Need help with procedural generation

I thought that was only supposed to happen at (2^31) or -(2^31)

 

No, imprecision of floating point numbers is always a problem, more the further you get from 0. You're probably thinking of the limits for signed integers.

 

But it's especially a problem if it's that bad at 500k already.

 

It starts to show already around 10k / 20k.

Jump to 4 min marker, there is explanation for your problem.

 

That's a very good video.

You can't have infinite precision in 32 bits. I don't know if you can find an arbitrary precision library compatible with Unity, but they will probably end up taking a lot of space in memory.

But yeah, do the thing in the video. It may be more of a hassle, but it'll be worth it in the long run.

 

OK. Thank you, I will do that.

 

OK so I have done some work on this. Firstly, for brainstorming, I created a project in Scratch:
https://scratch.mit.edu/projects/70031266/#player

That's what it's supposed to be. My system is generating everything but the co-ordinates correctly. The co-ordinates are almost there, but not spiraling, just going out in straight lines.

Code (CSharp):

1.  
2.         for (int i = 0; i < galaxy.armcount; i++) {
3.        
4.             c_x = 0;
5.             c_z = 0;
6.        
7.             c_rotation += galaxy.armcount/swirl_value;
8.        
9.             for (int j = 0; j < c_armlength; j++) {
10.            
11.                 c_x += (int)(Mathd.Cos (c_rotation) * 11); //110
12.                 c_z += (int)(Mathd.Sin (c_rotation) * 11); //110
13.            
14.                 galaxy.arm_waypoints[i,j] = new Vector3d (c_x+x_offset,y_offset,c_z+z_offset);
15.            
16.                 c_rotation += (galaxy.armcount/c_armlength)*swirl_value;
17.            
18.                 //log positions
19.                 Debug.Log ("X: " + c_x + " Y: " + c_z);
20.            
21.             }
22.        
23.         }

The values it logs are so (input in the Scratch project, that's what it should look like):
Arm Length: 6
Arms: 31 (I realise 31 arms in a galaxy may be unrealistic)
Swirl value (bend_factor in my Scratch project): 4

This is what it creates:

 

I'd probably switch your loops, so you're doing one "disc" at a time, moving outwards, and rotating that disc a bit each time. It could also make it a bit more intuitive if you do it in polar coordinates, and translate them later.

 

I don't think you understand that you will have to write an entire physics wrapper/simulation once you get out of Int32 ranges. That is a massive undertaking. Once again, just one solar system, in astronomical scale an invisible dust mite on a dust mite on a dust mite, brought an incredibly skilled team to their knees at times. These numbers you present are absolutely ridiculous in real terms.

 

Nope. I didn't.
Oh well. I can keep Int32. I'll upgrade if I absolutely need to

 

You should watch the entire Kerbal Space Center video.

They present solutions to many of the problems they ran into while simulating our local solar system. What you propose would be tenfold.

 

I've been working on my galaxy generator. I managed to break it and I am unsure what the problem is.
It says this:

Code (CSharp):

1. NullReferenceException: Object reference not set to an instance of an object
2. GalaxyGenerator.GenerateGalaxy (Vector3 input_position) (at Assets/Scripts/GalaxyGenerator.cs:233)
3. GalaxyGenerator.Start () (at Assets/Scripts/GalaxyGenerator.cs:78)

Here are the 2 areas causing errors:

Code (CSharp):

1. GalacticChunk galaxy_1 = GenerateGalaxy(new Vector3 (-250, -250, -250));

Code (CSharp):

1.     Galaxy GenerateGalaxy(Vector3 input_position) {
2.  
3.         Galaxy c_galaxy = new Galaxy ();
4.         c_galaxy.min = input_position;
5.         c_galaxy.max = new Vector3 (input_position.x+500, input_position.y+500, input_position.z+5000);
6.  
7.         Vector3 position = new Vector3 ((input_position.x + 250), (input_position.y + 250), (input_position.z + 250));
8.    
9. --->    int x_offset = generator.generate_int (2, (int)position.y, (int)position.z);
10.         int y_offset = generator.generate_int (2, (int)position.x, (int)position.z);
11.         int z_offset = generator.generate_int (2, (int)position.x, (int)position.y);
12.  
13.  

It says I haven't set input_position as an initialised Vector3 when I have

EDIT: It didn't mean Vector3. I rewrote the sequence of generating, and forgot to initialise the number generator. *facepalm*