is how a chaotic system looks trustworthy.
7. Signs: arrows and information at the decision point. Signs say what you need where you need to choose, not before or after.
8. Map (Vignelli 1972): abstraction pushed further. Every station a dot, every line a clean coloured stroke, the city reduced to a diagram. Cost: New Yorkers know their streets and parks, and the map's geography felt wrong. It was replaced within seven years.

What they left out

Beck left out geography, scale, streets and most landmarks except the river. The New York signs left out the decades of mixed lettering and ad-hoc notices that had grown on the system. Vignelli's map left out almost all geography, which was one step too far for that city.

Limits and criticism

Diagrams shape behaviour: people ride trains between stations that are quicker to walk because the map makes them look far apart. Vignelli's map shows the limit of abstraction: you can remove the truth people don't need, but not the truth they already hold strongly (here, the shape of their own city). Later maps of both systems have negotiated between diagram and geography ever since.

In the wild

• Git branch graphsGitKraken, Fork, GitHub network views
  Commits drawn as stations on coloured lines, merges as interchanges. Beck's diagram language for version history.
• CitymapperCitymapper, 2011
  Journey results show legs, changes and walking time as a compact diagram rather than a map first. Answer "how do I get there", then show where.
• Progress steppers in checkout flowsMany e-commerce sites
  Steps shown as evenly spaced stations regardless of how long each takes. Topology over distance, the right truth for the question "how much is left?".

Across very different objects, the same few moves repeat: rank by importance and make the ranking visible, show state at rest, borrow behaviour people already have, spend one accent, design a system with a fixed interface, and leave out whatever doesn't serve the one experience you must win.

Read the comparison table by row to remember an object, and by column to see a lens across objects. Columns are where patterns appear: notice how often "what was left out" is the most distinctive cell in a row.

The comparison table

The master translation table: physical to software

Thirty translations from this guide, each in the form "physical lesson, software pattern, example". Use it as a lookup when a design problem feels familiar.

In the wild: software built from object lessons

• Things 3Cultured Code, 2017
  Neutral canvas, sparing colour, one obvious add button, careful details. Reads like a Braun and MUJI collaboration in software. Several translations combined coherently.
• iA WriterInformation Architects, 2010
  Empty vessel, excellent defaults, a "while" state, plain files as the rail. MUJI, Kindle and 606 lessons in one writing app.
• HalideLux, 2017
  State on the surface, controls under the thumb, a camera's logic without its costume. Leica and T3 lessons translated to glass.

The final exercise

Wrap-up

Reflection prompts

• Which object in this guide is closest in spirit to your product? Which is furthest? What would your product look like if it moved one step toward the furthest?
• What is your product's "left out" column today? What should be in it?
• Where do you copy surfaces in your current design, and what reason were you really after?

Practice assignment for the next two weeks

1. Days 1 to 2. Write your constraint budget (Exercise 2.1) and your settings purge table (Exercise 9.1).
2. Days 3 to 4. Frequency map and redesign of your main screen (Exercise 4.1), plus the colour-role audit (Exercise 5.1).
3. Days 5 to 6. Power-off test for visible state (Exercise 14.1) and the hardest-hands test (Exercise 17.1).
4. Day 7. Ship the smallest three improvements from the week. Write a changelog entry for each, naming the object that inspired it.
5. Days 8 to 11. Keep the day-30 diary (Exercise 10.1). Meanwhile, do the final exercise (20.1) on an object from your desk.
6. Days 12 to 13. Protect your "while" state (Exercise 11.1) and add one ink window (Exercise 18.1).
7. Day 14. Run the "object lessons" checklist in chapter 20. Note your score, and book a weekly 20-minute breakdown in your calendar from now on.

Reading list

• Dieter Rams: As Little Design as Possible
  Sophie Lovell, 2011, foreword by Jony Ive · the most complete visual record of Rams's Braun and Vitsœ work, with the products in this guide photographed properly.
• Less and More: The Design Ethos of Dieter Rams
  Klaus Klemp and Keiko Ueki-Polet (eds.), 2009 · exhibition catalogue with many drawings and prototypes; good for seeing decisions in progress.
• Designing Design
  Kenya Hara, 2007 · MUJI, emptiness and the Horizon campaign in Hara's own words.
• White
  Kenya Hara, 2010 · a short, beautiful book on emptiness and the colour white; read it before you design your next empty state.
• Super Normal: Sensations of the Ordinary
  Naoto Fukasawa and Jasper Morrison, 2007 · the ordinary objects they chose, and why. Trains your eye for "settles in" over "stands out".
• Naoto Fukasawa
  Phaidon monograph, 2007 · the CD player, ±0 and "without thought" projects explained.
• The Design of Everyday Things
  Don Norman, revised edition 2013 · the vocabulary (mapping, feedback, signifiers) you need to describe what these objects do right.
• Mr Beck's Underground Map
  Ken Garland, 1994 · the story and evolution of Beck's diagram by a designer who knew the material well.
• Helvetica and the New York City Subway System
  Paul Shaw, 2011 · a careful history of the subway signage, correcting popular myths about the typeface.
• Game Over: How Nintendo Conquered the World
  David Sheff, 1993 · a history of Nintendo that includes Gunpei Yokoi and the Game Boy; read it for the trade-offs behind a winning product with a losing spec sheet.
• "Choose Boring Technology"
  Dan McKinley, essay and talk, 2015 · Yokoi's lesson restated for software teams; essential for a solo creator choosing a stack.
• "The Curb-Cut Effect"
  Angela Glover Blackwell, Stanford Social Innovation Review, 2017 · why designing for the margin helps everyone; the OXO idea at policy scale.

Next: take the same method to screens in Breakdowns: Software, and to reading products in Breakdowns: Reading.