What programming language would be the best to learn to develop a game from scratch?

If you are new you probably want to decide which game engine you want to use first, then learn one of the languages that engine supports.

Most popular choices would be: Unreal engine, Unity and Godot. Hovewer if your goal is to get into game dev industry by far the most popular language would be C++ followed by C#.

Depending on the scope of your game(if it is a very simple 2d game like 2048) you might want to consider Python too. it is known for being one of the easiest languages to learn and you will have easy time researching the points you get stuck on.

One of the most important things to learn about programming is not to make things from scratch unless you have a very good reason to, imo. Programmers build off of the things that other programmers make, for example using some code library to achieve a specific task or building a game on top of one of the many solid engines that are out there. (Godot, Unity, Unreal, etc.)

Writing software is hard, and making a fun and functioning game is even harder, so unless you have some very specific reason to write a game "from scratch", I would just get something like Godot or Unreal and start learning the ins and outs of game development and design.

Why Godot and Unreal over Unity? Because having source code access means that, if you at some point want to learn more about game engine development specifically, you can start by browsing and hacking away at the source code of an existing engine. So, do yourself a big favor and just start learning how to make games using an existing engine.

So! If you don't have much experience in programming, you DO NOT want to write your own engine. Period.

"I wrote my own game engine" is the kind of thing you'll see masters/doctorate computer science students (or crazy industry veterans) do. While it may be possible to write a simple text-based game that only uses the command line, it gets complicated fast.

There are some libraries out there like PyGame which let you set up "toy" games quickly (in Python), but no shipping game is going to be built entirely in PyGame.

When you're out applying for jobs in the industry, having a studio you're applying for say "We built our own engine" in 2023 is a red flag. There are multiple battle-ready game engines that have made thousands of games. Most places want to build games in either the Unreal Engine (C++) or the Unity Engine (C#). There is a third one I should mention - Godot - which is a flexible FOSS game engine. But most places use Unity or Unreal.

There is so much that goes into making a game engine. Not only are you making a game, you're making a tool that lets you make a game. You're making stuff that can read model and animation data. You're making something that can handle a bunch of different input methods. You're making something which needs to calculate lighting and collision, parse images, run scripts, save and load data, multiplayer games need a full networking model with local prediction, correction, and latency mitigation, etc.

By definition, making your own engine is untested. You are going to run into issues, whether you have 1 person or 1000. What starts off simple quickly balloons as you want to do more than just show white text on a black screen. Something like Unity has had a bunch of production games (like Hearthstone) use it and find all the issues already so you don't have to. There is literally zero reason to make your own engine today.

I myself work at a AAA game studio, as a programmer. I've worked on the Battlefield series in the past, although it's not what I work on now.

Let me give you the advice I wish I had 15 years ago, when I was starting out: think small. It is far better to have made 10 projects in 1 year than 1 project in 10 years. The only way to "make it" as an indie dev is to be incredibly talented, incredibly lucky, and have an incredible amount of funding. Even supposed "one-man teams" like Toby Fox had help making their games; it is very difficult to make a game with 100 people working on it, let alone 1.

Make small toy projects that you can do in a weekend. Drop it if you spend more than 2 weeks on it. Don't be like me where I spent years working on a dream project that I never got in a good spot to show to anyone. When I talk to people now, when I talk to interviewers or coworkers, I don't really mention my white whale of a dream project I never finished. I mention the little games I made for gamejams, the ideas I had and how I played around with them.

It is so much more impressive to show an interviewer an active GitHub and a bunch of free games you've put on itch.io. I've literally gotten jobs because of it, but it took me years to realize I was doing the wrong thing and needed to pivot.

With that rant out of the way. C++ is industry standard. Any programmer will need to know C++ inside and out. Even if you don't work in it directly, you're almost guaranteed to be working with something that works in C++. But C++ is a hard language to learn.

If you have taken a programming class already, I'd recommend Unity. Unity isn't as common as Unreal, but C# is easy to learn and somewhat similar to C++ (not that similar, but a lot can carry over). It is code, though, so you need to know syntax.

If you've never taken a programming class before and you're self-taught, then I'd actually recommend Unreal. Unreal has "blueprints", which is a visual scripting language. This means you don't need to know the syntax of what you want to do; you just grab nodes and connect them together. It's very easy to understand and intuitive, and it helps you build the foundation you'd use if you ever delve into the code side. You can make a whole game in blueprint, without touching code - the game won't be huge and mega-performant, but it'll be relatively easy to make and doable by a single person working on a very small project.

Bear in mind that there are other disciplines in game development other than programming as well. That's sort of the best part about making your own stuff - you have to learn to do everything, from art to design to programming. Designers typically aren't expected to know much about code, but they are expected to be creative, collaborative, and intuitively know what makes something a fun game to play. If you find out that programming isn't for you but you still really want to get into game development, making all these tiny projects is a great way to exercise your design muscles as well.

Back in the olden days I learned programming and made games with BlitzBasic and PureBasic. A simple programming language, but with included functions to easily display sprites on a screen or play a sound. Everything else you had to do yourself, so it wasn't a game engine, but you still could have results quickly without too much boilerplate code. It was great for learning, together with the community around it.

• PyGame seems similar. And you're learning a useful language to boot.
• PICO-8 is a little virtual console with everything built-in to write tiny pixel games in Lua.
• Plain JavaScript with the Canvas element. You can do everything yourself, only the HTML/CSS stuff around the Canvas makes it a bit more cumbersome. But you can always write your own functions to easily draw a sprite etc.
• JavaScript with PixiJS. Higher level Canvas functions (demo). Looks like fun.

Or you could start with the Godot engine. That way you could eventually make bigger games. But for learning game-making and especially programming from scratch it's too high level. Better learn the basics first, write a lil' Tetris, Game of Life, Screensaver etc. before using complicated huge tools.

Building a game from scratch sounds nice on paper, but if you are also trying to learn programming I'd suggest smaller steps.

Start by looking at some really simple games (think Tron, or snake) and figure out how they work, adapt them, break the code, fix the code, change the rules of the game and make it work differently...

Once you are confident changing code then try writing stuff...
Much as I'd love to recommend a language from 20 years ago (e.g. Pascal) realistically you want to be looking at something like Python as an entry language these days.

Depends on what you want. I've been liking Godot, but I'm an "Open Source" person. There's definitely more of a community around Unity or Unreal.

But Godot is free in both ways and relatively user friendly, and since you're uninterested in hiring a hundred people, using a tool that you like is fine, even if it's not the most popular.

There's a course I've never used called Learn GDscript which teaches the inbuilt language for Godot (GDScript) in the browser with fun interactive tasks. It looks neat, but I've never tried it myself. You can use other languages with Godot, but I recommend the GDScript. It's very similar to Python and is well integrated into the engine.

So from there it's about screwing around! Like other people have said, you're not going to whip up the game that's in your head in anything like the time frame you probably think. Even if you think you're being realistic, it's probably even worse than that. But I don't say this to discourage you, I say this to prevent you from discouraging yourself!

If you can get a game where a green circle goes through a maze and then text shows up on the screen that says "you did it", that should be viewed as an accomplishment! It's simple, sure, but it's something you did. Try to break your game's features up into micro chunks that are playable. It's easy to spend 6 months working on something and making progress, but not in any way you can show friends or whoever, and can't even really "play" yourself. That can be demotivating. Try as much as you can to have something playable as often as possible. It will feel much more like real progress if you constantly have something you can demo.

And also don't underestimate how much a bit of art and sound effects can change an experience. Silent 2D boxes is fine to test things out, but even a free art and sound effects pack makes a huge difference in how fun a game can feel. It can make even a simple premise suddenly feel like a game.

Good luck, have fun! Oh, also once you're done tripping over your feet, maybe try a game jam! They're good exercise.

If you want a serious answer, C#. It's a general purpose language with a great ecosystem in both the regular Dotnet Stack and stuff developed specifically for unity.

Ada

Sorry just a little programmer humor

I would suggest that you try something different. Do a little tutorial game in the Godot engine and learn all the things that you would need to implement yourself if you wouldn't have the Godot engine doing the heavy lifting for you. This might help you to get a feeling for the scope of your endeavor. I think it's hard enough to build a game where you don't have to implement sprite movement, update ticks, physics, collision detection, etc yourself.

If you want to do it anyway because it's awesome, there's a running joke that there are like 50 game engines currently being developed in Rust. Go all in!

Not a developer but one time I decided to make my own game and I got one character sprite done for walking to the left and it took me 18 hours.

My best friend is a senior developer and one of the brightest people I know. I’ve seen him code for me some arduino sketches that are absolutely next level (even after spending about 400 hours getting decent at python).

He’s currently 3 years into developing his game and has a few more years to go.

I’m not saying it’s impossible I’m just gonna say that maybe you should start small and see what the time commitment is and such and make sure you enjoy it enough to buckle up for a few years, overcome frustrations and are able to see it through.

If it’s your absolute dream it is entirely possible, just don’t let the length of the road to get there be your deterrent. One step at a time and then onto the next problem or step.

I'd highly recommend you start with a game engine like Godot. Very few studios, let alone individuals, code games from the ground up. Besides having lots of functions and libraries specifically designed to do the math required for games, these engines will have lots of tools that will be useful for performing all kinds of game-related work.

If you're looking for a true challenge and really hate yourself, and you really want to go from scratch, you could write your own engine in C or C++. Again, this is a Bad Idea™.