Building an Infinifactory: How to turn science into a successful game

At the end of Janurary Zachtronics, headed up and partially named after Zach Barth, released Infinifactory on Steam Early Access. Bucking the trend, it was almost entirely complete: A release that had no visible gaps in content, or major bugs. It was essentially a release version, soft-launched through Early Access. 

Gamasutra talked to Barth back in October about his motivations for this approach, and the benefits of having people play a game that isn't considered finished through Early Access; it appears to have worked out exactly as he wanted. 

With its swarm of conveyor belts, pistons, and welders, Infinifactory extrapolates the ideas that Barth established with SpaceChem, translating them into 3D which understandably increases the complexity of the puzzles Zachtronics has created. 

I talked to Barth about how things are going with Early Access, as well as why the small indie studio decided to invest in a fully voiced story in an abstract puzzle game, and what it's like to have Jon Blow and Notch publicly playing your game. 

It seems your motivation for doing Early Access was slightly different to that of most developers, in that you wanted to sidestep people judging the game before you’d had a chance to adjust it to the public. 

Premature reviews, certainly. The primary reason, though, is that it’s hard to telegraph to players that, "Yes this game is out, but the scope of content and what you see in the game isn’t final, and it might change, especially early on." Letting players know that we are listening to them and want to build the game around their reaction. 

And you’ve got Steam workshop support with Infinifactory, right? How’s that been?

It’s been pretty good. So… kind of like SpaceChem, we have this quirk where people think our games are hard with the puzzles we make, but the user-created levels are always so much harder, and already that’s the case.

Ours are designed to be very open ended, and not about "tricks," necessarily, but just building different kinds of things which require thinking about how you’re going to construct things. A lot of the puzzles in the workshop are: "Here’s something we managed to do that’s really strange, see if you can figure it out too!" So it’s a totally different kind of puzzle, and I often have no idea how to complete them. 

It seems that often with SpaceChem and Infinifactory there’s a wide spectrum of success. You can create a really clumsy and convoluted solution that just barely works, or you can make a really tight and elegant solution. Is that always your intent?

Absolutely. That’s the point of the whole game. Any puzzle game can create a solution, tear it apart and then get you to put it back together, but with our puzzles they’re so unlike traditional puzzles that they’re more like a task -- a puzzling task -- than a puzzle. 

There have been a lot of games that have used a coding-like interface to communicate hacking, or interacting with computers, but the product is often not much like coding. Whereas with SpaceChem and Infinfactory it seems much more like you’re trying to convert coding into something more understandable and less conceptual.

So SpaceChem had a slightly more direct comparison to computer architecture, as it was based on a sort of very convoluted computer architecture that doesn’t exist. This is much more directly distilled from factories, and industrial engineering. It’s more engineering than programming.

There are a lot of "programming" games around there, with air quotes around that, and they sort of superficially are about programming, or their interface is about programming, but I don’t think any of them communicate what it’s really like to do programming, especially the problem-solving aspect of being a programmer. I feel like that’s the really amazing thing about SpaceChem and Infinifactory, that other games, even if they’re about programming, can’t do. 

And was that your intent when you started making these games, to communicate to people who don’t or can’t program or code, what it’s like to do these things?

Probably not, because the people who play these games are probably good at this kind of thing, or are already on that track. It really was an earnest expression of the fact that I really like this kind of problem-solving, and I like engineering systems -- I have a background in engineering -- and it just seemed like a fun thing to make. 

The reaction in the game development scene to Infinifactory has been very positive; I’ve seen Jon Blow and Notch both talking about playing it, with the former streaming some of his sessions. I guess that’s a kind of vindication, where these prominent developers are enjoying your game. 

When I go to game dev meetups, I find that a lot of people haven’t heard of SpaceChem, which isn’t surprising, and they ask “Oh, what’s SpaceChem?” and the first thing I ask them is “Are you a programmer?” and if they say "yeah," I’ll tell them to just go play the game.

[Ed. note: If you're curious about SpaceChem, you can also read Barth's postmortem of it.]

I think the thing with these games is that they really just click -- and it’s not that they’re not for other people -- but it’s that they really do click with programmers, and people who do technical problem-solving as their thing in life, or one of the things that they do. You’ve just named two people who have long programming backgrounds. They’re both programmers who like games, and this is straight up the game for them. 

Have you seen any direct positive response to Infinifactory’s sales in the wake of their attention?

I’m sure it’s not hurting. There hasn’t been a noticeable effect, but the biggest spike was releasing the game. But I’m sure there’s been some sales from it. 

In general, it hasn’t seemed like you’ve aggressively marketed the game, and I know when you talked to Mike, you said you’d barely put out even more than a few screenshots. I know you’ve put out a trailer now…