The ability to do this would benefit any multiplayer game, hence why it seems CIG's future business plans are to licence this engine technology to others.
AFAIK, CIG has never once said they plan on selling their tech. That's just randos wildly speculating. CIG doesn't own lumberyard or CryEngine, and it's not a popular enough engine to make much selling tech stuff for the engine.
MMO "servers" haven't been running in actual single server computers for a very long time. The very idea of shards/realms/servers is a remnant of the earliest games in which those problems were actually a thing, and games still have them because of a sense community and the tradition.
Are you sure CIG didn't actually reinvent a technology that everyone else's been using for a good while already? I'll grant that no other company has been insane enough to want to simulate the entire fabric of reality in a videogame, so that does make server load a tad heavier.
EDIT: oh, does crytek even still exist? Is CryEngine even still a thing outside SC? I guess they took so long with this that UE5 has left cryengine very far behind in terms of looks and optimization. i wonder how fast SQ42 would get done from scratch starting today with UE5. I guess if they just had made it back then we'd be up to SQ45 now
The ability to do this would benefit any multiplayer game, hence why it seems CIG's future business plans are to licence this engine technology to others.
AFAIK, CIG has never once said they plan on selling their tech. That's just randos wildly speculating. CIG doesn't own lumberyard or CryEngine, and it's not a popular enough engine to make much selling tech stuff for the engine.
Unless I am misremembering prior bouts of thread activity, haven't they already been licensing out other techs they've developed to have an income stream outside of nerds buying ships? It's not exactly a stretch to assume that this tech would also be up for grabs, assuming it's not somehow locked into their specific fork of cryengine.
It was that somehow, from within the derelict-horror, they had learned a way to see inside an ugly, broken thing... And take away its pain.
Warframe/Steam: NFyt
I made a joke almost a decade ago about training my kids to man turrets on my constellation. All of my children are old enough to perform this task and not all of them were even born when I made that joke. Holy fuck, this game makes me feel old!
MMO "servers" haven't been running in actual single server computers for a very long time. The very idea of shards/realms/servers is a remnant of the earliest games in which those problems were actually a thing, and games still have them because of a sense community and the tradition.
Are you sure CIG didn't actually reinvent a technology that everyone else's been using for a good while already? I'll grant that no other company has been insane enough to want to simulate the entire fabric of reality in a videogame, so that does make server load a tad heavier.
EDIT: oh, does crytek even still exist? Is CryEngine even still a thing outside SC? I guess they took so long with this that UE5 has left cryengine very far behind in terms of looks and optimization. i wonder how fast SQ42 would get done from scratch starting today with UE5. I guess if they just had made it back then we'd be up to SQ45 now
Even the earliest MMOs weren't single-server. Asheron's Call for example used a series of servers covering strips of land north-south across the map. Loading through them was seamless.
believe each world ran on something like 7 quad core p133s. I think they got upgraded at some point to 200mhz machines or something, but I can't say for sure as I never really dealt with ops.
I wasn't involved with the server code at all, but I will tell you a funny story. The server breaks up the world into strips to do it's load balancing. When this was first written, we were told to lay out the content such that things did not line up horizontally, so they'd be on different strips. However, turns out the server was coded with vertical strips.
Once we were live, there was one strip that often had 3 towns on it, so we needed to move at least one of them to lighten the load. So we blew it up in an event.
Your town got blown up to help us balance the server.
Phoenix-D on
+5
Options
daveNYCWhy universe hate Waspinator?Registered Userregular
I mostly ponied up cash for Squadron 42, so the past few years and all the server meshing talk hasn't really been that relevant to my interests, which is my biggest complaint. No small chunk of their backers paid money to help develop Ultimate Wing Commander, and Roberts spent the time and money working on not that. Sure, the mechanics they developed for Star Citizen will transfer, but Squadron 42 didn't need the full mining mechanics or all the walkable ship interiors they've made. It is annoying.
Shut up, Mr. Burton! You were not brought upon this world to get it!
+8
Options
NEO|PhyteThey follow the stars, bound together.Strands in a braid till the end.Registered Userregular
edited October 2023
Finally got around to watching that half hour Squadron 42 I held the line thing, I get the feeling that Chris Roberts is bad at using terms correctly, because there is normally a fair bit of work between feature complete and polish phase by the generally-accepted meanings, and he is talking like they have jumped straight from one to the other.
I didn't hop on the ride for squadron 42, my introduction was a friend going "hey this Kickstarter looks kind of neat but I am currently broke, anyone wanna buy me in and I can pay you back later" and I went "that does look neat" and bought in myself as well. Hopefully the man isn't bullshitting too hard about what progress they are at, both for the sq42 people and the part where he said polish phase on SQ42 means Star Citizen is also going to be seeing quicker progress
NEO|Phyte on
It was that somehow, from within the derelict-horror, they had learned a way to see inside an ugly, broken thing... And take away its pain.
Warframe/Steam: NFyt
Random friend linked me this. Must admit after multiple years (years!) of side eyeing SC every so often to keep vaguely abreast of it, it's hard not to get a little hype. Until I'm actually playing the finished version though, I won't get my hopes up too much.
NEO|PhyteThey follow the stars, bound together.Strands in a braid till the end.Registered Userregular
edited October 2023
It is definitely still a long ass road to 1.0, but some of the more hyperbolic wailing and gnashing of teeth can make it sound like we're still in the hangar module tech demo days where all you could do is look at a ship compared to what is currently accessible and playable.
NEO|Phyte on
It was that somehow, from within the derelict-horror, they had learned a way to see inside an ugly, broken thing... And take away its pain.
Warframe/Steam: NFyt
At the risk of sounding like an idiot, why is any of this server meshing stuff necessary to make a fun open space multiplayer sim?
Because Chris Roberts wants to pull an EVE Online and have all the players in the same world.
:edit: The specific reason this needs fancy unicorn tech is that his fancy high fidelity gameplay requires a LOT of information to be going to and from the server compared to EVE Online's movement/combat model. Currently, the entire Stanton system is run by a single logical server (I hope they're doing logical servers and not a literal single machine), so if a bunch of players decide to have a ship exploding party over at Grim Hex, all the calculations start grinding server performance down for everyone. IIRC the initial pass of server meshing would make it so that (as an example) each planet+moons is on its own server in a cluster, and as someone travels between them they get seamlessly passed around servers, and also the servers talk between each other so that you can see people on the Microtech server before crossing the invisible line that puts you on that server yourself. Dynamic meshing would let the backend shuffle server boundaries as needed to address whatever shenanigans players might be getting into.
NEO|Phyte on
It was that somehow, from within the derelict-horror, they had learned a way to see inside an ugly, broken thing... And take away its pain.
Warframe/Steam: NFyt
Finally got around to watching that half hour Squadron 42 I held the line thing, I get the feeling that Chris Roberts is bad at using terms correctly, because there is normally a fair bit of work between feature complete and polish phase by the generally-accepted meanings, and he is talking like they have jumped straight from one to the other.
It seems plausible to me that the, single player, S42 game is more or less good to go, given that most of SC's issues seem network/server related. AI going stupid, desyncs, server implosions, and latency on their ever widening constellation of data service, for example. But, as far as S42 is concerned, the guns seem to shoot, the ships more or less fly, etc.
0
Options
daveNYCWhy universe hate Waspinator?Registered Userregular
Finally got around to watching that half hour Squadron 42 I held the line thing, I get the feeling that Chris Roberts is bad at using terms correctly, because there is normally a fair bit of work between feature complete and polish phase by the generally-accepted meanings, and he is talking like they have jumped straight from one to the other.
It seems plausible to me that the, single player, S42 game is more or less good to go, given that most of SC's issues seem network/server related. AI going stupid, desyncs, server implosions, and latency on their ever widening constellation of data service, for example. But, as far as S42 is concerned, the guns seem to shoot, the ships more or less fly, etc.
Depends on how all the complex systems and lack of abstraction work when things really get going. I would be mildly annoyed if there's a mission that craps out because an AI crewmember had a pathing hiccup or whatever that prevented them from doing the important thing. Whole lotta moving parts.
Shut up, Mr. Burton! You were not brought upon this world to get it!
Hooh boy do I have thoughts about all of this, which I am now going to share.
First of all, on server meshing I'm setting my hype levels to "cautiously optimistic" from all of what we've seen at CitizenCon. For those of us paying attention, nothing they showed us was particularly mind-blowing. Yes, it was nice to see Server Meshing actually working, albeit in an incredibly simple cut-down "happy-path" scenario, but this is something we've known about for years and at this point I really can't get excited until I see it actually running on a live server. It is an exciting technology, for reasons which I will go into later, but I'm also very aware that the gap between "it works well enough to demo" and "its deployed and working on a live server that the public can log into" could still be measured in years.
On SQ42, what's in the demo they showed us was amazing and proof once again that Chris Roberts is a master of building hype, but once again we've been burned so many times that until people are actually playing it, its hard to let myself get too excited. And again, the gap between "feature complete" and "you can actually play it" could still be measured in years, and I suspect probably will be. "Feature complete" is not "complete". I will be shocked if its actually in the hands of the general public before 2025.
On CIG deserving nothing but cynicism... unironically yes, absolutely, and I say this as someone who still considers themselves a fan. Star Citizen might not be a scam, but it has been horribly mismanaged from day one, I feel like they only actually figured out what it is they're even building towards in about 2019. I have never seen a company so godawfully bad at estimating how long a thing will take to be done, and I'm a software dev, so I do understand that estimating the time things are going to take can be hard. Some guy on reddit posted a great summary of past CitizenCons, what was announced, what was demoed and what has actually made it into the game. Basically 2016 was the year of "peak bullshit" where a bunch of stuff they demoed was straight up faked (sandworm anyone?) and much of it still hasn't made it into the same, but to their credit, their estimates have gotten better over time, and they've learned to stop letting Chris get hold of the mic and promise that whatever pie-in-the-sky bullshit he just thought of will definitely be in the game within a year.
I still don’t fully understand a) why server meshing is such a huge focus for this game, and b) why successfully implementing any form of server meshing for this game would necessarily have any significant implications outside of this game.
As best as I understand it (and having come at this a few times, I do think I get the idea, because it’s the execution that seems to be the actually complicated part), there are going to be a series of trade offs and design decisions as a result of pursuing maximum seamlessness for a maximum number of people. Those decisions are going to be very specific not just for reaching that goal, but also for reaching that goal in the context of a space shooter/FPS. And any decisions about how to prioritize a specific agent or event on a given server over other servers is going to be pretty dependent on the context and goals.
So unless a dev is looking to make something pretty close to Star Citizen, I’m not sure all that much work has been done for them.
And admittedly, I’m still skeptical that any implementation is going to survive at scales far beyond current MMO instance levels without doing the EVE time dilation thing, so even allowing a higher player cap to interact doesn’t seem like it’s inherently opening up gameplay doors?
And I feel more confident that I’m not missing some tremendous implication of this technology when the features they’re emphasizing in their buzz reel are variations on ‘Press F to perform canned animation.’
Like, I’m not invested in this aside from finding the project fascinating, wanting it to succeed way back in the day, and the niggling feeling I must be missing something about this specific element that gets emphasized so hard, so definitely feel free to dismiss all this if those are your litmus tests. But if I am wrong about how this is meaningful, I would genuinely appreciate an explanation because the premise is interesting and I’m really not just here to cheer on a car crash.
Edit: To be clear, it would be fun to be wrong and we’re on the verge of a transformational technology for gaming. Genuinely, that’s exciting news. I just can’t see how that is what this is, and I swear I have tried with some real effort multiple times. Part of the problem may be the lack of details on implementation mean most explainers are speculative?
The reason Server Meshing is so interesting to me, personally, is that it allows them to take advantage of one of the biggest advances in computing of the last couple of decades and apply it to video games, i.e, autoscaling. Skip ahead if you already know what autoscaling is, but if not, history lesson! In the days of yore, when you went to a website like www.penny-arcade.com, your connection would be routed to a single, physical server somewhere that was hosting the website. In some places on the net it still works this way, but increasingly these days rather than a physical web server what you're dealing with is a VM (virtual machine); a bundle of storage, memory and CPU resources hosted on a small part of a much larger supercomputer that manages a bunch of these "pretend" computers. The advantage of this is that if you suddenly get millions of people logging in at once, in the case of PA for example because a new comic just dropped, you can ask the supercomputer if it can allocate more resources so that you can keep up with the increased load. If it doesn't have enough resources to spare it might shunt your site across to a different supercomputer with more capacity or, if even that isn't enough it can share the load between multiple. This is all entirely seamless to the end-user; two users on opposite sites of the world see the same website and as far as their browsers are concerned it is the same website, but they're probably getting their traffic routed to completely different edge servers (which are geographically closer to the user and thus load faster in the browser) mirroring the same content between each other. The reason this is/was a game changer is it meant that it was suddenly much easier and much more affordable for a site to scale up and down with demand; in the before-times if you wanted to be able to handle 10,000 concurrent users at peak usage, you'd need to have enough servers on standby to handle that load ready all the time. Now they can automatically scale up, adding new instances when there's a surge in demand, and dropping back down when demand drops again.
This works well enough in a not-real-time scenario like a website or a forum, what CIG are trying to do is get this technology to work in the context of a real-time online universe, spinning new servers up and down and passing data and players between them seamlessly.
MMO "servers" haven't been running in actual single server computers for a very long time. The very idea of shards/realms/servers is a remnant of the earliest games in which those problems were actually a thing, and games still have them because of a sense community and the tradition.
Are you sure CIG didn't actually reinvent a technology that everyone else's been using for a good while already? I'll grant that no other company has been insane enough to want to simulate the entire fabric of reality in a videogame, so that does make server load a tad heavier.
This is not that. Yes, MMOs have been using instancing since the 90's, that's nothing new, but there's always a clearly defined border between zones, usually with a loading screen. You go into a busy city and there's a gate you have to click on, fade out, fade back in, you're in the city. That fade in/out was to hide you getting shunted onto one of many instances of said city, because its where everyone starts and they can't realistically fit all of the newbies into one instance. What Star Citizen is proposing to do is dynamically split servers up into smaller chunks on-the-fly as needed, reducing load on individual servers and allowing them to share information between each other. How well this will actually work in practice we still don't actually know because no-one has ever been crazy enough to try this, and I'm confident that there are a whole bunch of edge cases they haven't considered... but in theory it could be just crazy enough to work. Or it could be exactly crazy enough to crash and burn catastrophically.
The object container system was specifically designed to work with this in mind; basically every area in Star Citizen is a container. The Hurston system is a container, full of ships, planets and space stations. The planets are containers, full of ships, outposts, and any players running around on foot. The outposts, space stations and larger ships are containers full of rooms, which in turn could contain ships if they're hangars, which in turn could contain more rooms if its a big ship, and those ships could also contain hangars which contain ships which contain people and oh no I've gone cross-eyed. But the theory is that each of these containers forms a discrete block of stuff that can be split out into its own server if needed. Too many players in the system? Spin off a server for each planet. Too many players on a planet? Spin off a server for the busiest city. Too many players in the city? Split the busiest areas off into their own servers.
That scenario I just described is still a long way off; what we saw demoed was a very simple example of three interconnected servers in a small area with pre-defined boundaries. But the fact that they're even to do that much is a huge milestone. The next step will be to introduce Pyro, hosted on its own server, and have players able to transition between Hurston and Pyro. Then it'll be to break down the Hurston and Pyro systems into discrete, static chunks, probably one server per planet for starters. That much I'm now confident that they will eventually achieve, the question is what happens when they start trying to make all of this dynamic.
At the risk of sounding like an idiot, why is any of this server meshing stuff necessary to make a fun open space multiplayer sim?
Because Chris Roberts wants to pull an EVE Online and have all the players in the same world.
:edit: The specific reason this needs fancy unicorn tech is that his fancy high fidelity gameplay requires a LOT of information to be going to and from the server compared to EVE Online's movement/combat model. Currently, the entire Stanton system is run by a single logical server (I hope they're doing logical servers and not a literal single machine), so if a bunch of players decide to have a ship exploding party over at Grim Hex, all the calculations start grinding server performance down for everyone. IIRC the initial pass of server meshing would make it so that (as an example) each planet+moons is on its own server in a cluster, and as someone travels between them they get seamlessly passed around servers, and also the servers talk between each other so that you can see people on the Microtech server before crossing the invisible line that puts you on that server yourself. Dynamic meshing would let the backend shuffle server boundaries as needed to address whatever shenanigans players might be getting into.
I think that was originally a possible, goal, but within the last year or so they've made it clear that Star Citizen will absolutely not be a single shared universe and there will in fact have to be multiple clusters. What we don't know is exactly how big these clusters will be, because nobody has any idea how many players a fully working-as-intended dynamically server meshed cluster will actually be able to support. Hopefully lots, but latency alone means that clusters will at least have to be split into regions. The idea is to expand the technology so that a cluster can accommodate more and more players and just sort of find out where the practical limits of the tech are.
Star Citizen is as much a research project as a game, and I'm excited to be a part of it.
Mr Ray on
+2
Options
NEO|PhyteThey follow the stars, bound together.Strands in a braid till the end.Registered Userregular
Now that I've actually put some thought into the whole server meshing thing, the vibe I get from it lines up with Dr. Zola from the first captain america movie.
"The sanity of the plan is of no consequence"
Whether he CAN actually do it remains to be seen.
It was that somehow, from within the derelict-horror, they had learned a way to see inside an ugly, broken thing... And take away its pain.
Warframe/Steam: NFyt
The ability to do this would benefit any multiplayer game, hence why it seems CIG's future business plans are to licence this engine technology to others.
AFAIK, CIG has never once said they plan on selling their tech. That's just randos wildly speculating. CIG doesn't own lumberyard or CryEngine, and it's not a popular enough engine to make much selling tech stuff for the engine.
Unless I am misremembering prior bouts of thread activity, haven't they already been licensing out other techs they've developed to have an income stream outside of nerds buying ships? It's not exactly a stretch to assume that this tech would also be up for grabs, assuming it's not somehow locked into their specific fork of cryengine.
Brian Chambers was just asked about this.
They have no intention of licensing out the engine for now. People have already asked if they can use it and CIG turned 'em down.
So after a decade or so I decided to dick around in this again. Trying some delivery missions. Apparently they want me to put it on top of some delivery machine? And the package just sat a couple of inches away from the delivery point because I'm not tall enough for whatever snap point might exist? And after much fiddling the package just fell behind the machine?
So after a decade or so I decided to dick around in this again. Trying some delivery missions. Apparently they want me to put it on top of some delivery machine? And the package just sat a couple of inches away from the delivery point because I'm not tall enough for whatever snap point might exist? And after much fiddling the package just fell behind the machine?
Assuming it's the right box and the right machine, theoretically interacting with the "drop off "button on the console and then placing the box in the slot should work, but I haven't tried it since they messed around with the personal inventory stuff so it might just be broken
0
Options
Ninja Snarl PMy helmet is my burden.Ninja Snarl: Gone, but not forgotten.Registered Userregular
Well, you gotta hand it to the devs, they've really captured the reality of the gig delivery guy being unable to figure out the delivery kiosk, then going "fuck it", hiding the box somewhere, and leaving.
I knew that walking down the spindle hallway in the Hull C during warp was a gamble, and like an idiot I still did it and ended up in space. RIP 500k credits.
Assuming it's the right box and the right machine, theoretically interacting with the "drop off "button on the console and then placing the box in the slot should work, but I haven't tried it since they messed around with the personal inventory stuff so it might just be broken
Yeah, there was no placing it in the slot. Tried the "place" command, the outline refused to go inside the slot.
0
Options
NEO|PhyteThey follow the stars, bound together.Strands in a braid till the end.Registered Userregular
been a hot minute since I hauled boxes, IIRC the correct point to place it should be an interactable point, rather than needing to try and manually place it.
It was that somehow, from within the derelict-horror, they had learned a way to see inside an ugly, broken thing... And take away its pain.
Warframe/Steam: NFyt
been a hot minute since I hauled boxes, IIRC the correct point to place it should be an interactable point, rather than needing to try and manually place it.
There was a spot where I got a green outline by the desk of the dude in that building, but placing it there did nothing. And wasn't where the objective pointed. Eh, I dunno.
Just had a "dang, that's neat" moment where I went to find some guy's corpse. Late at night, pitch black, nothing but my ships floodlights to properly navigate by since the suit light wasn't good for actually exploring the crash site.
+1
Options
MonwynApathy's a tragedy, and boredom is a crime.A little bit of everything, all of the time.Registered Userregular
Well, you gotta hand it to the devs, they've really captured the reality of the gig delivery guy being unable to figure out the delivery kiosk, then going "fuck it", hiding the box somewhere, and leaving.
What's the deal with melting and purchasing other ships with credit these days? Is that still a thing?
I might have to re-do my ships. When I first purchased them, I had zero kids. Now my kids are 8, 10, and 12, so potentially multicrew ships are more attractive now. Unless they all go to college, move out, and start their own families before this is actually launches.
What's the deal with melting and purchasing other ships with credit these days? Is that still a thing?
In your My Hangar click the blue triangle to the right of "contains" to reveal the Exchange and Gift buttons on any of your game packages or standalone ships. Clicking exchange will bring up a dialog showing you how much store credit you will get if you melt that item. You can use that store credit to buy any non-warbond package, ship, or upgrade in the pledge store. If you later change your mind, you can use a buy back token to repurchase a package or ship at the price you originally paid. About 4 times a year they grant new buy back tokens, but they don't stack so you can't hoard them.
I might have to re-do my ships. When I first purchased them, I had zero kids. Now my kids are 8, 10, and 12, so potentially multicrew ships are more attractive now. Unless they all go to college, move out, and start their own families before this is actually launches.
Well then you can just have your kids and grandkids as crew on your Idris.
Stuff isn't supposed to blow up when I turn the engines on, right?
+1
Options
NEO|PhyteThey follow the stars, bound together.Strands in a braid till the end.Registered Userregular
That's just the patented Drake Hard Start(tm) system
It was that somehow, from within the derelict-horror, they had learned a way to see inside an ugly, broken thing... And take away its pain.
Warframe/Steam: NFyt
What's the deal with melting and purchasing other ships with credit these days? Is that still a thing?
I might have to re-do my ships. When I first purchased them, I had zero kids. Now my kids are 8, 10, and 12, so potentially multicrew ships are more attractive now. Unless they all go to college, move out, and start their own families before this is actually launches.
I hope they sell foton torpedo shells so my grandkids can give me a proper send off when the game finally launches.
Ooh, OK. Those delivery machines have their own "Place" interactive, you're supposed to use that instead of the one on the box. Good thing the game told me this.
OK, how do you actually start making money as a new player (...with a 10yo account)? Tried some bounty hunting, getting paid a whopping 4k per mission right now. Those "find a random corpse" missions pay 18k... if you ever find a corpse.
Well, you gotta hand it to the devs, they've really captured the reality of the gig delivery guy being unable to figure out the delivery kiosk, then going "fuck it", hiding the box somewhere, and leaving.
A decade and half a billion dollars, folks
So Diablo IV except under budget, more interesting to play despite being in beta, and somehow with a feeling of less futility despite expected server wipes?
Posts
AFAIK, CIG has never once said they plan on selling their tech. That's just randos wildly speculating. CIG doesn't own lumberyard or CryEngine, and it's not a popular enough engine to make much selling tech stuff for the engine.
Are you sure CIG didn't actually reinvent a technology that everyone else's been using for a good while already? I'll grant that no other company has been insane enough to want to simulate the entire fabric of reality in a videogame, so that does make server load a tad heavier.
EDIT: oh, does crytek even still exist? Is CryEngine even still a thing outside SC? I guess they took so long with this that UE5 has left cryengine very far behind in terms of looks and optimization. i wonder how fast SQ42 would get done from scratch starting today with UE5. I guess if they just had made it back then we'd be up to SQ45 now
Warframe/Steam: NFyt
Even the earliest MMOs weren't single-server. Asheron's Call for example used a series of servers covering strips of land north-south across the map. Loading through them was seamless.
The servers couldn't cover for each other though so when too many people found Arwic useful as a trade hub, it was causing lag. So they blew it up. https://www.reddit.com/r/AsheronsCall/comments/5nd27n/one_of_the_original_asherons_call_developers_home/
I didn't hop on the ride for squadron 42, my introduction was a friend going "hey this Kickstarter looks kind of neat but I am currently broke, anyone wanna buy me in and I can pay you back later" and I went "that does look neat" and bought in myself as well. Hopefully the man isn't bullshitting too hard about what progress they are at, both for the sq42 people and the part where he said polish phase on SQ42 means Star Citizen is also going to be seeing quicker progress
Warframe/Steam: NFyt
http://steamcommunity.com/id/pablocampy
Warframe/Steam: NFyt
Time to dust off the shame fleet?
edit: holy crap, made this almost 9 years ago to the day....
Because Chris Roberts wants to pull an EVE Online and have all the players in the same world.
:edit: The specific reason this needs fancy unicorn tech is that his fancy high fidelity gameplay requires a LOT of information to be going to and from the server compared to EVE Online's movement/combat model. Currently, the entire Stanton system is run by a single logical server (I hope they're doing logical servers and not a literal single machine), so if a bunch of players decide to have a ship exploding party over at Grim Hex, all the calculations start grinding server performance down for everyone. IIRC the initial pass of server meshing would make it so that (as an example) each planet+moons is on its own server in a cluster, and as someone travels between them they get seamlessly passed around servers, and also the servers talk between each other so that you can see people on the Microtech server before crossing the invisible line that puts you on that server yourself. Dynamic meshing would let the backend shuffle server boundaries as needed to address whatever shenanigans players might be getting into.
Warframe/Steam: NFyt
It seems plausible to me that the, single player, S42 game is more or less good to go, given that most of SC's issues seem network/server related. AI going stupid, desyncs, server implosions, and latency on their ever widening constellation of data service, for example. But, as far as S42 is concerned, the guns seem to shoot, the ships more or less fly, etc.
Depends on how all the complex systems and lack of abstraction work when things really get going. I would be mildly annoyed if there's a mission that craps out because an AI crewmember had a pathing hiccup or whatever that prevented them from doing the important thing. Whole lotta moving parts.
First of all, on server meshing I'm setting my hype levels to "cautiously optimistic" from all of what we've seen at CitizenCon. For those of us paying attention, nothing they showed us was particularly mind-blowing. Yes, it was nice to see Server Meshing actually working, albeit in an incredibly simple cut-down "happy-path" scenario, but this is something we've known about for years and at this point I really can't get excited until I see it actually running on a live server. It is an exciting technology, for reasons which I will go into later, but I'm also very aware that the gap between "it works well enough to demo" and "its deployed and working on a live server that the public can log into" could still be measured in years.
On SQ42, what's in the demo they showed us was amazing and proof once again that Chris Roberts is a master of building hype, but once again we've been burned so many times that until people are actually playing it, its hard to let myself get too excited. And again, the gap between "feature complete" and "you can actually play it" could still be measured in years, and I suspect probably will be. "Feature complete" is not "complete". I will be shocked if its actually in the hands of the general public before 2025.
On CIG deserving nothing but cynicism... unironically yes, absolutely, and I say this as someone who still considers themselves a fan. Star Citizen might not be a scam, but it has been horribly mismanaged from day one, I feel like they only actually figured out what it is they're even building towards in about 2019. I have never seen a company so godawfully bad at estimating how long a thing will take to be done, and I'm a software dev, so I do understand that estimating the time things are going to take can be hard. Some guy on reddit posted a great summary of past CitizenCons, what was announced, what was demoed and what has actually made it into the game. Basically 2016 was the year of "peak bullshit" where a bunch of stuff they demoed was straight up faked (sandworm anyone?) and much of it still hasn't made it into the same, but to their credit, their estimates have gotten better over time, and they've learned to stop letting Chris get hold of the mic and promise that whatever pie-in-the-sky bullshit he just thought of will definitely be in the game within a year.
The reason Server Meshing is so interesting to me, personally, is that it allows them to take advantage of one of the biggest advances in computing of the last couple of decades and apply it to video games, i.e, autoscaling. Skip ahead if you already know what autoscaling is, but if not, history lesson! In the days of yore, when you went to a website like www.penny-arcade.com, your connection would be routed to a single, physical server somewhere that was hosting the website. In some places on the net it still works this way, but increasingly these days rather than a physical web server what you're dealing with is a VM (virtual machine); a bundle of storage, memory and CPU resources hosted on a small part of a much larger supercomputer that manages a bunch of these "pretend" computers. The advantage of this is that if you suddenly get millions of people logging in at once, in the case of PA for example because a new comic just dropped, you can ask the supercomputer if it can allocate more resources so that you can keep up with the increased load. If it doesn't have enough resources to spare it might shunt your site across to a different supercomputer with more capacity or, if even that isn't enough it can share the load between multiple. This is all entirely seamless to the end-user; two users on opposite sites of the world see the same website and as far as their browsers are concerned it is the same website, but they're probably getting their traffic routed to completely different edge servers (which are geographically closer to the user and thus load faster in the browser) mirroring the same content between each other. The reason this is/was a game changer is it meant that it was suddenly much easier and much more affordable for a site to scale up and down with demand; in the before-times if you wanted to be able to handle 10,000 concurrent users at peak usage, you'd need to have enough servers on standby to handle that load ready all the time. Now they can automatically scale up, adding new instances when there's a surge in demand, and dropping back down when demand drops again.
This works well enough in a not-real-time scenario like a website or a forum, what CIG are trying to do is get this technology to work in the context of a real-time online universe, spinning new servers up and down and passing data and players between them seamlessly.
This is not that. Yes, MMOs have been using instancing since the 90's, that's nothing new, but there's always a clearly defined border between zones, usually with a loading screen. You go into a busy city and there's a gate you have to click on, fade out, fade back in, you're in the city. That fade in/out was to hide you getting shunted onto one of many instances of said city, because its where everyone starts and they can't realistically fit all of the newbies into one instance. What Star Citizen is proposing to do is dynamically split servers up into smaller chunks on-the-fly as needed, reducing load on individual servers and allowing them to share information between each other. How well this will actually work in practice we still don't actually know because no-one has ever been crazy enough to try this, and I'm confident that there are a whole bunch of edge cases they haven't considered... but in theory it could be just crazy enough to work. Or it could be exactly crazy enough to crash and burn catastrophically.
The object container system was specifically designed to work with this in mind; basically every area in Star Citizen is a container. The Hurston system is a container, full of ships, planets and space stations. The planets are containers, full of ships, outposts, and any players running around on foot. The outposts, space stations and larger ships are containers full of rooms, which in turn could contain ships if they're hangars, which in turn could contain more rooms if its a big ship, and those ships could also contain hangars which contain ships which contain people and oh no I've gone cross-eyed. But the theory is that each of these containers forms a discrete block of stuff that can be split out into its own server if needed. Too many players in the system? Spin off a server for each planet. Too many players on a planet? Spin off a server for the busiest city. Too many players in the city? Split the busiest areas off into their own servers.
That scenario I just described is still a long way off; what we saw demoed was a very simple example of three interconnected servers in a small area with pre-defined boundaries. But the fact that they're even to do that much is a huge milestone. The next step will be to introduce Pyro, hosted on its own server, and have players able to transition between Hurston and Pyro. Then it'll be to break down the Hurston and Pyro systems into discrete, static chunks, probably one server per planet for starters. That much I'm now confident that they will eventually achieve, the question is what happens when they start trying to make all of this dynamic.
I think that was originally a possible, goal, but within the last year or so they've made it clear that Star Citizen will absolutely not be a single shared universe and there will in fact have to be multiple clusters. What we don't know is exactly how big these clusters will be, because nobody has any idea how many players a fully working-as-intended dynamically server meshed cluster will actually be able to support. Hopefully lots, but latency alone means that clusters will at least have to be split into regions. The idea is to expand the technology so that a cluster can accommodate more and more players and just sort of find out where the practical limits of the tech are.
Star Citizen is as much a research project as a game, and I'm excited to be a part of it.
"The sanity of the plan is of no consequence"
Whether he CAN actually do it remains to be seen.
Warframe/Steam: NFyt
https://m.twitch.tv/clip/RamshackleObedientPieKAPOW-2rDuBajYtONFEIh0
Brian Chambers was just asked about this.
They have no intention of licensing out the engine for now. People have already asked if they can use it and CIG turned 'em down.
I believe you push those buttons and it opens?
It has been a hot minute.
And the marker points to the top of the machine. ¯\_(ツ)_/¯
PSN: ShogunGunshow
Origin: ShogunGunshow
Yeah, there was no placing it in the slot. Tried the "place" command, the outline refused to go inside the slot.
Warframe/Steam: NFyt
There was a spot where I got a green outline by the desk of the dude in that building, but placing it there did nothing. And wasn't where the objective pointed. Eh, I dunno.
Just had a "dang, that's neat" moment where I went to find some guy's corpse. Late at night, pitch black, nothing but my ships floodlights to properly navigate by since the suit light wasn't good for actually exploring the crash site.
A decade and half a billion dollars, folks
I might have to re-do my ships. When I first purchased them, I had zero kids. Now my kids are 8, 10, and 12, so potentially multicrew ships are more attractive now. Unless they all go to college, move out, and start their own families before this is actually launches.
In your My Hangar click the blue triangle to the right of "contains" to reveal the Exchange and Gift buttons on any of your game packages or standalone ships. Clicking exchange will bring up a dialog showing you how much store credit you will get if you melt that item. You can use that store credit to buy any non-warbond package, ship, or upgrade in the pledge store. If you later change your mind, you can use a buy back token to repurchase a package or ship at the price you originally paid. About 4 times a year they grant new buy back tokens, but they don't stack so you can't hoard them.
Well then you can just have your kids and grandkids as crew on your Idris.
Warframe/Steam: NFyt
I hope they sell foton torpedo shells so my grandkids can give me a proper send off when the game finally launches.
I mean, technically yes, but it’s supposed to stay inside the enginr
I haven't changed any keybindings so I'm not sure what the deal is, there.
Obviously I can get around it by shutting off the engines before leaving the pilot seat, but it's weird.
PSN: ShogunGunshow
Origin: ShogunGunshow
I mean, the front nacelle ended up inside the cargo bay. Just needs some duct tape.
...hang on a minute...
So Diablo IV except under budget, more interesting to play despite being in beta, and somehow with a feeling of less futility despite expected server wipes?