Forums » General

Sector and asteroid sizes of the future

Jan 25, 2023 HunPredator link
Currently, sectors are not very big. The range at what rocks are still present is very small usually, 15-30 km from the center, or the opposite, nothing in center, with roids spread out further away from it (bractus c5 is a good example of it, and dont get me wrong, i love that place).
While its not the case with every sector (there are some places with very long strips of asteroids), most sectors are like this. But what about going full on big?
Like by having very big sectors, potentially 200+ kilometers across, filled with sparse, but huge roids, or heck even packed to shit with small, outrageously long strips of rocks (some places in deneb really feel like that).
It was mentioned previously that in the future, render range could be basically endless, but how big you can go, before having game stability issues, with so many asteroids visible?
How rendering potentially hundreds of big or thousands of small ones would go, from dozens of kilometers, opposed to the like 25-ish kilometers we can see nowdays?
Also, would the radar range be increased, if this gets done, to like 15 kilometers?
And how big asteroids can truly be?
Some of them are already big, like some in bractus or sedina, but how much bigger they can be made?
I like the idea of still playable, but enormous sectors, but is there are limit on how big they can be?
Jan 25, 2023 Luxen link
Have you seen the unknown system, HunPredator? You should check some of the sectors there 😉
Jan 25, 2023 HunPredator link
I did, explored it quite a lot.
It still feels small... idk why
But compared to main space, yes, its big
Jan 25, 2023 incarnate link
Like Luxen says, check out the unknown system. (Saw your post, after making my response).

There are a lot of different technical challenges to different aspects of what you're asking about, and we've been working on many of them for a long time.

I certainly want denser, more complex and diverse sectors and environments. But, I also had to wait for the rendering capabilities of the average device to be able to speed up a bit as well (particularly with mobile).

There isn't much of a real limit to "how big can you go", except like.. pixel resolution of your screen, and the relative detail you can perceive (although there are trade-offs to different technological choices).

I've also been looking at intra-sector navigation options as well. Because once sectors become more extreme in size, then navigating them becomes more of a challenge. And, more server-side computational resources will be used by a single sector, if it has a lot more "going on within".

Anyway, this is something we're working towards.
Jan 25, 2023 HunPredator link
New intra sector navigation options? Can you give more details? I wonder what navigation would look like in the future.
Also, the Unk. System is much more modern compared to the main ones. Lots of bots, new asteroids and their debris, and a generally newer look
I went there to bombard guardians several times before, and i like it.

I wonder how very big sectors would look like with hundreds of rocks, with limitless view range and god knows how many bots (even without them)
Certainly interesting

Will there be test sectors, in either the unk. System, or bractus, what we can see these for ourselves in the near (or bit farther) future?
Jan 25, 2023 Sid123 link
Bractus P1 was a test sector for NPC flight AI iirc. Has a really dense and huge asteroid field.
Jan 25, 2023 incarnate link
Development of the kind we're discussing is "very difficult", which I say only to underline why it's such a "meticulously staged and measured" process in which one incrementally rolls out features, then does a ton of testing, then another feature, so on and so forth.

For example:

1) The sector has to be able to generate pathfinding data and fly NPCs around very complex asteroid fields, without using a ton of server-side resources (CPU or RAM).

2) Sectors need to load very quickly. The sector is expected to "be online" before the player's jump-effect is completed. We don't want "loading a million asteroid positions / states" to exceed this, which requires some thoughtful implementation.

3) Anything can happen in any sector, and we have a "true MMORPG" expectation of supporting "at least" 500 concurrent players per sector; we also have no idea what those players might do, or how impactful their gameplay choices might be. So.. performance scalability is important.

4) Your instrumentation and analytics on your server components has to be really good, so you're able to understand "what is happening" when it hits the limits. Because, "players be crazy" and eventually someone's going to do some weird stuff that's a problem (a-la "crack botting").

5) The game client has to be able to render huge numbers of asteroids, across all supported devices and platforms (including mobile and VR, which is not a small challenge).

6) Individual-asteroid scale is also an issue, as GPU-rendering of "extremely large objects" has different efficiencies from "extreme quantities of smaller objects". Open-world "terrain" engines, for instance (think: normal FPS games), are designed differently from open-world "space" engines; they use different techniques based on their respective strengths and requirements.

7) Any game-feature based around "individual asteroids" (like, say, mining) needs to be able to scale to more extreme environments.

8) Everything has to be periodically rolled out and tested in a way that gains exposure and testing by real-world users, but without breaking the whole game (or important parts thereof) if something goes awry.

Hence, you make a "secret" system in the middle of nowhere, and put some drops and stuff in there to attract a little attention, and see what blows up.

Then, over time, you add more things. But, at the moment, the renderer changes (#5, #6 above) are not really in-place. I mean, they're partially in place, but not everything needed for extreme draw-distances or very high counts. There isn't much point in trying to push extreme asteroid counts on the server, if the client isn't really ready to render them.

But, as of this week, we're also testing different spatial partitioning algorithms on the Sector (server-side), to see if we can get better performance than what we've been historically using. This directly bears on the performance of any system that needs to deal with large numbers of objects and answer questions like "is asteroid #859,944 in-front-of or behind asteroid #989,123, from my current position?". And then might have to a version of that question a thousand of times per second, for some reason (this is just a vague example to get across the idea).

Anyway, yes, there will most-likely eventually be more complex "public test spaces" where we see how things work in practical reality for real-world users, but in a way that's pretty "optional" and doesn't break the game for people if they can't participate.

One might recall, when we released the Debris Field test sector, a bunch of mobile players were pretty upset about how slow it was (similarly, there were complaints about the capship pirate sector)..

But, then we made it faster, and did so in a way that made the client's performance much more stable for the future. That was the point. There'll probably be more things like that.
Jan 26, 2023 HunPredator link
Thanks for the answers again, Incarnate