Finally, on 27th July, Blow noted the “whole industry is having a hard time” and then, when asked how many of his development team are working on the compiler for programming language Jai, Blow replied: “None, because we can’t afford to pay anyone because the sales are bad.”
Oof. I don’t like Jon Blow, but I do like Braid. Sad to see nobody really cares about it anymore.
I believe it was a desperate attempt to get a new source of revenue. His upcoming Sokoban game is taking forever to make, so it’s not going to bring them any new revenue anytime soon. In large part because he made the arcane decision to create a new programming language for it (as a replacement for C++), because apparently Sokoban is the type of game where you really need that high performance.
Sometimes writing the game engine is just more fun than making the game itself, ok…
Yes that can be true, but fun doesn’t pay the bills.
Yes, but read that again, he’s making a new language, not a new engine… To put it in terms of food, using things like Unity is equivalent to eating industrialized food, you have absolutely no control and you get what you get; Using other engines like Unreal or Godot that have open source is like cooking at home, some work but you can get it just the way you like; Building an engine yourself is like having a little farm in your backyard and doing everything from start to finish, it’s slow, you’ll face problems that have nothing to do with cooking that were handled by the farmers before and at the end you’ll get something only slightly better than what you could using store bought products; Building a language from scratch is the personification of the saying “to make an apple pie from scratch first you have to invent the universe”.
And you know the worst part? It won’t be any faster or better in any mensurable way, large groups of developers spend decades to develop the languages we have today.
I think his use case is that the new language allows for more rapid iteration in development. Years ago now, I saw his demo of the language, and it compiled so quickly that it may as well have been done by the time he pressed Enter. For all the gains he got from that, it still hasn’t helped him release a game by now, but I do see the problem he’s trying to solve, and I do think it’s worth solving.
Faster compilation is probably nice, but making a new language with all its tooling from scratch is a huge endeavor. Props to him for actually doing it.
The problem is that all this work takes away time from the actual game development. I’m not sure about the scope of his next game, but from what I’ve seen I don’t really understand why his Sokoban adventure game can’t be made in Unity. I don’t think he’s pushing any hardware limits with it.
Unity also got hot reloading nowadays, which is about as fast iteration you can get.
I’m just armchair guessing, but I believe he would’ve been done with his game by now if he just used Unity.
Hollow Knight: Silksong is a Unity game, last I checked, and it’s not getting done any faster. As per The Witness, it’s probably far more about how he’s retooling puzzles rather than his language, if I had to guess. Plus, it’s not just iterating within the editor; this thing exported a build in well under a second. I worked on a Unity game a few years ago, and it definitely took me far longer than that. It even had a bug for a bit there where we couldn’t see the game when run via the editor on Linux, so the only way we could test it was by exporting a build until we got an update to Unity.
Fair point.
Time will tell when Sokoban and Silksong releases. It’s hard to know what’s happening internally at the studios and why it’s taking so long.
Making an entire programming language is a bold move, and I’m skeptical it’s a move that’s going to pay off.