Baba is You has garnered a lot of attention.What do you think is the recipe for its appeal?The dichotomy between cute, minimal looks and fiendish mechanics?Maybe the crazy combinations one can make, while keeping the game very accessible?

A design aspect that I usually try to strive for when making games is surprising/amusing the player; I think Baba’s concept lands in a lucky spot where the mechanics provide some interesting situations without the basic system seeming too complicated to players, so there are several low-bar sources of surprise/amuse that catch people’s interest. For example, I’ve noted that showcasing “Baba Is You” turning into, say, “Wall Is You” and have the player controlling the walls seems cool to a lot of people, while being very simple in terms of in-game implementation.

I guess this means that in my eyes the game’s main “meat”, (i.e. most of the puzzles), might not be what players’ll be interested in the most.The game gets really tough and while I’ve tried to offer amusing interactions throughout the game (and I think I’ve succeeded ^^), the difficulty curve and the effort required to see some of those will probably turn many non-puzzle lovers people away. A tester suggested giving the players the ability to “opt out” of the rest of the game by offering an ending quite early, and I liked this idea a lot because it meant that people not into super-tough puzzles could play the game and appreciate the concept without having to worry about the difficulty curve as much.

Some people have commented that they like the artstyle, but since the game’s aesthetic is still quite minimalist, I’d imagine that it’ll be a somewhat divisive aspect.

Did you ever feel like your game, as you kept enhancing it for a full version, was becoming a kind of Frankenstein’s creature?I’ve seen a lot of crazy GIFs from your twitter account, where Baba’s behaving, well, not quite you envisioned ^^

There was definitely some feature creep in the mix when designing the rules, but I feel that the game managed to mostly avoid becoming too chaotic, one factor being that the most complicated words do so much on their own that they usually benefit from appearing alone. Therefore I could design areas with a specific new rule/word in the spotlight, with the more basic words acting as a supporting cast. This meant that I ended up having to have quite a few levels to teach those more basic “supporting” words first, though.

Even with the words that are in the game right now, the whole rule system had to be rewritten multiple times to make the game logic consistent(-ish). On top of this, I have multiple ideas for words/puzzles that didn’t make it in for various reasons, and during the development I had to cut a couple existing words for being too confusing or not interesting enough (e.g. “Back”, a word that made objects move backwards in time).

Some of the words that did make it into the full game had to be adjusted/simplified over the development for similar reasons!To give a fairly confusing example: “Baba Is Push” makes Baba pushable (and in turn push other pushable objects when pushed). “Baba Is Shift” makes Baba act like a conveyor belt, moving objects on top of it. Since the objects on a conveyor-Baba weren’t technically “pushed” rather than “moved”, it made sense at one point that they wouldn’t be able to push other objects, and consequently not care about walls, either. As a result there was a level about moving through a wall with the aid of conveyor belts. Tester feedback helped me realize pretty soon afterwards that this wasn’t a very good logic to pursue.

How did you go from jam build to full game version?Was your initial game engine open enough for all those additions, or did you basically have to redo the whole game?

I‘ve built’ the full version on top of the jam build, but very very little has stayed unchanged between then and now. As mentioned earlier, I’ve had to re-implement some of the most fundamental systems several times, and changing the graphical style has meant that things like the game’s resolution, visual code and so on have had to be changed as well.

I use a game-creation tool called Multimedia Fusion 2 by Clickteam along with the language lua, and MMF’s relatively easy workflow combined with how easy lua is to approach have made the process of rebuilding everything mostly a fairly painless experience. The jam version supported only square-shaped levels and all the objects were hardcoded. Breaking these limitations were some of the larger undertakings during the progress from the jam version to a finished Baba!