← All notes · 4 min read
The AI said 22 days. It took one.
What Apple likes to see in a game, a folding iPhone I don't own, and an estimate that was honest and still wrong.
Yesterday morning App Store Connect showed me a small notice: "Preparing for iPhone Duo." I don't own a Duo. I had not thought about it for one minute.
That notice started the day. By midnight Game Show had things in it that were on no list the morning before.
I asked the AI two questions. Does the game really work on the folding iPhone? And what does Apple like to see in a game before it puts one on the front page?
The number
The second answer came back as a plan. Game Center. Accessibility. A widget for the home screen. A live card on the lock screen. Each one had its risks written out, and at the bottom there was a total: 22 to 26 working days. Do it after launch, in two updates.
I asked why so long, based on what.
The reasons were fair. We couldn't build an iPhone app on my desk. Every build went to a rented Mac somewhere, waited in a queue for twenty or thirty minutes, and only then told us if a line was wrong. The lock screen card is a second small app living inside the app, with its own signature. The plan counted many failed rounds, and days of me testing on a phone between them.
I said today. Before the lock.
It took one day. Game Center went in with four leaderboards and seven achievements. The lock screen card went in. If you leave the game in the middle of a match against a friend, the score stays on your lock screen and around the camera. The first build check sat in the queue for twenty minutes, ran for 3 minutes 43 seconds, and passed on the first try. The failed rounds the plan had paid for never came.
Accessibility and the home screen widget are not done. Those wait for an update.
Since then I have seen both new things on my own phone. They work.
A phone I have never held
The answer to the first question was honest too. The game ran on the Duo, but in a small window in the middle of the big inner screen, because our package was built with last year's tools. The advice: don't hold the first release for this, put it in the first update.
Apple gives a badge on the store page to apps that are ready for the Duo. That badge matters to me. A new device has a short shelf, and I wanted to be on it from the first day.
So we designed for a screen neither of us can touch. I looked at everything through a simulator. The build machine now opens the game on a pretend Duo after every build and sends back screenshots.
Game Show is a TV studio, and a studio on a wide screen should not stretch. Now the stage opens from the center, on the question screen, the guest screen, the card selection, the backstage wall. The menu video is vertical, so on the wide screen its edges continue to both sides, blurred. The Skip button in the intro was hiding under the clock. It moved.
The tablet layout I had given up on the night before got fixed on the way. Same problem, it turned out.
One thing I left out. When the Duo sits half folded on a table, the question should be on the top half and the keyboard on the bottom. That is the first update.
Dressing for the editors
I worked at magazines. The App Store front page is one. People choose what goes on it, and they choose things that show what their own devices can do.
So besides the code we made what an editor can pick up. Events on the store page: a New Year's party, a weekly face-off, the champion of the week. For the pictures we made a real looking living room and put the game's actual screens on the TV and the phone in it, so nobody is looking at a game that doesn't exist. Seven achievement images in the game's own objects. A store page in English next to the Turkish one.
Then I sent Apple the nomination. What they do with it is up to them.
What the number was made of
The 22 days were not a lie. They were the price of a careful team, a queue, and a lot of things going wrong. I work alone, I was at the desk until midnight, and almost nothing went wrong.
If I had believed the number, version one would have launched without any of this, and the nomination would have described a plainer game. That is the part that stays with me. I nearly didn't ask.
This morning the Mac on my desk builds the iPhone package by itself. The queue is gone. The next estimate should be shorter.
— Kemal