Online video calling app: what every control on the screen is for

The scepticism is earned. You used something like this in 2011, it was a text box with video beside it, and it was mostly people who wanted to be looked at. Then it closed, and the replacements arrived within weeks, which is itself suspicious.

Here's the half you're right about: the software is not better. It is, in most respects, the identical software wearing a different layout. Nobody solved matching, nobody solved moderation, and the fundamental exchange — two strangers, one button, no way back — is exactly what it was.

What you're wrong about is the crowd, and the crowd shifted because of the layout and not in spite of it. That's the thing the tour is for. Every control below moved or grew or disappeared since 2011, and each of those changes pushed a different sort of person in or out.

The first control, and the one you reach before anything exists at all. On the old services it was a pair of buttons — text or video — and the existence of that choice is most of what separated 2011 from now.

One button means the camera is assumed. Which sounds trivial and isn't: it removes everybody who wanted to watch without being seen, and those people were a large share of the old crowd and the source of most of what you remember badly.

The two seconds after the press are the browser asking for the camera. That's the browser, not the service — a permission attached to the address, which is why it asks again on a different domain and why a browser video chat can't quietly keep the camera after you leave.

Second control, and it isn't really a control. It's where you sit between connections, and it is the part people complain about most.

The reason usually given is boredom, which can't be right: eight seconds is too short to be bored in.

Here's what I think it is. The wait is the only stretch of an online video calling app with nobody else in it, and memory keeps those stretches and drops the rest. Ask somebody how an evening went and they'll describe the waiting, because the waiting is the part they were alone for and therefore the part they were present for. The service is remembered as emptier than the log would show.

1. You press, and you enter a pool of people who are also pressing at that second.
2. The matcher pairs you with one of them. There is no list and no queue — the pair didn't exist a moment earlier.
3. Both sides get the same picture at the same time. Neither was shown the other first.
4. If nobody is pressing, you wait — and the wait measures how fast people leave their current conversations, not how many people are on the service.

The instruction that follows is daft and works: turn your head. Find an object across the room, and look at it until the picture changes. You arrive at the next connection having been elsewhere, which is a different state to arrive in.

Third control, and it lives in the same screen: the small rectangle in the corner with you in it, which is the one people most often switch off.

Leave it on. The small rectangle isn't there as a mirror; it is the one gauge an online video calling app gives you. Without it you can't tell that the lamp behind you switched off, that you've slid to the edge of the shot, or that the lens locked on to the wallpaper instead of your face.

So the rule is: keep the gauge, glance at it the way a driver glances at a mirror, and never study it.

Fourth control, and the one that has changed the most since 2011 — because in 2011 it was the whole product and now it's an accessory.

What it's for now is single words. A place name, a spelling, a number, the thing neither of you can pronounce. It is not for conversation, and using it for conversation closes more calls than any other habit here, because typing takes you out of the picture and the other person watches you stop being present.

Fifth and sixth controls sit next to each other on most layouts — the skip and the report — and that proximity causes a specific confusion worth clearing up.

The skip ends the pair and does nothing else. No message reaches anybody, no record is made, nothing attaches to a name that doesn't exist. It's instant and free, and far and away the most honest thing in the interface — a service that makes leaving enormous and obvious is telling you that leaving is normal.

The report does something different and on a different clock. It goes into a queue that people read, in order, days later. It will not end your evening better and it will not visibly punish anybody tonight. It is still worth using, and the disappointment people describe comes entirely from expecting it to behave like the skip.

Seventh control, usually paid, and the one I'd argue you should leave alone for the first several evenings.

It works. That's not the objection. The objection is arithmetic: every filter cuts the pool, and a cut pool means longer waits — not proportionally longer but multiplicatively, because the matcher now needs two suitable people pressing in the same second rather than any two.

A second reason the online video calling app hour matters, and it concerns the wires rather than the people. A browser video chat runs on whatever the network happens to be doing, and networks are busiest when everybody is at home. The same address delivers a visibly worse picture at nine in the evening than it does at one in the morning; the service then gets blamed for what is really the street outside.

And the clock does the same job for nothing. Somebody's evening is a better filter than their passport, because what you actually want is people with nothing after this, and that's a property of the hour rather than of the country.

1. Work out the local time where you'd like to be talking.
2. Wait until that place is two hours past its dinner.
3. Open it then, with no filters at all.
4. Compare that evening against a filtered one at your own convenient hour. The difference is not subtle.

Both are worth a paragraph, because both are commonly misread.

The first is a language preference, which is not the same thing as a country filter and is usually more useful. A country tells you where somebody's body is; a language tells you what will happen in the first ten seconds, and the opening ten seconds are the whole of the problem.

The second is a flag for the conversation rather than the person — a way of saying this call was good, which a few services collect and none of them visibly do anything with. My guess is that it feeds the matcher, quietly, and nobody has ever shown me evidence either way.

Both are worth trying once and neither is worth paying for, on the same arithmetic as the country filter: every narrowing costs waiting, and waiting is the one cost that compounds.

A word on phones against laptops while we're here, since the tour has assumed a screen without saying which. The camera on a phone is better and the microphone is worse, because the microphone is at arm's length and pointed at the room. A laptop reverses both.

Neither difference decides anything, and people spend a surprising amount of argument on it. What does decide something is whether the device can be put down. A phone held in a hand hands the far end a room in constant small motion, and that reads as agitation whatever the camera can do.

Eighth, and my advice is don't, which surprises people.

It strips away your own small picture, which we've just established is the one gauge you have. And it strips away the room you sit in: with the browser in view you're somebody at a table who happens to be on a call; without it, the call is where you are. An hour of the second leaves a different mark on an evening than an hour of the first.

There's one more thing fullscreen takes, and it's specific to a mobile video chat rather than a laptop: on a phone it hides the address bar, and with the address bar gone you lose the only quick way to tell whether you're in a real browser or inside some app's built-in one. That distinction decides whether the camera works at all.

The exception is a conversation that's already working. Twenty minutes in, with somebody showing you something, fullscreen is right and the instrument doesn't matter any more. Use it as a reward rather than a default.

This is the part of the tour where there is nothing to point at, and it answers your original scepticism better than the rest.

Nothing counts who is online. No ranking and no badge, no streak, no tally of how long tonight has lasted, no profile, no past. On anything else on your phone all of those exist, and each is a reason to stay.

The absence is a matter of structure rather than modesty. Money in a mobile video chat turns up when a reader crosses to a paid service, so your minutes pay nobody, so nobody ever had cause to build a reason for spending more of them. It's an absence, so it never gets advertised, and it's the clearest difference between this and the rest of the device.

One more thing the tour has skipped, because it isn't a control and has no button.

No indicator anywhere tells you how you sound, and there's no preview for it the way there is for the picture. You can see yourself and you cannot hear yourself, and that asymmetry accounts for a surprising share of the evenings that go badly for reasons nobody can name.

The remedy is wired earphones, and the reason isn't privacy. Without them the speaker feeds the microphone, so the far end receives its own voice again half a beat out of step — intolerable, and never said aloud. They leave, you file it away as one more skip, and you invent a story about your face.

Which closes the tour of things that exist. What follows concerns why anybody presses the button at all, and the layout has no part in it.

Now the part that isn't interface, because the tour doesn't explain the behaviour on its own.

People say they open these things when they're bored, and everybody nods, and nobody looks at what boredom is. It isn't emptiness. An empty hour is fine — an empty hour is a bath or a walk. Boredom is the refusal to choose: an hour spent cycling through options and starting none of them.

Which explains the appeal exactly. A random video chat makes the choice for you, and remakes it every ninety seconds, and the relief of that is not something the old services ever advertised because none of them understood it either. The first roulettes sold novelty. What actually sells is the removal of the decision.

Which also explains a thing people report and can't account for: that a random video chat is more absorbing than a feed despite containing less. A feed asks you to choose forty times an hour, quietly, and the choosing is the tiring part. Here the choosing has been taken away, and all that remains is attention.

I'd put that as the strongest argument for the whole format, and it's an argument from weakness rather than strength, which is why nobody makes it in marketing.

One example, because the abstract version above is unfair to the actual experience.

A cook in Hanoi after the kitchen had shut, standing in front of a microwave with his own reheated dinner in it, waiting the way everybody waits in front of a microwave — which is to say with total attention and no purpose whatsoever.

We had perhaps sixty shared words. He turned the phone to show me the counter, then the pass, then the microwave again with eleven seconds left on it, and the eleven seconds were genuinely the most interesting part of my evening. Nothing came of it. He ate. I left.

The whole exchange happened inside a browser video chat with no application installed on either side, which is worth saying because people assume an app is involved and then go looking for one in a store. There isn't one. There's a page, and the page asks for the camera, and that's the entire apparatus.

That's the product, and I want to be honest that it sounds thin written down. What it isn't is boring, in the specific sense above: for eleven seconds neither of us had a single decision to make, and that is rarer at eleven at night than it ought to be.

Here's where I argue with myself, though, because there's a version of everything above that's bleak, and I'd rather not pretend otherwise.

At three in the morning the composition changes. A larger share of the people awake are by themselves and would prefer not to be, and the removal of the decision I praised two sections ago starts to look like something else — a way of not choosing to do anything about being alone, for an hour, and then for two.

There isn't a clean answer to that from me. The one honest thing I can offer is the distinction: this is good at filling an hour that was already going to be empty and bad at filling one that wanted something else. Nobody can tell which kind of hour you have except you, and the service will never ask.

Which is, in its way, a point in its favour. It doesn't pretend to be medicine.

Last observation before the answer, and the one I'd most like you to keep.

The awkwardness of the first thirty connections is real and everybody has it. What's interesting is what makes it stop, because it isn't courage and it isn't practice in the ordinary sense.

It's repetition with a reset. Thirty first-seconds in a row, none of which remembers the previous one, is a training apparatus you cannot assemble anywhere else in ordinary life — a party gives you six of those in an evening, and it remembers all six. Here you get forty, and the forty are independent.

People come out of a fortnight of this measurably easier with strangers offline, and I'd have dismissed that claim before I noticed it happening to me.

A last note on the tour itself before the answer. Every control described above was added by somebody trying to fix a complaint, and every one of them fixed the complaint and produced a new one.

The filter fixed «I never meet anyone from there» and produced «why is the waiting so long». Fullscreen fixed «the window is too small» and produced «I can't see myself». The report fixed «nobody does anything» and produced «nothing visible happened». Which is not an argument against fixing things — it is a reason to expect the next version of this to have more controls and the same evenings.

So, the question from the first paragraph. Which control actually decides how the hour goes?

None of them. Not the filter, not fullscreen, not the text box, not the skip. Every one of those changes the texture of a connection and none of them changes who you meet, because who you meet is decided before any of them: by what hour it is where the other people are.

The control that matters is a clock, and it's on your wall rather than on the screen. Pick the hour when somebody somewhere is two hours past dinner with nothing after it, and every other control becomes a detail. Pick eight in the morning and no amount of filtering will save the evening.

Which is, incidentally, why the old services and the new ones feel so different despite the software being the same. The layout moved the crowd, the crowd moved the hours that are worth opening, and the hours are the only control you actually have.

Pick an hour and try it