Two screens never add up to one

Two 27-inch monitors have less panel and fewer pixels than one 40-inch monitor, plus a seam down the middle that nobody can move. The standard two-screen desk is not more screen. It is less screen, cut in half.

That half of the argument is just arithmetic, and it is in the table below. The half people raise an eyebrow at is that mine stands up - a 40-inch 16:9 on an arm, rotated by 90 degrees, which is how it spends most of its life. Four years ago I wrote that this screen has no successor and praised it for working like two big, almost square screens side by side. That was the compromise talking: 16:9 is the only shape that does both jobs, and at 4K it is finally tall enough to be worth turning.

A big screen is divisible, small ones are not addable

A big screen can pretend to be any number of small ones - FancyZones, a tiling window manager, or two maximised halves - and the split changes in a second to suit the task.

Small screens cannot do the reverse. Two 27-inch monitors never become one 40-inch one: the seam is permanent, and so are the two stands, the two calibrations and the window that has to pick a side.

That asymmetry is the whole argument. One surface divides any way you like; several surfaces never add up.

Graphics cards taught the same lesson. Two mid-range cards never rendered one scene as well as a single strong one, and Nvidia stopped writing SLI profiles after 2020. Different reasons - frame pacing rather than plastic bezels - identical shape: one strong card drives several screens, and no number of weak ones makes a strong one.

Not work against play, but column against grid

The web is a column. Mobile-first indexing finished the job in 2023, and horizontal scrolling is a bug. Source files, unified diffs, logs, terminals, chats, pull requests, documentation: same shape. Content lives vertically, and width is mostly margin.

It is tempting to call wide screens the entertainment shape and leave it there, but that is not the line. Cinema and games are wide - and so are spreadsheets, side-by-side diffs and video timelines, which are work by anyone’s definition. The split is between columns and grids, and columns are what most of us stare at all day. (Entertainment only stayed wide on the desk anyway; on a phone it went vertical years ago.)

Upright, my 4K panel is 2160 x 3840: around 200 lines of code in one editor at a typical 19-pixel line, against about 110 in landscape. A 400-line review becomes two screenfuls instead of four.

16:9 is the only ratio that survives rotation

Rotate an ultrawide and you get a chimney: one column wide and taller than a door. Rotate 4:3 and the standing shape is fine, but lying down it pillarboxes every film. 16:9 turns into two text columns and turns back into cinema - not ideal in either orientation, just the only one that is usable in both. That is the whole case for the ratio, and it is why the screen I would replace this one with is still 16:9.

The honest caveat: upright does not fill itself either. A page whose container stops at 1200 pixels leaves 900 unused beside it. Height comes for free; width only pays off when you tile - documentation beside the editor, the app beside devtools. Upright just moves the empty space to where a second window fits.

On a laptop, the ratio is the only rotation you get

A screen on an arm can be turned, so its shape only has to survive both orientations. A lid cannot be turned: whatever height it has, it keeps for good - and height is what runs out first.

So the ratio should track the size, in the opposite direction to the market: the smaller the panel, the squarer it should be. A small laptop wants 3:2, a larger one 16:10, and 16:9 only starts to pay where the height is there anyway. No laptop is big enough for that. The largest ones sold are 18-inch 16:10 machines - 24 cm of panel height, less than half of what a 40-inch 16:9 has lying down, and a quarter of what it has standing up.

On a lid the width is not even a choice - the keyboard sets it. Height is the only free variable, so the ratio is how you spend it. In the same 13-inch chassis, 16:10 is 11% more panel than 16:9 and 3:2 is 19% more, all of it height, and the number on the box grows with it, from 13.3 inches to 13.9. The panel is not what you read on, either: the tab strip and the address bar take about 2.7 cm off the top whatever the shape, so the page grows faster than the glass - 13.9 cm of web page on 16:9 against 16.9 on 3:2, 22% more page for 19% more panel. Centimetres rather than pixels, because zoom moves pixels and leaves the glass where it is; set the same text size on every screen and this is the comparison that survives. Then hold a phone next to it. An iPhone 15 keeps 12.5 cm for the page against the 16:9 lid’s 13.9, so whatever a laptop wins there, it is not height: it wins on width, on long lines rather than more of them. Vertically, the machine most people work on gives a web page about as much room as the phone in their pocket.

The objection is video, and at a fixed width it does not hold: a 16:9 film is exactly as large on the 3:2 panel, with the extra height left over above it. It only shrinks if you compare at equal diagonals - which is the comparison shops make, because for the same glass a wider ratio prints a bigger number. The same panel area sells as 13.3 inches at 16:9 and as 12.8 at 3:2, so half an inch of the diagonal is marketing rather than screen.

Which is why 3:2 and 16:10 coming back is the display trend of the last few years I am happiest about. It is the same argument as the big screen’s, arriving from the other end: on a laptop you buy height with the ratio, because you cannot buy it with rotation.

Geometry, not marketing

My screen is a curved 40-inch 16:9, so 88.6 x 49.8 cm of panel. Against the two dual setups people usually compare it with:

One 40-inch panel has more surface and more pixels than two 27-inch ones at the same density - and the pixels are continuous.

Eye level is a rule about staring, not about height

OSHA wants two things at once: the centre of the screen 15-20 degrees below eye level, and the top of it at or below eye level. On a tall panel those two halves pull against each other.

At 70 cm, both halves hold only up to about 51 cm of panel height - which a 40-inch 16:9 lying down (49.8 cm) just squeaks under, and which nothing standing up can meet. The rule has no answer for a tall screen, so the thing to read is its mechanism.

It is quoted far more often than it is explained. It rests on two measured things: the resting posture of the eyes is about 15 degrees below horizontal, and looking up opens the lids wider - Tsubota and Nakamori measured 1.2 cm² of exposed eye in downgaze against 3.0 cm² in upgaze, with tear evaporation about 240% higher. Blinking drops from about 22 a minute to 7 in front of a screen.

Both mechanisms are about sustained gaze, not about where the panel ends. Holding your eyes up for an hour dries them; glancing up for two seconds does not. This is a comfort optimum averaged over a working day, not a health limit - and the reported prevalence of dry eye among screen workers, anywhere from 9.5% to 87.5%, says more about questionnaires than about eyes.

The sharper objection is not to the rule but to the people it is aimed at. Hardly anyone spends the day looking up; they spend it looking down, at a laptop on a desk or a phone in their lap. If the common error is a bent neck, a screen that lifts the eyes is a correction rather than a hazard.

Up to a point, and the point is where eyes stop and vertebrae start. Gaze shifts of 20 to 30 degrees are made by the eyes, the head joining in mostly to keep them inside a comfortable range - and a head held tilted back is not a cure for a head held forward, only a different fixed posture. The thing being corrected is shakier than it sounds, too: the link between forward head posture and neck pain holds in adults, vanishes in adolescents and is confounded by age. So what a tall screen offers is not a better angle but the absence of a single one - the eyes travel a range all day instead of being pinned to a spot.

Mine sits with its centre roughly at eye level, which puts nearly half the panel above it and the top strip about 32 degrees up - far enough that reading there earns a small nod, which is rather the point. Seven years, no complaints: one desk’s worth of evidence, not a recommendation.

So the height costs nothing, and the movement it provokes may be worth something. The editor lives at eye level; the shelf above it holds a running build, a log, a browser I glance at. Two things to check on your own panel first: a curved screen now bends top to bottom, and a VA panel’s viewing angles rotate with it.

The one honest argument for a real second screen is narrower than it looks: on a video call, a single screen means the meeting covers the work you are talking about.

Fewer screens, fewer invitations

The “42% more productive with multiple monitors” figure is vendor-sponsored, and Coding Horror picked it apart years ago. The quieter results are the useful ones:

• NEC/University of Utah: one 24-inch widescreen beat two 20-inch screens by 6% on text editing - and lost that lead on spreadsheets, exactly along the column-grid line.
• Colvin et al.: no significant difference in completion time between one and two monitors.
• Hutchings et al.: more screen space raises window-management overhead, sometimes to counterproductive levels.
• A 2021 survey of developers at home: the strongest influence on perceived productivity was not the setup, but interruptions.

Underneath all of them sits a confound worth naming. When these studies ran, a second monitor was the only way to buy more pixels, so “two screens” and “more working area” were the same variable and no result could separate them. That is not the choice anyone faces now. The NEC/Utah setup is the one that pulls the two apart - and it pulls against the dual desk: the pair of 20-inch screens carried 48% more panel and 67% more pixels than the single 24-inch widescreen, and still lost the text-editing task.

So no percentages from me, just a distinction. A second screen is not extra space, it is a standing invitation: somewhere to park a chat window where it costs nothing. On one surface everything that wants my attention has to take it from what I am working on, which makes it my decision rather than the desk’s.

How it sits on the desk

A VESA arm behind the screen, the original stand gone, the desk underneath it free again.

Rotating takes a few seconds - pull it towards me, turn it, push it back. Upright for work, landscape for films and for gaming with my daughter, which is still the best argument for a big 16:9 that I know.

None of which is a love letter to my particular monitor. The successor I asked for in 2022 still does not exist, and turning the old screen on its side was only the upgrade I could have had back then. Upright, 4K is enough, not plenty: put a 6K or 8K 16:9 on the shelf, curved, and it goes on the arm the same afternoon. Only the band at eye level has to behave itself - and shelves can always be taller.

• email
• facebook
• twitter
• linkedin
• reddit
• tumblr
• pinterest
• pocket
• vk.com