Jump to content

The merits of being able to choose asteroid traits


Recommended Posts

First off, huge thanks to @Cairath for creating and maintaining this awesome tool.  I use it every time I start a map, because I refuse to let RNG dictate the difficulty of my map.

With that said, once things are finished, we are going to be left with a completely unpopulated tool.  So I encourage anyone who's interested in using the new seed DB to download and use the mod to help re-populate, once things are ready of course.  We also need to make sure that anyone contributing knows that modifying the contents of the worldgen/traits folder will make their contributions invalid.

Looking forward to using the new tool to help find a new home for my duplicants!

3 minutes ago, Junksteel said:

@Cairath Hey! I want to say thank you for your work. I got so used to it man... not having that quality of life beauty in the preview made me feel naked.

Sorry! It's taking quite long and collided with my vacation x( 

14 minutes ago, nessumo said:

@Heister this is really funny... Not working for me (-: Your seed and my test seed on Arboria attached. I am running LU-351503 on OSX 10.13.6. Any idea?

Works for me on same version running on Win10.  My guess is either it's a difference with OSX, or you modified your worldgen/traits files.

I am running vanilla game - no mods, not modifications (AFAIK). And please don't start standard flamewar - we all know in our hearts that macs are better (-:
Anyway, when I know the twist it is not a problem. It would be good if any other mac user could check if it is mac thing or "my" thing, tho.

Just now, nessumo said:

I am running vanilla game - no mods, not modifications (AFAIK). And please don't start standard flamewar - we all know in our hearts that macs are better (-:
Anyway, when I know the twist it is not a problem. It would be good if any other mac user could check if it is mac thing or "my" thing, tho.

I have the same problem, also with OSX 10.13.6.

20 minutes ago, nessumo said:

we all know in our hearts that macs are better

At being overpriced paperweights, maybe. 

Anyway, it wouldn't be the first time that differences in OS have caused the RNG to have different results. 

29 minutes ago, Yunru said:

Anyway, it wouldn't be the first time that differences in OS have caused the RNG to have different results. 

Anyone more experienced could explain this to me? Because I thought seeds are used in deterministic function to generate the world. How is RNG involved here?

14 minutes ago, nessumo said:

Anyone more experienced could explain this to me? Because I thought seeds are used in deterministic function to generate the world. How is RNG involved here?

It's deterministic as long as the used method is the same, which is not always the case if the implementation used isn't custom. It still *is* RNG - it's seed based, as there is never true randomness, in reality it's what we call pseudo rando  - there's always a seed based calculation involved, in this case it's visible to the player, but in in instances where it should be "random" various things can be the seed, such as current timestamp, seconds the program's been running, anything that would generate "randomness". 

There's differences between algorithms used in Mono and raw .NET framework, and even between various .NET framework versions. For the seed validation /trait search purposes I've taken the algorithm from the decompiled sources attached to the game (because the default system.random generated different results) and I'll be validating seeds against that algorithm.

 

It certainly would be easier to manage and nicer for the community if all platforms generated the same results, but sometimes getting that to work might be a pain in the butt. Would be nice if someone running linux could post their results for that seed

Maybe I wasn't precise - could you please assume I am software developer, I know that "seed" name was taken from pseudo-RNG generators etc etc... But in every game I know generation of the seed mimics randomness, but for given seed world generation is later deterministic. I thought here is the same - for given seed (and game version, because obviously code can change) geysers, planets, biomes etc are the result of fully deterministic function. There is a layer of "randomness" also, for example every time you generate the world plants have different growth, but when you say RNG could be possibly involved in this bug I am lost.

Tldr: It's RNG but it's deterministic if you use the same seed. It's not the same across different OS because most likely it's not using the same algorithm orrrr system architecture is getting in the way, but that part is a bit beyond of what I can correctly explain

 

@mathmanican Looks like what I've expected, same result on Linux and mac, different from Windows. Not surprising... except for the fact that someone just had test it for me on ArchLinux and got the same results as on windows. Could you check what seed 1 gives you please?

Unfortunately, as this is exactly the part that interests me. I would be very surprised to learn that algorithm is different on different OS, therefore I guess that it must be, as you said, "system architecture getting in the way" of fully deterministic algorithm. That is something unexpected! nvm, we will probably never know... 

2 hours ago, nessumo said:

@Heister this is really funny... Not working for me (-: Your seed and my test seed on Arboria attached. I am running LU-351503 on OSX 10.13.6. Any idea?

Yeah I know why this happens.

This is an issue with how Klei made the traits list before getting them for the game. They just read all the files on a list, and I thought they obeyed alphabetical order, which seems to be true for anyone running ONI on Windows. But when reading the Mac/Linux files it's not ordering them the same way before getting traits randomly.

This pretty much means that seeds between people running Windows/Linux and Macs might no be interchangeable :(, this is true even for the TNI tool.

Since my tool runs on the web, I don't have access to your files which means I can't read the files on your Mac on the same order as in Windows

The only way it will work is if I compile a standalone version for you guys and you run on your computer.

Give me some time and I think I can get a version for you.

edit: Maybe it's this or maybe I'm wrong, let me just run some tests before.

46 minutes ago, nessumo said:

Anyone more experienced could explain this to me? Because I thought seeds are used in deterministic function to generate the world. How is RNG involved here?

You can use a (C)PRNG in your software or you can use the OS one. Does not make any sense for Klei to use an OS generator. Much more likely this is a generator from the game engine and there is some difference on the Mac. Has anybody tried seed portability to Linux?

Although the geodormant/geodes thing actually looks like it might be file ordering (this is... silly), there have been major biome differences in map uploads to TNI before traits were a thing. Once I am done updating (bear with me, it's still going to take a while) we could compare it in detail. If it turns out to be file ordering, I think we could ask Klei to fix that quirk.

On 7/15/2019 at 3:11 PM, Junksteel said:

@Cairath Hey! I want to say thank you for your work. I got so used to it man... not having that quality of life beauty in the preview made me feel naked.

Big thanks from me too.

1 hour ago, Cairath said:

Could you check what seed 1 gives you please?

I used seed "1" and got the following:

Spoiler

frz1.png.ca7c4105c632e879280c5fe2bfe4594b.png

bad1.png.dd33687ef8832ea5d975c488c82103bc.png

frst1.png.274b4e1b16a4a81be631397c85f6c43a.png

lush1.png.b3197438993a8af28587c512ba70ad02.png

sndst1.png.e7bb6d30e69d293b7e8d0c8f2b35b135.png

I stopped taking screenshots, as this demonstrated to me that the traits are definitely seed dependent (and independent of type of world).

 

Archived

This topic is now archived and is closed to further replies.

Please be aware that the content of this thread may be outdated and no longer applicable.

×
  • Create New...