File: 0190763a6365e89b062c8502b(...).png (447 KB, 701x1088)

447 KB PNG

>>103010955
first for abby chan the hotest loli

>>103008673
That's great anon. DId you wow them with a nice portfolio? Currently, I'm working in education, but I think it would be neat to pivot into graphics programming. Not sure if I'm qualified though but I think its something I could learn how to do if I study more.

>>103010983
more math that is

File: 1711651243546.png (211 KB, 1563x1391)

211 KB PNG

>>103011028
I like math so I wouldn't mind it. But I've heard it is a hard career to break into, so I wouldn't mind trying something more "entry level" friendly.

File: Swimsuit Sophie.png (126 KB, 300x425)

126 KB PNG

Does anyone have experience with the 3d id tech games and how good their physic engines are? I've been looking at Quadrilateral Cowboy, Quake 1 and 2 and Doom 3 and I'm wondering how good some of their components are and if they can work well with modern physics engines if they're garbage, especially the physics-related portions of the engines.

>>103011081
I dont understand your question

>>103011390
I'm just wondering how good the 3D id tech engines are in terms of their physics. I'm also asking if it's possible to throw a modern physics API at them or if that's even worth it.

>>103011081
Integrating a modern physics API, like PhysX or Bullet, into these older engines is technically possible but would require extensive modification.

>>103011445
Quake 1-3 have simple bounding box collisions
Doom 3 has a rigid body physics system
It's possible to integrate a rigid body physics system in the old Quake games but I don't know why you would

>>103011081
What about the physics would you want to change in id tech? The physics were good enough for AAA of the day, and I can't imagine trying to rip out the physics of id tech to replace it with something new. That seems like a large under taking. I think you'd be better off integrating your preferred physics engine into 3D rendering engine of your choice at that point.

>>103011465
I imagined so. It's an old ass engine made before physics engines were standardized. I don't think I could do that and make a game at the same time unless I have more people on my one man team. Thanks.

>>103011496
>>103011500

I'm trying to find a physics engine to implement in an action-adventure game where I want a lot of cool explosions and a large scale destruction ala Just Cause. I'm using raylib for audio-visual part of the right now but I'm trying to find a physics engine that's easy to work with and can do the job. I'm leaning towards PhysX, but the installation process isn't clear.I guess I just though about the Doom3 engine out of desperation to get something off the ground.

File: lll.gif (2.75 MB, 640x640)

2.75 MB GIF

>>103010955
thinking about making simple games in javascript to complement my webdev learning

what would be a better environment
kaplay or phaser? does anyone have any experience with either?

or should i just ignore that and move to something like love2d

>>103011588
The Doom3 engine is not a solution for your problem

>>103011588
It sounds like you're just beginning so I'd recommend two paths to you, either pick up Unity and make your game or stop trying to make a game for now and just learn the basics of low level game programming by using raylib and other libraries without a particular goal in mind
You're going to be pretty lost with a goal like that to start with

File: Screenshot_20241026-023101-353.png (463 KB, 493x636)

463 KB PNG

>>103011596
Phaser is nice, but try making games with pure JS first.

File: GO BURNICE GO.gif (200 KB, 510x510)

200 KB GIF

>>103011629
I've already made simple 2D projects (not full games but robust sample sized programs related to mouse events and drawing polygons and circles on the screen), I'm more inclined to ride it out with raylib to learn it in a more robust fashion.

Thanks for the advice, mate.

>>103010955
you beat me again you tripcode freak

>>103010955
Based gedg

Alright /g/ays, I need something to build right fucking now. I'm so fucking bored.

>>103011873
MMORPG capable of hosting 10,000+ players on the same server.

>>103011883
Uh, that's quite a bit above my pay grade.

>>103011895
Okay, a game where you guess heads or tails.

>>103011926
>award the player for guessing correctly by giving them a crate that contains a skin for the coin
>rig the game so your guesses always have a super low success rate
>sell crates in the game's shop
>make crates require a key to open
>sell keys in the game's shop
>only allow the player to flip the coin so many times a day
>lock flipping the coin additional times behind an in game currency that can be purchased through the shop
>add a battle pass that gives the player uselessly small rewards for each "correct" guess
>add a better battle pass that gives the player slightly better rewards, but is locked behind a subscription
This could make millions.

>>103011686
shit i dunno why i didn't even consider that
yeah i'll do pure js first

>>103010955
I'm trying to prototype a bullet hell game, how do they generally handle collision for thousands of bullets? and how do they handle the map?
I have two approaches for collision, the first one is to split the screen in half, upper half that will never hit the player won't even have collision. The second approach is to model the bullet as a line, all bullets on the same line will only test for collision once.
As for the map I don't even know how do they handle long maps. It seems that actually doing a long map would be wasteful, so maybe unique sections with some inbetween sections that loop until enemies are cleared?

>>103014087
2D collision detection isn't THAT expensive, assuming it's implemented in a non-stupid way. Some kind of data structure like a quadtree (this is not the only viable one) will make it so you only need to test any object against things that are relatively close to it.
Plus you're unlikely to want to test bullets against each other, so 1000 bullets doesn't explode into 1,000,000 checks.

>>103011081
They are relatively simple and easy to understand, the weaknesses are usually in rigid body stuff and resolving collisions (though even modern physics engines have weird edge cases). If your going for some glitchy retro physics (for skiing and whatnot) stuff like that is a good reference.

>>103010955
Nice to see a webm of 3D dwarf fortress dev. May God bless this anon's autism.

>still cannot figure out how to manually set the color of vertices in code in godot
Like I have the ArrayMesh.ARRAY_COLOR array, now what? How to set the material?
Last thread died after I've asked this question, I hope this one won't

File: 1727972429930188.jpg (9 KB, 183x275)

9 KB JPG

>>103015084
>now what?
Learn OpenGL and become one of us!
ONE OF US!

>>103015084
>now what?
learn rust

>>103015084
Not a clue, never used godot. I know how to do it myself though. Easy enough to have another 3-float vector as part of your vertex data that you send to the shader.

>>103015192
>become one of us
i have finished games, no thanks

>popular youtuber was on /gedg/ and I didn't even know it
https://youtu.be/RKjcx-8R86c?si=3UbVILPNpueqfXfK

>>103016713
another fool tricked into using Godot

>>103016713
never seen him on gedg but I might have seen him on agdg. I mean he reference agdg in his vid, so...
also god damn he was blessed by the algorithm.

I turn into such a nuclear brainlet when I'm working with very nested statements

>>103011895
It most likely isn't, depending on how you define MMORPG

>>103017403
He submitted to one of our progress days from last year

https://covalentbonds.itch.io/robot-game-early-prototype

I fucking hate naming things
I have 6 different compiler passes that do various things and thinking for appropriate names for them has had me stuck for weeks

>>103017841
>pass1
>pass2
>pass3
>niggerfaggot
>pass5
>pass6
Done, go do something else and rename when inspiration hits.

>>103017841
I stopped caring at a certain point.
I'll write down a variable name and I'll think
>wow that's a fucking awful name. If anyone sees this they'll think less of me
and then I just continue on, because fuck it

>>103016734
What should he be doing?

>>103017634
Oh ok I totally missed it by 1 progress day.
>t. started submitting to progress day 3

>>103018103
This. I used sit there for minutes trying to come up with good names, now I just name them whatever and just get it over with.

long live GEDG.
GEDG > AGDGGERS

File: no engine.png (667 KB, 900x506)

667 KB PNG

>>103018880
uhhhhh BASED my good sir
shitposting /vg/ nodevs in SHAMBLES

>>103018958
I remember quiting this anime after like the 3rd episode because it was bullshit

File: nene mad.png (394 KB, 856x481)

394 KB PNG

>>103019034
I fast forwarded to the parts with Nene or the programming lead, the rest is boring

>>103018185
Using Unity

>>103018103
If you don't name things then you end up not knowing what your code does

I will continue to ignore their existence

>>103019069
The most bs part was that MC got a gamedev job at a company straight out of highschool
>Verification not required

>>103018880
the /v/ gamedev thread is trash too, it's just a shilling thread

File: 2024-10-29 12-50-53.webm (3.63 MB, 1280x720)

3.63 MB WEBM

not quite right but at least there is some actual acceleration now

File: 1729187412861885.png (210 KB, 369x573)

210 KB PNG

Let's say I want to make an engine for pokemon fangames in C/C++ like how that one anon was making a SMB1-style engine a while ago. What should I do?
What graphical style(s) should I support?
>Gen 1/2
>Gen 3
>Gen 4/5
>3D
>gen 1/2 but tiles and sprites are rendered in 3D

>>103019355
sugoi driftu
>>103019360
use raylib. Choose one graphical style and only implement that first. Make your own game rather than guessing what would be useful for a game

File: 2024-10-29 13-08-23.webm (3.82 MB, 1280x720)

3.82 MB WEBM

ignore the messed up tunnel windows, I know how to fix it, just need to draw transparent things last

>>103010955
what is better JS/TJ or C/C++ for a simple 2d VN game engine?

I know JS. front end dev

>>103019503
GLFW + OpenGL

>>103019503
HTML + JS

>>103019503
you can do even full 3d games with js, it's fine for that application, and also much more portable than c++

>>103019494
>148fps for approximately 4 triangles dangling on the screen
it's time to optimize

So I discovered that I should make a thread dedicated to vkQueueSubmit.

Any other Vulkan devs discover this?

Thank you for putting my game in the OP.
God Bless.

>>103019503
>2d VN game engine
Honestly? Lua + CPU Rendering. If you need cool VFX, then switch to OpenGL ES. No thirdparty libraries. Trust me, it'd be fun.

>>103019503
If you already know something, and what you're comfortable with is capable of doing what you want to do, then it's foolish to learn something new to end up doing the same thing
Notch would've never finished minecraft if he had listened to retards telling him not to use Java (what he knew and was comfortable with)

A visual novel is just pictures and dialogue, hell as much as I hate webshite, CSS is pretty damn good and you're going to have a much better time making UIs than with any other framework.

File: 2024-10-29 14-35-03.webm (3.82 MB, 1280x720)

3.82 MB WEBM

>>103020189
amazing what running a release build on its own can do for your framerate when compared to a debug build launched from the visual studio debugger

I don't want to make a game I want to make a 3D modeler that runs in the browser with WebGPU because Blender is AIDS

any gedg discord?

>>103020374
Ooh, nice game. Can you tell me more about it?

>>103020837
I can't think of anything more AIDS than a 3D modeller than runs in the browser

>>103020837
javascript is gigaaids

File: whiplash nemesis.webm (2.97 MB, 640x397)

2.97 MB WEBM

>>103020887
It's an old DOS racing game I'm trying to reverse engineer from disassembly. Got the graphics and track files 95% figured out and I have written a pretty fully featured track editor application. Working on making a shitty game engine now so I can start figuring out the physics. I am coming up with my own crap version first, this helps me to better understand what I'm looking for and possibly even simply plug in the constants when I find them. Project is here: https://github.com/Zizin13/Roller

>>103021090
Emscripten exists

>>103020838
We can ERP in here if you want.

I slide my hand slowly down your back. You shudder with delight when I whisper in your ear and call you a dirty rust tranny as I overstep my c array bounds.

File: pkm2.webm (3.82 MB, 1920x1080)

3.82 MB WEBM

Y-sorting? Fuck no.
I made a tile postprocessor to calculate z position of everything based on placement and tile type. Further down the line I'll actually need a 3d grid, but this will do for now.

>>103021498
wtf anon

>>103021498
Personally I don't appreciate that retro filter, but other than that it looks great, anon

>>103021498
>Y-sorting
only naive devs do sorting, literally every game that isn't 3D on retro platform never sorted. All you really need to do is to draw the whole ground, then check and redraw the exact tile that overlap the player/entity after you draw it. Basically a bottom/mid/top layer. This is literally the case for tactical games such as Tactics Ogre on GBA.

>Y-sorting
use the depth buffer

File: anonymous.png (1.17 MB, 1693x845)

1.17 MB PNG

>>103010955
Always props to a fellow engine chad, it already looks good and with a good gameplay idea you're already there. But god, if not, you got a looong way head of you, I see you suffer from copy-buffer lagging problems too, feelsbadman

File: ec8a2319737a1b33930f582b4(...).jpg (33 KB, 468x528)

33 KB JPG

So, after 13 months, I think I finish my own engine. A very simple 2D engine programmed in C and C++... My conclusions:
>windows sucks
>visual is a kick in the balls
>microsoft sucks
Now hearing: https://www.youtube.com/watch?v=6LB7LZZGpkw

>>103022068
so where's your engine?

>>103022081
I won't share it. It's basically RPG maker but for create platforms games.

>>103022068
>>windows sucks
>>visual is a kick in the balls
>>microsoft sucks
all true
this is why you develop on linux, cross-compile to windows, and if a customer tells you there's a bug on windows tell them to get their rags off

>>103022157
>this is why you develop on linux, cross-compile to windows
How? With CMake?

>>103022057
Thank you for the words.
I'm trying to solve the lag now by putting chunk building on another thread, and then also have a 3rd thread dedicated to vkQueueSubmit.

I'm willing to hear any advice if you'd have it.
God bless.

>>103022183
yeah

>>103020374
>launched with debugger
FYI this is what causes most of your performance loss, not the fact of being a debug build. Mostly just being a debug build will bloat file size but not impact runtime performance noticeably. It's actually attaching the debugger that tanks performance.

File: wat.jpg (42 KB, 313x355)

42 KB JPG

>>103022240
what? Debug builds don't use optimization flags when compiled. That makes them slow as fuck

>>103022157
>this is why you develop on linux, cross-compile to windows
how is that less hassle than developing on windows

>>103022256
everything about working in windows is cancer, from the terminal to package management

>>103022280
If you're using the terminal or package managers to build a game engine you're doing it wrong

>>103022290
lmao ok bud

>>103022293
A game engine should be a self contained piece of software, you don't need to be pulling in dependencies or doing devops shit
It's been at least a year since I've touched a build system

>>103022309
just use docker, its not hard
you're making it seem more complicated than it is
its a self-contained piece of software, but you should have iron-clad build reproducibility

>>103022253
The compiler absolutely still does a ton of optimization, and the performance difference between a release build and a debug build with no debugger running is comparatively minor. Attaching a debugger causes massive performance differences.

>>103022318
>just use docker
>you're making it seem more complicated than it is
lmao fucking webdevs man, is your engine in JavaScript?

>>103022332
do you know what docker even is? I've never done any webdev
but now I'm remember why game devs like you are some of the worst paid software devs in the world kek
anyways bye

>>103022280
This. Windows is not KISS.

>>103022339
Yes, Docker is a tool used by webdevs

>>103022350
yeah I'm not spending time teaching someone that makes less than 6 figures

>>103022354
What's a webdev going to teach me?

File: anonymous.png (1001 KB, 1472x868)

1001 KB PNG

>>103022184
ha, i'm still on OpenGL so, life's hell. In theory even in OpenGL it's possible to avoid if I limit buffer copying sizes to some budget that allows for my desired framerate, even though it's the main thread. DESU, my current system is so shitty I want to just tear it out. I'm learning OpenCL so maybe i can create some of my assets on-GPU.

Don't know much about vulkan other than I thought the copybuffer could happen in a different thread. In fact I was hoping it would resolve my issue.

OpenCL...OpenCL is gud. 0.0

File: anonymous.png (4 KB, 397x88)

4 KB PNG

>>103022472
Can't tell if OpenCL is really 7000x the speed of single threaded or I'm doing something wrong, ofc the CPU can multithread a little bit, but still

>>103022057
>>103022445
nice aesthetic

>>103022494
why use opencl and not a compute shader? opencl is not very portable, OS and driver support is a shitshow

>>103022505
Oh, maybe you're right, idk, I guess I liked the syntax.

>>103022494
Watch out with a lot of that gpu stuff, the compilers love to optimize the hell out of your code so much that if it can omit a long winded calculation because you didn't store it in a buffer at the end, it will.

>>103022509
well, anyhow, if your problem is an "embarrassingly parallel" problem then 7000x is pretty standard

File: anonymous.png (89 KB, 250x226)

89 KB PNG

>>103022519
>>103022525
If I read you right, my example is likely optimized away (equivalent to -O3 in gcc) on the GPU side, and I need a better example. However, if I do make a non-trivial example, it can still be embarrassingly parallel, it's possible it will still be a big savings.

>>103022569
as long as you're not hardcoding data into your code it shouldn't be optimizing away calculations much unless they're really not used in your final output
just test it with different input-output data sets

if they all match, you're fine. write a test for your CI flow while you're at it

>>103022635
>write a test for your CI flow
where are all these webdevs coming from

File: anonymous.png (274 KB, 486x368)

274 KB PNG

>>103022635
>ci flow
nigga...

>>103022652
its funny that you think CI is something only webdevs do
you'd think the JPL was just praying to the moongod every time they updated code

File: 1494131241133.gif (3.75 MB, 280x302)

3.75 MB GIF

>>103022635
>CI flow

>>103022680
>its funny that you think CI is something only webdevs do
Am I wrong?

>>103022702
I take it you've never worked professionally in any kind of sw dev or engineering
That's cool, whatever. Just get more experience.

File: anonymous.png (2.71 MB, 1024x1024)

2.71 MB PNG

>>103022680
>JPL
I never want to hear about the useless NASA faggots ever again. Wake me up when SpaceX, ChatGPT, Deepmind, or literally anyone relevant does it

>>103022707
Yes I have, that's why I was asking

>>103022713
kek

File: perfect human.jpg (102 KB, 508x512)

102 KB JPG

writing tests and ci are software development humiliation rituals

>>103022716
???

>>103022717
They're things very serious employed people do

File: anonymous.png (1.28 MB, 1453x856)

1.28 MB PNG

>>103022525
Ok I asked o1-preview for a non trivial problem and it LGTM. 750x speedup over -O3 code, that's pretty impressive. I guess compared with 64 cores tho things get a little weaker. Jeeze, a 10x speedup, how lame.

>>103022833
well consider most end-users are more likely to have a beef gpu than 64 cores
you also may still have trivial problems you can exploit as well, there's a lot you can use on the GPU - just make sure your don't ever get killed on transferring data back to CPU

if you can leave it on the GPU (usually you can with rendering pipelines for example) leave it there, don't read it back

again, tests would help the maintainability of all this

File: anonymous.png (1.05 MB, 1337x839)

1.05 MB PNG

>>103022872
Trust me when I say...copybuffer is the bane of my fucking existence

>>103022884
best you can do with that is try to amortize the cost across frames with some async shenanigans using a transfer queue

>>103022290
>>103022309
You can statically link against libraries installed through a package manager. The only real problem is fuckhueg libraries like ffmpeg that you really want to slim down yourself with build flags.
I build self-contained windows executables with MSYS2 and it works perfectly fine.

File: gabe laptop.jpg (537 KB, 3031x1705)

537 KB JPG

This is an actual photo of Gabe Newell in the middle of his daily 4chan run. Every morning, before he starts the daily meeting on how to further screw over Artifact fans (attendance is mandatory for every Valve employee and takes about three hours), he visits /g/'s /gedg/, /vg/'s /agdg/, and checks /v/ for game dev threads. Each game he sees in these threads, he checks if it's on Steam, and if it is, he manually suppresses it in the algorithm to make sure nobody can see it without specifically searching for it.

Do NOT post your games here. You have been warned.

I got what I think is decent PC (5600x + rtx 3060)
What's a good game engine for newfags? My long-term goal is to make a 3D action game

>>103010966
Cute abby

>>103020838
there's an irc but i don't know how active is it

File: vegetagrass.gif (2.8 MB, 498x371)

2.8 MB GIF

>>103021498
Goddammit, someone else has the same idea as me to make a pokemon engine

>>103024803
>Goddammit, someone else has the same idea as me to make a pokemon engine
I already did that as my beginner programming practice project long ago, so I'm not interested in doing one again. I'll be trying to make a mix of monster taming, survival and zelda gameplay. If this risky idea doesn't work out I'll experiment with pixelart by doing procedural generation.

>>103023665
Stop lying, he’s putting all the 4chan projects on the front page

File: kyriakos-grizzly-grizzly.gif (75 KB, 220x164)

75 KB GIF

>>103023665
>Gabe saw my game


>ain't verifying my email
>ain't buying the pass
suck my dick

Alright give it to me straight. Zig or Odin?

>>103025609
C

>>103025635
no. I don't want to use old language

>>103025609
Odin > Zig
Zig is too verbose and those fuckers can't keep themselve from changing everything every update.

>>103025731
Odin cannot even attach methods to structs like how zig and Rust do. It honestly kills the language for me.

>>103025779
>attach methods to structs
that's just syntactic sugar for

thing_do(thing)

vs

thing.do()

neat feature but the rest of zig is meh.

File: 5639758356834.png (652 KB, 1134x637)

652 KB PNG

How do I improve my cities? Obviously need to design more building types and will be adding in generation for small details like cars and billboard things along with street lights

>>103025838
I build roads first. Allows me to get the visual layout, then I try to concentrate the taller buildings together and smaller/less dense buildings around them as it fades into residential as it spreads outwardly. Tight alleys between buildings that have spaces are cool but it looks like your buildings are spaced rather oddly for any sort of road pattern. That's my 2 cents anon. Love ya

>>103025885
Yeah the way I'm doing it is pre building the cells since I don't trust proc-gen to make a realistic looking layout, probably do need to put more effort into the road areas to make it look more lively

>>103025651
C23

>>103025837
except you need to pass the struct as the first parameter for everything. gets ugly super quick.
Might just use zig because I really dislike the C way which is what odin does.

thing_do(thing, parameters)

vs

thing.do(parameters)

>>103025891
I'm expanding my town currently and the biggest thing that ties together the "realism" (mines first person) is the scale, the road feels proper scaled, the houses and buildings and cars all feel scaled proper. Next is making sure there's not strangely empty areas around. But no anon, it doesn't look bad by any means. And I feel ya on procedural gen. All my game is hand placed.

>>103025908
see >>103025779

>>103025934
>>103025779
bloat

>>103025934
>>103025912
That oop shit is awful and makes the code a lot harder to follow

>>103025912
understandable. Odin wanting to be a better C, don't want to add methods on struct. You will be better of in Zig then. personally I don't mind that so I prefer Odin over zig.

>>103026018
>That oop shit is awful and makes the code a lot harder to follow
It's not OOP, though it does follow the principle of encapsulation

>>103025838
I'm just opening up google maps and copying the city I live in

File: sdfsdf.png (406 KB, 633x479)

406 KB PNG

>>103016062
did just that, it shows up in the editor, but not in the game

>>103021498
Cool lighting effect. Looks a lot like Khas, did you use him as a reference?

>>103026018
>makes the code a lot harder to follow
only if you're really stupid

>>103026293
>>103016062
The funny thing about Godot is that the drama is a month old at this point and yet their funding keeps jumping above 57k and then below 57k again, which suggests that the autists are still fighting their donation war

>>103025609
Nerd trash

>>103026547
I have no idea what Khas even is, but thanks.

File: khas.webm (545 KB, 1280x720)

545 KB WEBM

>>103026879
Khas is a custom script for vx and vx ace that looks almost exactly like this.
It uses ruby and manipulates bitmap in memory though (no shaders or anything and ruby is very slow) so I have no idea how it manages to reach 60FPS

File: dfsgdfg.png (851 KB, 1141x643)

851 KB PNG

>>103026293
Used a material override, it displays fine but it broke completely lmao

File: pkm-light.webm (1.83 MB, 1920x1080)

1.83 MB WEBM

>>103027259
Oh, I see. I couldn't find it through search.
Mine can do more actually because it supports penumbras. It can also use many light sources and cover massive areas.
It still has some artifacts and can't light front of the wall though because it wasn't meant to be used in that context.

I'm looking for a game engine not associated with some sort of twitter or discord tranny cabal which is made for actual game making, not as a theoretical structure that could be made into something to make games with enough redundant setup of 500 things everyone will obviously use but some GNUtard thinks "well we can't just ASSUME that we have to leave it non functional for practical reasons and force everyone to go through a time wasting humiliation ritual to set it up!"
I want something that simply works and what I DON'T want is some fag to go "hurr durr Scratch RPGmaker xdddd"
OKAY?
Shill me GOOD ENGINES ONLY

File: sdfsdfsfd.png (3.72 MB, 2535x1329)

3.72 MB PNG

>>103027442
it was just an issue with ALPHA = COLOR.a in my shader
threw that out since it's useless anyways

>>103027472
this isn't the right thread for this

>>103027472
unity
unreal
if you want to go open source and not chose godot, then bevy or defold

>>103027492
>Game And Engine Development General
>>103027508
first two violate my very first stipulation, so does godot, I will try the other two thanks

>>103027515
bevy is most certainly a theoretical game system made largely by literally trannies (you can look at the contributors, its been tranny central since day 1)
as far as I know no one has shipped a real game with it yet that anyone cares about

>>103027515
No one here is very interested in shilling and flamewars and buzzwords, that's exactly what the /vg/ thread linked in the OP is for, you would get a lot more replies and discussions about those topics there.

File: 1730303663908.png (43 KB, 1028x666)

43 KB PNG

I am done with the level/stat system though it will probably be improved later on. Now I just need to implement auto tiling and animation system upgrade. After that I am basically done with the editor

>>103027472
Unity, Unreal, Game Maker

>>103027472
Use MonoGame, it is the game engine of choice for Stardew valley and afaik it has a similar API to Unity, except it is FOSS

>>103027605
Defold it is then I guess.
>>103027695
I'll check this one out too, thanks.
>>103027617
I don't care about the offense you take to me wanting to avoid psychopath pedos who constantly shill their fringe beliefs at everyone, censor or shame anyone who calls them out for it, and generally must always "control the narrative".
Acting world-weary about the divisiveness you go out of your way to cause by slandering unrelated people incessantly and constantly "pushing the envelope" is a worn-out defense mechanism.
Give it a rest.

>>103027472
GLFW + OpenGL

>>103027472
SDL3

>>103027472
raylib if you want to just work

>>103027725
>Acting world-weary about the divisiveness
Yet you enter this thread and bring your retarded childish /pol/ shit in with you

>>103027798
Nope, I just wanted to make game and avoid having to be involved with a "community" of idiots like you. I don't care why it makes you seethe so much and I do not want to hear your complaints about people rejecting your psychotic subculture.

>>103027811
You're bringing your gay culture war into this thread
You're imagining culture war where there is none (Unity and Unreal have nothing to do with Twitter or Discord cabals)

>>103027674
Are you making a RPG maker? You should join
https://ophanimkei.com/you/RPGmaker

>>103027830
I asked a question, you seethed, your ignorance isn't my problem, take your fucking poppers and go

>>103027811
>Nope, I just wanted to make game and avoid having to be involved with a "community" of idiots like you
Same, and then you showed up. Every so often a retarded tourist from god knows where starts shitting up the thread and it's getting tiring.

>>103027861
rust solves memory leaks

>>103027861
There's been too many of these lately, somebody needs to tell them to leave

>>103027852
>>103027859
I accept your concession and I hope you will either learn to make better posts or try another website such as Twitter.

>>103027861
It IS tradition on /g/. This board desperately needs flags and IDs, that is the reason why /bant/ is so much higher quality than /b/, even though the same people use it

>>103027887
Go back to /pol/ you stupid fucking child

>>103027908
I'm gong to have to ask you to calm down and discuss video game engines, rather than your feelings and board politics. There's already a board for that - you mentioned it.

>>103027897
Don't worry 4chan is on it's way out. hiro is making sure of it. We'll soon be free from this place.

>>103027929
What will I do then? Go outside? Do real work? Where will I post my toy projects then? Will I be forced to find real people to keep me company?

File: 1727972811297113.jpg (53 KB, 554x816)

53 KB JPG

>>103027937
I wish I knew

>>103027832
Yes, I am, it has been two years since I started but I had a lot of other work, my codebase for the editor is 16kLoC large and 10kLoC was written in the last two months because of that.
> NOTICE Ring not accepting new members at the moment because I need a break. I won't be mad if you email me to inquire though. I might just take a really long time to respond. Please don't comment on my profile about this.
So I should email the owner or something? Also you need a website for this, don't you? That is what webring is

>>103027746
>>103027472
GLFW + GLEW + GLM

>>103028026
Pill me on OpenGL vs GLEW+GLM

>>103028067
Glew is just a cross platform wrapper that makes it easier to use. Not really necessary but nice.

GLM is just a math library. Handy for all the linear algebra shit you're gonna be doing, no point in re-writing all that.

>>103028106
Sounds pretty good actually. Do you think I could make a game that compiles on OpenBSD with this toolkit?

>>103028125
I don't see why not

>>103028106
>>103028125
Use SDL instead of GLEW because you will need audio and image support anyway and SDL has the best audio library (mixer) and SDL_Image supports the most formats of any library (except maybe ffmpeg, though I am not sure if image support is even in the actual library)

>>103028147
I use nothings stb for image and audio.

>>103028157
suit yourself, but with SDL you get zero crackling with audio and you can export to any platform, including web

>>103028147
Did they ever fix SDL so the built-in function for displaying PNG or GIF or other various file transparency actually works? It was totally fucked when I tried it a few, maybe five years ago.

>>103028403
With that you mean displaying the SDL_Surface or the SDL_Texture? For the latter, you have to do

    SDL_SetRenderDrawBlendMode(renderer, SDL_BLENDMODE_BLEND);

The reason for no blending by default is that SDL supports multiple backends and not all of them are hardware accelerated. With hardware acceleration it doesn't matter because every pixel runs in parallel so blending will be just as fast as no blending. But for software rendering you really shouldn't use blending if you don't need it because the number of CPU cycles is multiplied by 20 or so, not counting the memory access delay. And even if you use it, you probably don't have half-transparent pixels, just visible and invisible ones so blending is overkill

>>103028939
At the time I tried both, neither worked, I did a bunch of different fixes from StackOverflow and then finally found a comment that there was a long-standing bug with it.

I do not personally care if I have to use a Surface or a Texture, I would just like to have an image with transparency displayed in the game window with transparency rather than surrounded by white pixels as if it did not have the transparency.
So what I would like to know is the least costly method of displaying transparent images as transparent images in the game window, in such a way that they actually display as transparent images, not images with white where it should be transparent. Yes, I am very sure I made them transparent correctly.

File: 1730309985479.png (329 KB, 572x482)

329 KB PNG

>>103028976
I don't know anon, it works for me

>>103028147
>SDL instead of GLEW
Isn't GLEW just an OpenGL loader? SDL doesn't come with one, they do completely different things.

FOSS is too modular who can keep track of what all these little trinkets do

>>103029277
Right, I mistook it for GLFW, been a while

>>103029315
I don't I just look for the trinket I need when it comes up and then choose one.

While I think enginedev for the purpose of making a game is a pointless and timewasting endeavor, I respect the technical prowess and work of the /gedg/. However can you not intrude on /agdg/? That general is about making and SHIPPING games not making game engines.

File: cris_no.gif (1.77 MB, 520x450)

1.77 MB GIF

>>103029513
>/agdg/ shipping games

>>103029513
if you can make a game, you can make a game engine.

there are far more successful indies that made their own engine than those who used unity or godot.

>>103029821
There is only one successful game made by one person (Indie), that is Undertale and the author used Game Maker

>>103029842
pretty sure eric barone made stardew valley by himself and he made his own engine.

>>103029842
notch wasn't one person?

>>103029861
Stardew Valley uses MonoGame
>>103029872
was he? didn't he have Jeb and other fags with him?

>>103029524
Over the past tens years they have had 6 releases

>>103029884
monogame isnt an engine

>>103029910
It is a framework which is almost the same
Stardew Valley is 2D and even then the creator didn't want to use raw OpenGL or Vulkan like the retards here advocate for.

>>103029938
>>103029910
desu if your framework is something you just plug scripts and assets into, its an engine
does monogame even let you write your own tick()?

>>103029964
>does monogame even let you write your own tick()?
surely it must, even Unity allows that

File: 1725378805060851.jpg (35 KB, 640x513)

35 KB JPG

>if you didn't write your game with C + OpenGL, you aren't an enginedev

>>103029938
>It is a framework which is almost the same

lmao

is raylib a game engine

is sdl a game engine (im pretty sure that 90% of the posters here use sdl)

i know you web devs have a cancerous crab bucket culture so let me try and explain this.

would you rather have a 2d platformer with 500 levels written in godot (or some shit) using godots default scene editor and using godots opaque formats.

or would you rather have a 2d platformer with 500 levels in your own engine with your own tailored editor and your own file formats that you can automate changes and updates upon.

if you've ever spent any time in a game engine like godot or unity you'd know they scale horribly for this very reason, and it quickly makes gamedev an absolute slog.

>>103030028
Reddit spacing
>would you rather have a 2d platformer with 500 levels written in godot (or some shit) using godots default scene editor and using godots opaque formats.
>or would you rather have a 2d platformer with 500 levels in your own engine with your own tailored editor and your own file formats that you can automate changes and updates upon.
Godot because I hate dealing with physics. Though if I just wanted to make SMB clone it would be easier to roll my own

>>103030016
well I think unity just lets you overload update() (or something named similar) function for gameobjects right? I mean actually writing the mainloop

>>103029821
>there are far more successful indies that made their own engine than those who used unity or godot.
That's so absurdly untrue

>>103030292
but it is true thoughbiet

>>103030316
Over 90% of successful indies are using an engine like Unity
Go look at Steam you idiot

>>103030277
oh that. I don't know

>>103022445
in Vulkan you can record commands to a command buffer from any thread as long as the command pool which the buffers come from are made on the thread you're recording from.

These commands are recorded and then submitted through vkQueueSubmit ( very expensive operation ).

This is the same thing with Queues, but the problem is there are devices which have only 1 queue so if you want to support them, you would instead dedicate 1 thread to only Queue submission instead of making a Queue per thread.

And by the way, I just removed all uses of copy_buffer from my program ( on the white board. Still need to implement the removal ).

>>103030350
and the vast majority of those aren't actual indies but small studios of around 20 people.

im talking about actual indies.

>>103030400
As the size of the studio goes up, the likelyhood of them writing their own tech increases, not decreases
You literally have no idea what you are talking about

>>103030415
no most small to medium sized studios have abandoned inhouse engines and just use unity so that hey can have a revolving door of employees to burn out.

>>103030428
Most one man projects use an existing game engine
Most small size studio projects use an existing game engine
Most medium sized studio projects use an existing game engine
When you make a post, please post things that you know are true, instead of making things up and assuming they are

>>103030022
>OpenGL

:/

File: 1729008856429641.jpg (39 KB, 640x640)

39 KB JPG

>>103030472
Look at you big guy using Vulkan.

>>103030448
that vast majority of single-man developed games use their own game engine. if you actually have used game engines and aren't just some web dev crab, you'd know the reasons as to why developing your own engine is the best option. thet vast majoirty of game development isn't on graphics, or physics or engine systems it's on content generation and creating a good content creation pipeline.

when you are a single dev, this pipeline needs to immaculate. you cannot waste time onm shitty default editors or opaque file formats, you need something that scales and something can assert a massive amount of control over. i.e you need your own tools.

>>103030502
>that vast majority of single-man developed games use their own game engine
No they don't, they use Unity, GameMaker, RPG Maker, etc
Stop posting bullshit

>>103030488
The only backend I've written in my engine so far is Metal. But I'll probably do D3D11 and Vulkan sometime later.

>>103030523
assmad web dev lmao.

what i said is true, almost all the famous hit indie devs used their own engine.

you can cope all you want. but this isn't web dev, making games irregardless of wether you use and engine takes a lot of effort. you can't just react sloppa a game, you have to actually try.

File: 1696014910861585.jpg (166 KB, 629x1023)

166 KB JPG

>>103010955
How far along should a game be before trying to stream development and doing vlogs and stuff?
I'd hate to start posting and lose steam or get busy with my job and family, then lose all credibility.

>>103030578
This isn't about me, it's about you posting things that are blatantly untrue
Almost all the famous hit indie devs use Unity

>>103030602
name 11

>>103030594
It's good to get yourself out there and under the eyes of other people even if you do end up cancelling a project, people aren't going to hate you for it (unless you made ridiculous promises that you never keep)

>>103030618
Open up the Steam store and scroll through it I'm not here to spoonfeed you the obvious

>>103030631
so you cant name 11?

>>103030618
Nta but can you name 11 that made their own engine?

>>103030638
I can

File: 1714574045346569.png (253 KB, 598x628)

253 KB PNG

>>103030620
I think I'd rather have a vertical slice finished at least. That way I can stream the fun stuff like making content (enemies, weapon types, level design)
Streaming/vloging the boring boilerplate UI, fighting with a game engines physics and AI, and general bookkeeping type stuff doesn't seem like it would draw in viewers and potential sales

>>103030594
Once it starts looking like a game

File: Ampoohros.png (91 KB, 270x301)

91 KB PNG

What's agdg's opinion on using assets from asset stores? I'm not talking about asset flips, I know those are gay

>>103030691
Most gamedev streams aren't interesting content either way I don't think you should hold yourself to high standards, just try to get yourself out there and network with other developers

>>103030723
I recommend asking them.

>>103030723
Trannies love stealing things and pretending it's something else

>>103030652
minecraft
binding of isaac
stardew valley
celeste
shovel knight
dead cells
rimworld
dwarf fortress
kenshi
terraria
the witness

>>103030734
why would networking with other devs help?

>>103030785
>binding of isaac
flash is an engine

>>103030785
A lot of the games you listed are using frameworks, not rolling their own.

File: 1730229882741939.png (588 KB, 900x506)

588 KB PNG

>>103030785
you can't win
you're arguing against a "no true scotsman"er.

>>103030851
not the original web game you doofus

>>103030881
frameworks aren't engines.

is raylib an engine?
is sdl3 an engine?

https://youtube.com/watch?v=RJoLcfnguO0
He actually did it. He's not one of us anymore.

>>103030915
Raylib and SDL are libraries, frameworks are closer to engines then they are libraries.

>>103030923
>your only goal is to survive
Why do so many indie games forget to actually make the game interesting? That's hardly a unique gameplay loop

>>103030939
framework = engine without GUI

>>103030950
But bro the generic droning rave music bro

>>103030755
What's your opinion on using assets on asset stores?

File: disdain-for-plebs.jpg (43 KB, 600x693)

43 KB JPG

>>103030958
Seriously though, why are so many indie games just a platformer, or an RPG, or a generic 3rd person action game?
Making your own game means you can do weird stuff like combining two genres or making a mechanic that wouldn't work right for modern AAA movie games.

Worse than those generic games are the artsy games that expand on a mini-game from some popular AAA game

>>103030995
Check out liminal galery

>>103030995
Same reason most art or music is generic and tasteless. The average person, even creatives, are devoid of souls and can only shit out a basic concept at best.

>>103030848
Outside of exceptional circumstances consumers aren't interested in you or your work as an amateur developer
Every developer starts out by networking with other developers

>>103030995
Because people who create things are inspired by other things, they don't set out to create with an experimental mindset of lets do this and see what happens, they have an idea that they think is cool (which is inspired by other things they've seen) and that's what they want to make

>>103030995
What if my game is a platformer RPG 3rd person action game?

>>103030995
Combination of the fact that these are pretty basic, easy genres that don't really require programming knowledge (platformers / JRPGs are simplistic as fuck, 3D action games are not but the standard game engines come with them as premade projects basically so you skip the hard parts and just tweak them), and the fact that most indie game devs are tasteless hacks who grew up playing shitty platformers and JRPGs on their sega gamecube or whatever japcrap console they had as a kid and that's what they think video games are about.

>>103031050
But why does it help to network with them? What is it helping? Surely not sales

>>103030785
>stardew valley
>celeste
>terraria
these 3 are all XNA, which was unity-tier popularity in it's day

>rimworld
this is unity retard

>>103031080
Mirror's edge gameplay in third person with RPG progression would be pretty dope desu
Your joke is a better idea than most indie games

File: Come and take it bullets first.png (22 KB, 557x475)

22 KB PNG

>>103010955
Alright, I'm coding a 3D engine for the NES.
Why? Autism.
I need some help figuring out a few things.
>What 3D rending system runs the fastest with kb's of RAM?
Reycasting has been proven possible, though from what I've heard rasterization is faster.
>Do you guys have any tricks to speed up rendering?
I know about EOR texture 'unrolling', frame buffers, and spaced line interlacing. Researching portal rasterizers on the c64 shows shit really hits the fan when rendering over 20 tris. (see freescape).
>For reference, please assume I have 128k of CHR RAM to work with and 256k PRG ROM and a CPU clock speed of 1.79 MHz.
Any suggestions are appreciated. Assembly adjacent systems (like c64 demos) are more likely to work with the modified RAM.

>>103031119
How do you think you get noticed? By being part of peoples social circles. Do you really not understand networking?

>>103030785
Anyways, to replace them I'd say,
Slay the Spire (LibGDX)
Factorio
Faster than Light
Noita
Teardown
CreeperWorld 4

File: joe.png (78 KB, 219x195)

78 KB PNG

>>103031145
I really like challenging ideas like what you're doing, that's a really cool thing to try and do, good luck

>>103031152
Get noticed by other indie devs? Which is a group of a few thousand people? How does that help you get noticed by the people you're trying to sell your game to?

>>103031050
Networking with developers is only useful to find someone to work with you or for you to stalk and see how they handle things

>>103031145
Do you want it to run on real hardware? Because if not, bank switching with mappers that couldn't work on real hardware (but would give you access to a shit load of ROM) could let you use prerendered wall shapes to cheat the environments

Faceball on the Gameboy worked, so you should be able to do what they did

>>103031190
>>103031212
You can't network with consumers as an indie developer because consumers do not care about you or your game. You network with other developers, which gives you opportunities you wouldn't have alone

>>103031251
>which gives you opportunities you wouldn't have alone
What are those opportunities?

>>103031251
Networking with other developers provides no benefit for attracting consumers, other than seeing how they market and advertise themselves and maybe getting close enough to actually discuss their (working) strategies. However most devs doing this have no audience but other devs, Pissing into an ocean of piss.

>>103031231
>Do you want it to run on real hardware?
Yes, the chipset I'm currently targeting MMC5 EW-ROM.
>>103031190
Thanks, I'm gonna need it.

>>103031251
How do you attract consumers if you don't approach them? You need a way to put your game infront of them. Sure devlogs are not the best way since it tend to attract people interested in gamedev and are not necessary your target audience, but it's way better than nothing for sure.

>>103031263
>>103031265
You know a developer who knows a guy in the media who can get you an article or a video
You know a developer making a similar game to yours that can funnel their audience into your audience

>>103031327
You can reach out to media, streamers or youtubers yourself and have a much higher success rate.
Washed up devs *might* be willing to throw you a bone but any active dev will not put a spotlight on a competitor.

>>103031336
Why do you think cold calling someone has a better success rate than knowing something in the business?
Jesus christ the retards in this thread today

>>103031347
"cold calling" someone who's job is writing about or making videos on games about your own that is similar to things they cover has a much higher success rate then begging literal whos to get in contact with another literal who who might know a third literal who to maybe get in touch with you

>>103031145
cool project.
So it is official, rasterization is faster. I knew it.

>>103031327
>>103031336
>>103031347

Seems like the era of devlogs like wolffire is over these days and the only way to get attention is to make a cool looking trailer when you're 90% done with development, post in gaming spaces if they don't ban advertising yourself, and hope those spaces lead to attention from various media

And the only reason I say when you're 90% done and not when you're totally done is because it will take some time to get organic attention, so you'll need to post every so often while you're finishing up your game.

This is the advice I've found in the current decade for making a indie games

>Jesus christ the retards in this thread today
>>User asked a very good question
>>You explained it poorly
You're the retard bro

>>103031357
You're incorrect
As anyone who has had any sort of job should know, knowing people who know people is always better than approaching a stranger

>>103031367
You can approach 100 strangers (who are desperate for new things to cover) meanwhile you are on your knees begging and crying for someone twitter to do your job for you.
NGMI end of conversation.

>>103031347
>...
>be friend with someone that knows someone
>...
>profit!
there's some steps missing

>>103031384
Nobody is desperate for things to cover. There is a complete oversaturation of games right now. Most media will just ignore you

>>103031395
Am I talking to literal children here
If you know somebody who knows somebody, you ask them if they can put you in contact with them

>>103031425
>ask them if they can put you in contact with them
"nah bro sorry too busy"

>>103031433
You are literally inventing hypothetical scenarios to dismiss advice because you don't want to believe it

>>103031425
>Am I talking to literal children here
>If you know somebody who knows somebody, you ask them if they can put you in contact with them
fucking nigger retard, you're once against just jumping straight to already knowing somebody who knows somebody.
>how to be rich? Just be rich bro!

Forming a friendship with someone influent to then have them refer to their friends will be alot more work than cold calling influencers and media.

Just go hit up toby fox or whoever on X and ask him to shill your game, bro. it's easy, bro.

>>103031460
Yes you are inventing hypotheticals where you must suck someone off on twitter and nothing else would work

>>103031468
That's true, but friends are what get you places, cold calling has a low success rate
Making friends within the game development scene is the best thing you can do outside of obviously making games

>>103031470
I'm not inventing hypotheticals you petulant moron, I'm speaking from experience

File: file.png (22 KB, 714x130)

22 KB PNG

Does your game have stuff you don't understand?

>>103031504
Yes and it's not even a game I'm just drawing a triangle to the screen

>>103031425
Anon if it makes you feel better I get what you're saying. I would've never had my current job if I didn't network, and the same principles can apply to just about anywhere in life.

>>103030400
>>103030428
>>103030523
>>103030602
>>103030652


only faggots use pre-made game engines, all the best games ever made rely on one thing: coming up with some new wacky wahoo way of doing something specific and then writing a game around that
wolfenstein 3d
doom
portal
half-life

telling people that making their own engine is a waste of time is some truly retard level arguing, it's not hard to make a game engine, but writing it yourself teaches you intricate things you otherwise wouldn't know, it restricts you in a way that makes you more talented, it gives you something no one else has
my game will always be a thousand times better then ever rpg maker faggot because I didn't rely on pre-made images, sounds, music, OR code outside of the basic graphics libraries im using

>>103032810
>it's not hard to make a game engine
Actually it is

File: Screenshot at 2024-10-16 (...).png (1.01 MB, 1548x904)

1.01 MB PNG

>>103032817
well damn you sure told me, I'll just throw out all these engine designs I've been working on since they clearly don't exist

>>103032868
I didn't say it was impossible or you shouldn't do it, I said it was hard

>>103032868
based
>>103032810
based
>>103031716
>>103031425
>talking to literal children
yes, most have never had a job in their lives let alone a job in any kind of SW engineering field, and of that small minority most are just jr devs that haven't even cracked 6 figures (in 2024)

take what you read with a huge grain of salt, this thread is better for helping people than getting help

>>103032979
have you ever even tried to make a game engine or are you actually just talking out of your ass.

you can literally make a 2d sprite batch renderer in opengl in like 30 mins, its like 100 lines of code.

>>103033024
I have made a 3D game engine

>>103033034
then you know how actually suprising easy it is then. basic blinn-phong lighting is 20 lines of shader code.

>>103033043
No I know it's much harder than it seems to a beginner
a 2D sprite batch renderer or blinn-phong lighting is not a game engine

>>103033075
how?

>>103033090
Because you look at all the components and you say "oh that's easy" but the complexity of your game engine explodes when you have to take all these pieces and put them together into a coherent whole
2D and very simple 3D is managable, if you want to move beyond games that look like Quake it starts to become very complex

>>103029513
I'm not opposed to using an existing engine if it's the right tool for the job, but for my project of reverse engineering an old game so far I've had a much easier time writing a bespoke engine and editing tools than trying to get UE5 to pretend it's 1995. Even ignoring that hurdle, there's also something to be said for having the entire codebase in your head and knowing it like the back of your own hand versus a colossal project made by hundreds of engineers over the course of decades that probably no one fully understands in its entirety.

>>103033116
your actually talking out of your ass.

your not even giving specific examples or talking about anything concrete just "it gets more complex dude trust me". another fucking crab-bucket webshitter in this terrible thread.

>>103033159
That's because the complexity does not come from the implementation of specific pieces of the engine but tying it all together

>>103033168
okay how so? tell me about this 3d game engine you "made". give me some examples of problems you faced.

>>103033024
Being a programmer shouldn't be a string together certain libraries and code kind of affair. How much of that was really you doing something?

I urge everyone to here to be more humble. Learn as you go. You are not just someone that connects pipes. You are the one making new parts and a builder of constructions at heart. So learn to always go deeper and never be content with how it is right now.

>>103033274
what are you on about. opengl is a graphics api. its how you program the gpu.

>>103033281
"you" did so little it's not even funny.

>>103033281
opengl has a stateful driver maintained by a third party, its training wheels
real men use vulkan

>>103033289
>>103033290
a 2d batch renderer in vulkan is still about the same level of complexity one you get past the vulkan initialization.

i was making a point you fucking retards.

>>103033314
Stop coping. Be humble or just shut up.

File: pepefact.jpg (72 KB, 1200x614)

72 KB JPG

>>103033314
I don't really care desu, I'm just here to lay out the fact that people that use opengl are webdevs

>>103033327
actual retard lmao. the funny thing is you didn't even start this shit with low effort trolling. you just got ass-blasted and now are desprately trying to pivot.

you will NEVER accomplish anything in your shitty lives you react sloppa web-shitters. you cope and crab-bucket for the rest of your lives.

>>103033351
I am not even that anon. You are just unhinged. Go touch grass man.

>>103033370
but you are thoughbeit

File: 1722538843546953.png (154 KB, 680x680)

154 KB PNG

>thread hit bump limit
time to shitpost!

>>103033425
rust solves memory leaks

>>103033245
Well to start with there's the different subsystems of an engine:

Making a 3D animation system with layers, blending, tweening
Making a 3D physics system
Making a render pipeline. 3D renderers are well documented, but actually designing your content pipeline, material formats, and so on is open to interpretation and requires domain knowledge
AI systems. Navmeshes or grids, optimizing these is quite involved and also requires domain knowledge

These are the "easy" problems because you can compartmentalize them. They can still take you literally months of work. The hard part of an engine is bringing it all together. What's your entity system going to look like? Are you going to use embedded scripts? The answer to your question depends on your game, and there's lots of blind alleys you can go down because architecture isn't anywhere close to a solved problem. I first opted to use a system which I think is like Godot (I haven't used Godot) where everything is a hierarchy of nodes and you attach nodes together to form game objects. But I found that approach to be far too generic. Also tried working with embedded scripts but the found the cost of writing glue code made the whole thing not worth it. When you're making a 3D game of even moderate complexity there's so many things to consider when you're putting the whole thing together it borders on untenable for a solo dev to manage