[Megathread] Star Citizen

[Jastall]

Do you actually follow development at all, because if you actually payed attention you would know the AI will be handled in the universe simulation quanta, a background simulation that will deal with everything and only load in whatever is needed close to players, they have shown this system in action already, the AI in the alpha is just for atmosphere.

"They can't get a simple script to work, this means they'll be able to implement hyper advanced AI then!"

Some people are begging to be fooled I suppose.

[Kyanion]

Fair enough, this all stemmed from people claiming it was pushed to 2023. We don't have enough information to make that claim yet. So that claim is wrong. If we agree with that there is nothing else to discuss.

Nah just a roadmap showing the bar almost to Jan 2023. That must mean nothing apparently.

[kenn9530]

For more advanced behaviors, yes. But very basic stuff like pre-set paths and animation sequences right now don't require quanta and should be fairly simple to implement.

Also: https://robertsspaceindustries.com/c...port-June-2020

You should tell Chris Roberts that all AI will be handled by quanta, because he seems unaware of this as of at least June 2020, talking about different AI - including bartenders! - and narry a mention of quanta.

The servers are under a heavy load so that will cause commands to not be issued properly because the server is doing too much, they serve the current purpose.

[Edge-]

The servers are under a heavy load so that will cause commands to not be issued properly because the server is doing too much, they serve the current purpose.

With 50 people?

Then why was CIG wasting so much time on that AI if the servers couldn't handle it and they need server meshing and quanta to be in place? Didn't they spend years talking up the AI bartenders only for them to release with all the excitement of a cold, moist blanket?

[kenn9530]

"They can't get a simple script to work, this means they'll be able to implement hyper advanced AI then!"

Some people are begging to be fooled I suppose.

The AI does work, there are hundreds of AI that the server has to handle, so if a few like doing weird things what does it matter since its just the basic system to add some life into the server and nothing more, when a system is working near max capacity lower priority commands will be ignored.

- - - Updated - - -

With 50 people?

Then why was CIG wasting so much time on that AI if the servers couldn't handle it and they need server meshing and quanta to be in place? Didn't they spend years talking up the AI bartenders only for them to release with all the excitement of a cold, moist blanket?

You do realise the server doesnt just handle 50 ppl, it handles every station, planet, moon, ship all at the same time and at least a few hundred AI, as soon as 1 person logs into a server it has to currently load in everything not just parts of the system.

[Edge-]

You do realise the server doesnt just handle 50 ppl, it handles every station, planet, moon, ship all at the same time and at least a few hundred AI, as soon as 1 person logs into a server it has to currently load in everything not just parts of the system.

Yes, I'm aware of that. If the server is buckling under the strain of simple solar system simulation, that's actually way worse for the game, dude.

Again, if the servers are already overloaded and, according to you, quanta is needed to address AI in the game, which itself requires server meshing to be functional, why did CIG spend years of time on NPC AI that will be "replaced" by quanta?

Or am I misunderstanding you?

[kenn9530]

Yes, I'm aware of that. If the server is buckling under the strain of simple solar system simulation, that's actually way worse for the game, dude.

Again, if the servers are already overloaded and, according to you, quanta is needed to address AI in the game, which itself requires server meshing to be functional, why did CIG spend years of time on NPC AI that will be "replaced" by quanta?

Or am I misunderstanding you?

When server meshing is in then there wont be much load on the servers as there is now, but knowing what a server can deal with is essential for maximising the game.

Quanta will handle the AI controls on its own in the background without impacting on the server and only load in AI when a player is near by, if a player moves away that AI wont be on the server but will still be doing its thing as data in quanta.

The work with AI still needs to be done its just going to be a different system that handles the load in the end.

[Yelmurc]

Nah just a roadmap showing the bar almost to Jan 2023. That must mean nothing apparently.

Do you understand that static server meshing could be delivered before the end of the year and the progress tracker is showing all work on server meshing in general?

[Edge-]

Do you understand that static server meshing could be delivered before the end of the year and the progress tracker is showing all work on server meshing in general?

Everyone does at this point, but whether they can even get static server meshing done by the end of the year is pretty questionable, again, given their track record and recent comments about not wanting to forecast more than beyond the next patch because too many things were slipping and that was pissing folks off.

[Xeenith]

fool me once, shame on you. fool me twice, shame on me... and some here are tired of being fooled dozen of times.
I for one will believe some tech is ready WHEN it's implemented and works without any major bug, until then, road map and delivery dates are just hot air, and I'll stay far from that!

[Thestrawman]

Do you understand that static server meshing could be delivered before the end of the year and the progress tracker is showing all work on server meshing in general?

It could. It could be delivered tomorrow. Or 2024. You have no way to know either way and your continual hissy fits about peoples very realistic estimations does nothing to prove yourself right.

[Kyanion]

Do you understand that static server meshing could be delivered before the end of the year and the progress tracker is showing all work on server meshing in general?

Do you understand that static server meshing could never be delivered as well? It might be this year, it might not. Signs point to not considering CIG's track record. What is your point exactly?

[Gorsameth]

When server meshing is in then there wont be much load on the servers as there is now, but knowing what a server can deal with is essential for maximising the game.

Quanta will handle the AI controls on its own in the background without impacting on the server and only load in AI when a player is near by, if a player moves away that AI wont be on the server but will still be doing its thing as data in quanta.

The work with AI still needs to be done its just going to be a different system that handles the load in the end.

Which very sums up what everyone's problem is with SC currently.

The game cannot exist without Server Meshing and 'Quanta' and we're almost a decade into development.

IF they can solve the magic bullet before the backers run out of money to throw in the money hole there is hope. But that's a big IF.

[DarkAmbient]

I saw an interesting post on Reddit, comparing the timelines of Chris Roberts and Hideo Kojima and what they both achieved in that time.

In the span of 11 years, Kojima developed the brand new FOX game engine, released MGS: Ground Zeroes and MGS: The Phantom Pain, released the Silent Hills demo P.T., got fired and then built a brand new studio, and then developed and released Death Stranding.

In the 12 years since pre-production on Star Citizen began, Chris Roberts built his studio and Star Citizen is still in alpha, using a heavily modified version of CryEngine.

[kenn9530]

Which very sums up what everyone's problem is with SC currently.

The game cannot exist without Server Meshing and 'Quanta' and we're almost a decade into development.

IF they can solve the magic bullet before the backers run out of money to throw in the money hole there is hope. But that's a big IF.

server meshing is not an issue it will be put in game when everything else required for it to work is in place, also the universe simulation to create a living universe has been shown in operation and is not required for SQ42,

[Mamoraki]

From a thread on the forum of robertsspaceindustries.com

Let's say i'm in a spacestation with about 50 other players wich would be serverinstance 1 for example.

Then there is a Spaceship near the spacestation with 50 other players in it wich would be serverinstance 2.

Everyone can see each other. That means i'll get all the state updates from all players from 2 different servers at the same time?!

yes

Reading that, how is that going to work at a reasonable speed with regards to latency just between servers (let alone any network switches/routers(/firewalls)), you would want most if not all servers as close to each other as possible preferably connected to each other directly without a switch. Sounds like expensive servers to have.

And I doubt that it is possible when they need more servers (on demand adding servers) that those are in the same datacenter as the existing ones, those would introduce more latency (?).

In a space battle, player X fires on player Y, server of player X needs to know where player Y is.. Now you can scale that 50 up to 500 or so, that just means a lot more data to throw around from/to instance.

So, what if 1 person moves from instance 1 to instance 2 (boarding the spaceship)

How is a database server going to handle that with hundreds of instances (on top of everything else)? (100 people isn't anything special, but I imagine we are talking about hundreds of thousands of possible players on hundreds of servers with hundreds of data transfers). Mostly concerned about speed here. (EG, server 1 needs to unload any unnecessary data (commit to DB) while 2 needs to load it, without the player(s) noticing)

Wouldn't every instance need know on what server another player is? Especially if they can see each other?

I guess every instance is 1 server, not 10 instances on 1 server, if not, that 1 server should be able to handle all those people without the extra overhead of the instance..

[Kyanion]

server meshing is not an issue

Server meshing is a very big fucking issue.

- - - Updated - - -

The issue with the tracker, at least the last iteration, is that they claimed it'd only show what they're super confident can be delivered when it says it will be on the tracker. I 100% see why people would assume that those timeframes are then expected/"promised." So missing them then becomes an issue for both sides, because one has said "this is what we can do," while the other side is now let down and has to wait for X feature even longer. CIG are their own worst enemy.

The sad part is people will keep tossing them money. I'd love to see the fallout if they fail to get server meshing in by the end of the year, but sadly we both know the invested backers will suck up more lies.

[Val the Moofia Boss]

CIG is now developing navmeshes so that NPCs/wildlife can walk on planets outside of cities. The planets/moons are too big to generate navmeshes for them, so the team is working on dynamically creating navmeshes. If the player lands on a planet and NPCs/wildlife are around, then the servers will start creating a navmesh for that area.

Also this week we got a livestream of the devs making a trolley that players and NPCs can push around. Nice tidbit. That being said, the livestream was unfortunately named "The Pyro Proposition", which implied that it was going to be actually about the pyro star system, not a trolley.

[Edge-]

If the player lands on a planet and NPCs/wildlife are around, then the servers will start creating a navmesh for that area.

This sounds like another layer of complexity for servers to handle on top of quanta and dynamic server meshing and a ton of other shit...why does it seem like they keep making things more complex as time goes on?

[Easo]

server meshing is not an issue it will be put in game when everything else required for it to work is in place, also the universe simulation to create a living universe has been shown in operation and is not required for SQ42,

I am preeeeetty sure you were the one saying that after server meshing jesus tech will be in the game everything will be built upon the top of it. Now you are saying that other things have to be implemented first.
So which is it?