Unity or Unreal for making a 3D game?

Unreal then.

It's fucking free

How can it be free if it's based on a licensed engine?

it's free, you will only pay if you use some amazon servers services

amazon can do whatever they want with their copy, also in this point is almost another engine

>amazon can do whatever they want
Oh, right

unreal is overkill for most stuff and they take a chunk of your profits meaning they partially own whatever you create.
Unity seems like the better choice for indies.

use godot

I've yet to see an indie game made in Unreal that was well optimized.

Why? Is it hard to optimise for Unreal?

It depends how many people in your studio. If a lot, use Unreal. If few, use Unity.

Not necessarily, but people are just shit at it. It's definitely the engine that's easiest to throw together half-decent looking visuals in, so it attracts a lot of incompetents who don't realize putting 4k textures everywhere, not implementing LOD switches or mesh culling, and putting shadow resolutions at their maximum is a bad idea.

yes, unreal indies run like shit

Torque3D
torque3d.org/

godotengine.org/qa/49206/get_simple_path-not-returning-shortest-path-3d
godotengine.org/qa/43645/simple-navigation2d-pathfinding-issue

Attached: godot.webm (1024x600, 1013K)

The engines are almost the same now. There's really only a few differences here and there, and most of that comes down to C# and C++ and how they interface with the engine systems.

its open source just fix it yourself!!!

>IT HAS THESE 2 SMALL ISSUES LITERALLY UNFUCKING USABLE!!!!

What's more interesting is where the two engines are going.

Unity with its data oriented tech stack, and Unreal with it's physics destruction and immense profiling.

It'll be interesting to see if you can even leverage ECS for gameplay code, vs if Chaos can even run performantly.

Unitys data orientated implementation is so fucking over-engineered
you literally have to use a different language embedded with in C# to get any use out of it
It doesn't even give you any big performance benefits unless you have tens of thousands of objects

God I want to make an engine so badly lads. Why does it have to be so hard? Why do I have to be so lazy?

it's actually one issue (so far) over both 2d and 3d
and yes it's unusable because of that

Easy mode: Unreal
Medium mode: Unity
Hard mode: Panda3D

I made one and it took me about 5 years

>turd vs shit sandwich

I made one and it took me about a year

Gameplay code is always all over the place. You can't get any good cache coherency out of it.
Like you said, it's good for making Super Monkey Ball with 30,000 balls, but, I don't know.

I know they're struggling with it, because it doesn't extent much outside of transforms, rotations, mesh rendering/instancing/batching and other linear updates.

What makes you say that?

That's literally what I've been trying to tell anyone who buys into the ECS meme but nobody else seems to get it

Any advice? I want it to be clean without some of C++'s shittier features and bloat. I'd use C but that would take even longer.

I'd recommend using ActionScript

This is how it really be.

Attached: Untitled.png (814x479, 20K)

What type of engine are you trying to make?
The best advice is not to make a generic one, and to make one that fits your requirements

You'll die if you do it. It's literally the hardest thing you can program, that being a fully capable 3D game engine. No joke.

exactly how does that issue make it unsuable?

First off, don't focus on making an engine, focus on making a game where you happen to be making an engine alongside it. Otherwise you're going to waste a lot of time. Read up on architectural concerns of game engines (Jason Gregory's "Game Engine Architecture" is probably the single best resource on it if you want a decent overview, though you will need to supplement it with other readings), and then read up on anything you need specific deeper knowledge of (3d rendering, etc). Then dive in to the source of some existing open source engines (id Tech stuff tends to be good, some of the stuff in them tends towards being out of date, especially if you go back to Quake 3 or older, but even then there will be some value) to see how things work in practice, maybe try to do some work in whatever engine(s) you're messing around with. If you're using C++, Read Effective C++ and Effective Modern C++, choose a moderately modern but not bleeding edge standard (C++11 is well supported, and is a good choice, you might be able to get away with something newer if you only care about targetting specific platforms), and then come up with a reasonable subset to target (there is plenty of advice on features of C++ to avoid for game engine shit out there, just search it up). Then just like make game.

>File that I was working yesterday disappeared from the machine
I'm scared of working on my project, backing up the important files.

Attached: 1548826742073.jpg (960x854, 99K)

>not using version control with multiple backups to begin with
What the fuck is wrong with you?

Money, and low HD space.
Was an image file,and a directory, that disappeared.

That book looks good. Know a decent place to find the third edition? I've found the second but looks like some good stuff got added.

Next make your virginity dissappear lmoa

Use Github

Nothing much got added, except for this.
gameenginebook.com/resources/ConcurrencyLecture.pptx
The book takes everything in that slide deck and goes over it again. You're not missing much.

The stuff that got added in the third edition is nice to have but not really essential. You'll be fine using 2e if you can't find the 3e.