Game Discussions Star Citizen Discussion Thread v12

CIG. CIG never changes...

1648716082939.png
 
1 - server meshing is not here. The actual server architecture can't handle more than Stanton+50 players.
2 - tools showed during the years were (for some still are) not mature enough to be used in an industrialization process with non technical employee.
3 - planetary tech was heavily worked till 2020 and at each iterations transformed the geographical structure of planets, forcing the manual relocation of most of the manually placed structures. Mass producing planets while knowing you'll have to redone all the manual placement of asset on each them at each iteration of the planet tech is a no go.

Where is the Quanta?
 
They showed off their super fast content creation software and skills 6 years ago. Since then, even going by those "old" techniques, they should have had all systems already going. So, why haven't they?

1. It doesn't matter how much content they have, if it breaks the game to have above a certain number of items, that wont change till their engine can handle it i.e. they cant even run 2 systems at the same time, let alone 101.

2. They dont need to provide you any more content, since the less they do give, the more money you give. Every new ship thats bought is more time to sit around and do nothing.

3. All that content creation software doesn't do what they want and essentially is horsecrap and wow, noone else even bothers to make/use similar software for games.

Do you mean those "tools" for dropping crabs and furniture on maps?

I was looking for the video of that but can't remember when they showed it.
 
CIG use proc gen to create static assets. Those assets then need to be loaded into memory all the time as people move around. Where are they stored? On disk or from the server. Some other games use proc gen to generate stuff on the fly from the seed. So the same stuff is always in the same place, but it takes a bit more CPU overhead as opposed to disk/network/memory overhead.
and it helps greatly if you're multi-threading that part of the code...although the band analogy comes into play then.
 
I think the more interesting questions at this point still centre on their approaches to planets though. IE are there costs to the techniques which make them a poor fit for scaling to 100+ star systems etc? And would those costs have been avoided by the 'no brainer' use of classic proc gen surface generation instead?

I'm won't be touching the more specific questions. :) All good questions though.

But chewing the fat on the more general questions above... I cringe at the methodology CIG are using. And the way I see it they have two main issues (and many others):
1. They're planning about 400 planets/moons across their 100 systems. Assuming those textures/terrain can't be part of the downloaded client software, that's a ridiculous amount of continuous streaming / bandwidth. It will be interesting to see if some of that data is held in local (client) data caches, eg. a player's most common or recently visited planets fetched from local SSD.
2. Their multiplayer networking is server-client, another huge bandwidth resource hog.

My engineering teeth are clenching just thinking about the running costs. But this will all be necessary to deliver CRobber levels of fidelity. They can't use pure procgen for planet generation and expect to do what they promise. But all that streaming data and running costs... although cost may not be a factor if their funding model continues to-the-moon (cough cough).

Looking at ED on the other hand, and again from an engineering perspective, ED does it exactly as I would do it, lean and efficient. The planet procgen is all performed client side, and the multiplayer networking is P2P. I know players hate P2P , but it's unjustly maligned IMHO. Handing over all that work (and most of the bandwidth costs) to the clients is exactly the solution I'd be considering in a first design, and only ditching it if more performant methods are needed.
 
Grab your popcorn, and if you're a monthly-pay-to-tester or $1k+ backer put on your Big Boy Testing Pants, because this 3.17 PTU Wave 1 is gonna be spicy! 😃🍿🌶️

That can't be true, LittleAnt assured us every build was more stable than the last.......
 
I bet their shareholders will be happy and the Steam reviews stellar. :ROFLMAO:

Nah, the motivation is distinct here, oh Elite-smitten one.

CIG push out 'quarterly' deliveries (even hurrying them to stay with that cadence), not due to impetus from their corporate structure.... But because such things keep players engaged in the short term. Even though quarterly live releases are obviously not a great way to build features over the long term...

Makes you think doesn't it :unsure:

(Despite having a structure that should allow for thoughtful long-form design, they've ended up with a development cadence which leans more towards instant gratification ;))

---

PS: CIG's 'shareholder pleasing' stage comes if/when they force out Squadron 42 in a sub-standard state to attempt full ROI for their investors ;)
 
Last edited:
Nah, the motivation is distinct here, oh Elite-smitten one.

CIG push out 'quarterly' deliveries (even hurrying them to stay with that cadence), not due to impetus from their corporate structure.... But because such things keep players engaged in the short term. Even though quarterly live releases are obviously not a great way to build features over the long term...

Makes you think doesn't it :unsure:

(Despite having a structure that should allow for thoughtful long-form design, they've ended up with a development cadence which leans more towards instant gratification ;))

---

PS: CIG's 'shareholder pleasing' stage comes if/when they force out Squadron 42 to properly pay back their investors ;)
It is staggering development, you hater.

Ant have you found the Quanta? What's taking you so long?
 
Back
Top Bottom