Craftomation 101: I spent an hour debugging a robot that refused to fetch wood, and I loved it

My first robot froze to death three metres from a campfire because I had swapped two blocks in its routine.

I should add that I lit the fire. For him. And that he died going to fetch the wood that was supposed to feed it.

There, you now know roughly everything about Craftomation 101.

The principle

You arrive on a frozen planet with a small rocket and a hungry robot. The eventual objective is to terraform the place.

The studio puts the job at nine thousand years, which is honest, and immediately proposes the solution: have robots do it for you.

So you program your CraftoMates by connecting visual blocks. Harvest, combine, deliver, in the order you decide. Then you watch.

And that is where the game gets interesting, because watching a robot execute your plan is the best debugging tool ever invented. You see exactly when it does something stupid, and you understand immediately that you wrote the stupid thing.

This is not a factory game in disguise

Be careful not to mistake the genre, because the confusion is easy.

In a conventional automation game you place a machine that always does the same thing. Here, you write the behaviour. That is not the same skill and it is not the same satisfaction.

The game takes you quite far too: reusable functions, variables, arrays, coordinating a growing workforce. The studio phrases it well, a useful program becomes the foundation for your next machine.

In other words you are not only building a base, you are building a library.

And one constraint gives all this its flavour: your robots need fuel, and your growing base needs heat. So you have to automate not only production but the survival of those producing. A robot that stops for lack of energy is not a calculation error, it is negligence.

The full loop

You harvest, you combine materials, you discover recipes, and you turn a frozen landing site into an expanding base.

Your robots are assembled and upgraded from what they find around them, which produces the game's most satisfying loop: programming a robot to harvest the materials that will build more robots, which will harvest more.

At some point your workforce feeds itself, reproduces and works without you. That is the goal, and it is also slightly unsettling.

Finally you plant trees, produce oxygen, and work towards terraforming the whole planet.

Replayability

The frozen planets are procedurally generated, with different layouts of resources, buildings and volcanoes.

The pleasing detail: you can share a seed with another player. Which means you can set each other a challenge on exactly the same planet and compare solutions. For a programming game that is the best possible social feature, and it cost almost nothing to implement.

What it is worth

Version 1.0 launched yesterday at 10.49 euros, and the game shows 523 positive reviews out of 621, roughly 84 per cent.

Luden.io is not new at this: it is the studio behind while True: learn(), which already managed to make machine learning fun by pretending you were training a cat. They know exactly what they are doing.

Two honest reservations.

This is not a relaxing game. It looks cozy, with its little yellow robots and its snow, and it will ask you to understand what an array is. If the mere mention of the word variable tightens your jaw, the cute layer will not save you.

And the opening is laborious, as in every programming game, while you learn the vocabulary of the blocks.

But if you have ever felt the precise pleasure of watching a machine do exactly what you told it to, this one is for you. We have written a guide to getting started, because the first three hours deserve a little hand-holding.

Your worst infinite loop? We collect disasters on the InsertCoins Discord.