How ipe makes games

This is the first post in the journal on play.ipe.rest, so it starts with the question underneath all the others: how do we decide what a game from ipe() should be?

The short answer is that a game should sit down inside your break instead of taking it over. ipe is short for idle process, the state a processor settles into when there is nothing left to do, and the "rest" in ipe.rest comes from there. That does not mean our games are quiet games. Strategy and duels still make you think. It means the game respects the time it lives in. Five principles come out of that, and we hold all four of our games to them.

None of the four is released yet. We will say that again below, because it is one of the principles.

Short rounds, few rules

A round should end in minutes, and closing the app should never cost you. Iltonggi, a short, fast, classic territory-conquest strategy game set in Korea's Three Kingdoms era, takes 5 to 10 minutes a round, and closing the app saves your game automatically. Inkisle, where you send fleets between islands and tint the sea your color, takes 3 to 5 minutes, played with one thumb in portrait. Claimball is built around short matches, and we are not quoting a time before playtesting. Sunmeld has no time on its site, and we will not invent one. Its round is over when the shared tub overflows.

If the rules don't fit on one page, the game is asking too much. Sunmeld is four lines: take turns dropping stones, matching stages merge, whoever merges scores, and if the tub overflows the player who placed the last stone loses. Claimball is a notebook grid where the tile your ball bounces on becomes your land, and a smash with the center of your paddle claims three tiles at once. Inkisle's rules are that islands make troops and a larger force arriving takes the island. Iltonggi is the deepest of the four, but its depth comes from what you cannot see: land you have not visited stays in the fog until you send a spy.

Short rules do not mean shallow games. Reading the board, and reading the other side, is where the thinking goes.

A quiet app

No streak nagging, no ads, no accounts, no tracking. This is the principle you can check most easily, so here is what each game card on play.ipe.rest says. No game is released yet, so the terms below are plans and may change.

Iltonggi collects no personal information, has no ads or tracking, and its core features work offline. Nothing nags you to log in every day. Inkisle needs no account and no network, and has no ads. Its pricing is not final before release, so we are not quoting one here. It also has optional help items, and we would rather say so than leave it out. Sunmeld is planned as free to try, with one purchase unlocking everything, and no ads. Claimball puts it in one line: planned as free to try, one purchase unlocks everything, no ads, no subscription.

We have not put prices in this post because the stores are not open yet.

Saying what is true

If it's not released, we say so. If we haven't checked, we say that too. Iltonggi and Claimball are marked "coming soon", and Inkisle and Sunmeld are in development. No game has a store page or a release date. Figures that have not been through playtesting, we leave out rather than guess.

Every game keeps a public making-of, and that is where we write down what is unfinished or unchecked. On play.ipe.rest each game card also lists who it is not for: Iltonggi is turn-based, so it is not for fast reflexes; Inkisle flows in real time, so it is not for slow turn-based play; Claimball has you returning the ball in real time, so it is not for anyone who finds busy-handed games a burden; and none of the four is played with friends yet. Every game here is solo or against an AI. That is a limit of the lineup today, and we would rather you read it before you download anything.

The same honesty applies to us. ipe() is an AI-native studio, and we say openly where we use AI.

An opponent that plays fair

Comebacks come from the rules, not from a secretly weakened opponent. This is a principle we hold all four games to: winning and losing should come from rules you can read, not from an AI that was quietly adjusted behind the scenes. In Inkisle, the way a clearly won round ends early is a rule you can read, not a hidden tweak.

This is a promise about the opponent. It is not a promise that nothing else in a game can affect how you play. Inkisle has optional help items, and we are not going to pretend otherwise.

What this journal is for

Principle by principle, that is how ipe() games are meant to work. Whether they work is something you can only judge by playing them, and we will keep writing about how they are made, what went wrong, and what we have not checked yet. Other posts here will look at specific questions, such as why a turn-based game suits a break.

// play(); rest();

 Back to the journal

Last updated: