At least that's a goal, a lot of people say "I want to make a game" but get stuck in the making an engine part. Which is also valid, don't get me wrong - if you enjoy it, by all means! However it's not building a game, it's building a game engine.

Anyone that wants to build an engine (or a CMS, or whatever) with the goal of making a game should try making one or more games with one or more existing engines first, if only to get a better idea of features and shortcomings. And if they really want to make games, or just the engine.

Literally that what I'm pursuing rn, I want to make my first (real) 3d game, I'm very aware of the loophole that most self-made engine game developers falls into but I genuinely found no 3d engine that suits my criteria:

* Be resources efficient (memory and cpu usage)

* Has primary 3D support

* stable

* Supports old devices (ideally embedded systems too)

* And of course be FOSS

I'm focusing on mobile, as you know mobile devices are very sensible to heating so the 1st option is a must, Godot isn't resource efficient, it heats very easily when I play one of the sample games. Regarding the second option, whilst Godot still supports gles2(which is the most widely supported api yet across old devices) it's further being pushed as second-citizent over Vulkan and looking through github no other engine checks my boxes...

If someone could direct me to an engine that checks all the boxes above I would happily try it out.

Bevy?

Problem: I want an engine that does x, y and z but none of the 4 can do that

Solution: make a new engine

Result: I want an engine that does x, y and z but none of the 5 can do that

it's more like 400. I'd be surprised if you couldn't find any engine that couldn't do XYZ.

but hey, your engine doesn't need to be a competeing "standard" if you do want to make a new engine.

I've been looking into Stride[0] for a bit and it seems to meet most of your requirements. I'm not sure at all about embedded support but it does seem to treat Linux as a first class citizen for once.

the main up/downside is that it's a C# engine. So it won't necessarily be as hyper-optimal as c++ without some discipline on what parts of the language you use (hint: avoid LINQ at all costs in the hot loop, and look at the newer ways in C# on how to utilize unmanaged collections).

0: https://www.stride3d.net/

No engine. Just write a game,

The only thing more impressive than writing a game engine is actually using it

It's definitely possible to do it - for me the key is to write the game and the engine together, and only implement the bare features you need. For the first game at least, the engine and the game are essentially the same thing. Also, forget about wiring an editor, just add debug functionality to the game.

That is truly a curse. I just want to make an engine so I can say I did everything from scratch

I hope by 'everything' that includes the text editor and compiler you'll used in the development process ;)

That's literally what I fall into, making an engine then falls deeper into making programming language and editors, it's an endless hole.

Buddy if I retire early I will ;)

I think a lot of them want to be like Carmack. :)

Or Commander Keen