I watch quite a few videos relating to SC. SaltEMike, DG360, GrumpyEye, and a few others. It varies depending on who i am interested in following. I like Drew's videos in general, but they tend to drag on, and as i explained, I wasn't up for watching 1.5 hours of a video when i had other plans (and beer), so asked Mole if he had anything that stood out for him. He answered, i gave him a like, and topic over. What is your beef with that?
Game Discussions Star Citizen Discussion Thread v12
- Thread starter Yaffle
- Start date
-
- Tags
- star citizen
I watch quite a few videos relating to SC. SaltEMike, DG360, GrumpyEye, and a few others. It varies depending on who i am interested in following. I like Drew's videos in general, but they tend to drag on, and as i explained, I wasn't up for watching 1.5 hours of a video when i had other plans (and beer), so asked Mole if he had anything that stood out for him. He answered, i gave him a like, and topic over. What is your beef with that?
Think i just found Mole's alt account!
Viajero
Volunteer Moderator
I think he is pretty clear and there is really no arguing with it: 9 years waiting, 400 million pledged in good faith by backers. No product. This is just a joke.
"I am laughing tears of sadness because I am the one who put in money"
Unless of course, CIG, or anyone else, considers what we have now as released somehow.
Last edited:
Viajero
Volunteer Moderator
Not saying this is extremely bad luck but I only had blank panel bugs during the alpha and possibly an occasion shortly after release if memory serves. But I have been doing all kind of missions extensively since then and have not encountered that particular bug at all, to be honest.
Last edited:
Well, I finished the free-fly event and have to say, it shifted my opinion a little bit, I didn't encounter too many bugs etc and FPS was solid, didn't feel like it was cooking my gaming laptop (like Cyberpunk does)so I grabbed a copy (the Aurora base deal, nothing more than that, nor will I ever go beyond it).
Possibly just bad luck. I've been past those panels dozens of times while doing other things and not once were they blank.
And so it begins...
*Totally NOT a cult!
MuppetsLordy
Source: https://old.reddit.com/r/starcitizen/comments/r8bwtl/two_weeks_inthis_is_a_problem/
An 'ED refugee' responds
Source: https://old.reddit.com/r/starcitizen/comments/r8bwtl/two_weeks_inthis_is_a_problem/hn4uih1/
As do many others:
Source: https://old.reddit.com/r/starcitizen/comments/r8bwtl/two_weeks_inthis_is_a_problem/hn586of/
Source: https://old.reddit.com/r/starcitizen/comments/r8bwtl/two_weeks_inthis_is_a_problem/hn5c5ya/
Source: https://old.reddit.com/r/starcitizen/comments/r8bwtl/two_weeks_inthis_is_a_problem/hn4y3bt/
Source: https://old.reddit.com/r/starcitizen/comments/r8bwtl/two_weeks_inthis_is_a_problem/hn4s8f0/
Source: https://old.reddit.com/r/starcitizen/comments/r8bwtl/two_weeks_inthis_is_a_problem/hn69dms/
Bonus perspective:
Source: https://old.reddit.com/r/starcitizen/comments/r8bwtl/two_weeks_inthis_is_a_problem/hn4ya0e/
Optimise CCU chains... what the hell is he on about? It doesn't take a masters degree to shuffle some ships around and get new stuff for zero dollar
Think i just found Mole's alt account!
I have a feeling he only has the one ship, a limited budget and yet he's dreaming of owning some bling barge or other through paying cash for it rather than working for it by playing the game. I suspect he's going to have issues down the line with spending the rent money on ships...it's a pure mugs game.
Those of us who play the CCU game every year are stupid enough to have already spent too much on buying ships and ship packages to juggle with. Being able to buy ships in game just by playing negates all this CCU palaver, I only do it because I can, I'm already a concierge muppet... and it's free. I don't spend any new cash, haven't done for years...$10-$15 here and there on cosmetic additions like paint jobs aside.
If I was starting out now as a newbie rather than back when I did, buying additional ships for cash wouldn't enter into it, I'd buy in with a Jack of all trades $100 Cutty Black and be done with spending. Considering all the ships I have, I still spend most of my playing time flying a Cutty Black. All the other stuff...even the Prospector... is just hangar decoration I could have worked my way toward.
Last edited:
Best way, stick to it regardless.
Viajero
Volunteer Moderator
Blockchain? Did you say blockchain? Ah not my bad, I´d have sworn... nevermindOptimise CCU chains... what the actual F is he on about? It doesn't take a masters degree to shuffle some ships around and get new stuff for zero dollar
You'll find he only has the one ship, a limited spending budget and he's dreaming of owning some bling barge or other through paying cash for it rather than working for it by playing the game. I suspect he's going to have issues down the line with spending the rent money on ships...it's a pure mugs game.
Those of us who play the CCU game every year are stupid enough to have already spent too much on buying ships and ship packages to juggle with. Being able to buy ships in game just by playing negates all this CCU palaver, I only do it because I can, I'm already a concierge muppet... and it's free. I don't spend any new cash, haven't done for years...$10-$15 here and there on cosmetic additions like paint jobs aside.
If I was starting out now as a newbie rather than back when I did, buying additional ships for cash wouldn't enter into it, I'd buy in with a Jack of all trades $100 Cutty Black and be done with spending. Considering all the ships I have, I still spend most of my playing time flying a Cutty Black. All the other stuff...even the Prospector... is just hangar decoration I could have worked my way toward.
.Good luck in the 'verse! Don't let the Bed Bugs Bite! (Seriously, don't log out in beds, you'll need to do a character reset)
Decent chat here with an MMO code monkey working at Zenimax:
Source: https://www.youtube.com/watch?v=C41vT-f5dSc
TLDR:
Notes:
On his dev background:
On being a space sim fan:
On entity groupings vs older approaches: [9m]
On Static Server Meshing: [15m25s]
On Dynamic Server Meshing: [15m50s]
On blockading full locations: [16m50s]
On combat across mesh boundaries: [17m52s]
On capital ship battles: [18m50s]
On being seated lowering demands on server etc: [20m46s]
On everyone in SC all going to one location: [24m45s]
Possible max player count in one place? [26m40s]
Possibility of more numbers in combat-free cities? [30m]
Base building: [34m50s]
Merchantman / Privateer marketplaces: [37m55s]
On large owned items persisting: [38m55s]
On devs versus stakeholders: [41m55s]
What will Server Meshing look like next year? [45m35s]
Personal Residence Buildings in Landing Zones: [51m25s]
---
All fits pretty well with what we know from the official Q&A.
TLDR:
- Static Server Meshing is comparable to systems used by current games, such as Destiny. (IE 50 people in Stanton, 50 people in Pyro, and the jump point being a loading screen). [15m25s]
- "With the best will in the world, I don't think we'll ever see a situation where combat can take place over a mesh boundary." [17m52s]
- Capital ship battles will all have to be on one server [18m50s]
- Max player count in one place? He guesses: 100, maybe 200. [26m40s]
- "I know from my own side as a programmer you can sit there and say: This is what the limit's going to be. And there will be stakeholders who go: No, I don't believe that. Until you actually prove to them what happens when those limits get exceeded..." [41m55s]
- "If the Static Meshes come in, and work as planned, you will see absolutely nothing different". Other than things that currently exist performing a bit better, and potentially opening up Pyro, not much will change. Tony Z might have more headroom for the background sim. [And Chris for coffee cup fidelity ] [45m35s]
Notes:
On his dev background:
- "I'm one of the code monkeys, backend, for Zenimax, on their games." [1m19s]
- Wolfenstein etc got him into netcode, and for a long time that was his thing professionally. Not a graphics programmer etc. Very different world to his server programmer world. Summarises his works and interests as: "Network, backend, trying to control the flow of things." [5m20s+]
- Hasn't had any contact with CIG devs. Just working off own experience here. [41m30s]
On being a space sim fan:
- Backed SC in 2012. Playing semi-regularly since 3.10
- Is playing less of 'another one'. IE Elite Dangerous [2m30s+]
- Has multiple SC accounts. Happy to play more villainous ways etc. [32m]
On entity groupings vs older approaches: [9m]
- In the old days bandwidth was the issue. Dial up etc. And even today lots of people don't have the best connections. So devs would send only the information they really needed to synchronise. Great for performance, but restricting. If you wanted to give someone a bag, you needed to work out how many things they could have in the bag. Needed to change all the netcode touching that player, and the bag etc. So anything you wanted to add, got hard.
- By comparison entities are much more "generic", especially when it comes to exchanging them with databases etc. When being moved around, all the nested entities inside don't need to be queried.
- In the old system it was easier to have checks and balances, validation rules, to make sure everything is consistent.
- Part of the reason the inventory UI went up so quickly, he thinks. is that all they're doing is showing you the nested objects inside an entity. (May even have had an internal tool to do that already).
- Recursive imbedding in entity system can become a problem in its own right. [Doesn't say how]
- Entity use far from unique. But scale that SC wishes to deploy them at may be.
On Static Server Meshing: [15m25s]
- A number of other games have, it could be argued, 'Static Server Meshing' in place. If SC was to launch with Static SM, Destiny is already doing the same sort of thing now. (IE 50 people in Stanton, 50 people in Pyro, and the jump point being a loading screen).
On Dynamic Server Meshing: [15m50s]
- Sub-dividing the game world very dynamically, and defining the static meshes much more 'generically' = the more ambitious bit. Games that are based around fixed transition points between servers have to be designed like that from the ground up. Using gameplay elements to hide the seams etc. SC is trying to do that much more dynamically. That's where the real challenge and the ambition is.
On blockading full locations: [16m50s]
- Ninetails blockades (blocking local quantum) would work as a way to prohibit entry to an area temporarily, making player take a longer route during which gameplay unlikely to occur (combat) during which a player can be moved to a different shard / instance etc.
On combat across mesh boundaries: [17m52s]
- "With the best will in the world, I don't think we'll ever see a situation where combat can take place over a mesh boundary.". There are so many edge cases. So it has to be a place that's relatively stable, where you have time for someone playing in Aus, on an EU server, to do the exchange of stuff they need to, from one instance to the next.
On capital ship battles: [18m50s]
- What about large capital ship battles? If they were actively in combat, almost definitely they'll all have to be on one server. If there's no combat going on, the actual carriers themselves could exist in a static 'shard', and have purely interiors existing in dynamic 'shards' [assuming he means server here]. Once you start firing turrets / where you need to start arbitrating decisions between different meshes, that's where there are so many edge cases. Doesn't think they'll try that.
- Lot of load on that single server? Yes. [Heavy sigh]
On being seated lowering demands on server etc: [20m46s]
- So much of SC's combat being seated, means it does limit what you need to communicate to others about you. Where is your turret pointing? Are you firing? What are your health stats. Not a lot else. (Far less, than you walking around. Which has a whole load of other attributes. You could be drunk, you could be leaping, crouching etc). [They joke about exaggerating force reactions to point where nobody can leave seats . And note that the Liberator came with 16 jump seats.]
- Being seated helps with desync. If you're seated it's much easier for the back end to give authoritative info, compared to all the info related to walking. Moving between meshes, would almost guarantee you'll have to be seated. When seated, you can basically be treated as part of the ship. Handy cheat / short cut. You're no longer responsible for your position. You're not costing the server or the networking anything at that point. (Underlines point about games normally faking a simulation, not actually simulating. It's all smoke and mirrors.)
On everyone in SC all going to one location: [24m45s]
- "Never, ever, ever, ever going to happen that you are all going stand there and see each other."
Possible max player count in one place? [26m40s]
- Guesses: 100, maybe 200. It's not what you can do with server performance per se. Some things have to be metered by the performance of the slowest player. (Min spec computers, bandwidth per region etc).
Possibility of more numbers in combat-free cities? [30m]
- While they're armistice zones, yes. But moment armistice zone goes away, no. (But says has never found an effective crime & punishment system in multiplayer game. Thinks more wallet impact would be an effective stick to guide player behaviour etc. [32m4s])
Base building: [34m50s]
- Still unsure as to exactly how they're going to work. If you can store ships, inventory etc there, he can't see them persisting in an attackable form when you're offline. [Even if it's on the shard that you were on]. What he could see, to tackle 'combat logging' style exploits, that if there were people / hostiles in the area, he could see keeping it around in that shard. But in general has to assume they're no different than bunkers. Even if you drop a bomb on them right now, it may destroy some defensive installations, or local players, but the base itself is indestructible.
Merchantman / Privateer marketplaces: [37m55s]
- Would have the same issues as bases, and as capital ships, on server meshing front. How are they going to exist in the game? Will they be attackable? Will it be that if it's in motion, it'll be attackable. If it's parked, maybe it'll be indestructible? (Similar to ED's Carriers etc).
On large owned items persisting: [38m55s]
- CIG will need to decide whether they allow these things to persist when owner not there etc.
- Items with low friction to ownership can cause issues. Imagine a world where every other person owns a Merchantman. (Not something that he thinks is likely, but something that could happen). Backend guys can't rely on players being sensible, you have to protect against the worst case. IE: What would happen if everyone bought a BMM and decided to park them all together?
On devs versus stakeholders: [41m55s]
- Won't hear more on the above until they've experimented with it and have some results.
- "I know from my own side as a programmer you can sit there and say: This is what the limit's going to be. And there will be stakeholders who go: No, I don't believe that. Until you actually prove to them what happens when those limits get exceeded... Some of the stuff is actually proving to both sides that there is an issue. And then seeing how both sides decide to solve it."
- "It's the battle. And why, you know, people always joke about the likes of Peter Molyneux, and other people who will say one thing... Anytime that I ever hear about any game, those things that they say this game is going to be is their vision.... Every game is an iterative process, and whilst you try to stay as close to that vision as you can, sometimes reality means you can't get that. Sometimes that's, y'know, reality brought on by deadlines."
What will Server Meshing look like next year? [45m35s]
- "If the Static Meshes come in, and work as planned, you will see absolutely nothing different."
- Thinks they're quite brave in making all of this detail public, because so much of this stuff goes on in a lot of games, but it's black box info. This stuff, along with load balancing, is normal behind the scenes stuff.
- Potentially will give a little bit more room to the servers. So you might start seeing more responsive AI, things spawning in when they're supposed to, increased item limits in the world.
- Other than things that currently exist performing a bit better, and potentially opening up Pyro, not much will change.
- Should allow Tony Z to do more fancy things in the Quanta background sim.
- The dangling pipes demoed in Pyro, which will have to be networked, will have more headroom. All those coffee makers, with their cups that you can pick up. You can bet that CIG would love for them to be there, working, but currently they don't have the headroom.
Personal Residence Buildings in Landing Zones: [51m25s]
- That could be one of the most trivial cases for a dynamic mesh. If only you, and/or a selection of friends, or even an org, have access. Lots of games do this for habitations. It's an instance. The tricky thing is faking that entrance into the room. Could imagine them using a very sterile L-shaped entrance hall for the transition etc.
- Can't have exact buildings for every potential player. But could have one entrance way behind which is a 'rotating pod' of options. (Could use rental system to make sure only active players were owning a location at any given time, in which case could have bespoke locations).
---
All fits pretty well with what we know from the official Q&A.
Decent chat here with an MMO code monkey working at Zenimax:
Source: https://www.youtube.com/watch?v=C41vT-f5dSc
TLDR:
- Static Server Meshing is comparable to systems used by current games, such as Destiny. (IE 50 people in Stanton, 50 people in Pyro, and the jump point being a loading screen). [15m25s]
- "With the best will in the world, I don't think we'll ever see a situation where combat can take place over a mesh boundary." [17m52s]
- Capital ship battles will all have to be on one server [18m50s]
- Max player count in one place? He guesses: 100, maybe 200. [26m40s]
- "I know from my own side as a programmer you can sit there and say: This is what the limit's going to be. And there will be stakeholders who go: No, I don't believe that. Until you actually prove to them what happens when those limits get exceeded..." [41m55s]
- "If the Static Meshes come in, and work as planned, you will see absolutely nothing different". Other than things that currently exist performing a bit better, and potentially opening up Pyro, not much will change. Tony Z might have more headroom for the background sim. [And Chris for coffee cup fidelity
Notes:
On his dev background:
- "I'm one of the code monkeys, backend, for Zenimax, on their games." [1m19s]
- Wolfenstein etc got him into netcode, and for a long time that was his thing professionally. Not a graphics programmer etc. Very different world to his server programmer world. Summarises his works and interests as: "Network, backend, trying to control the flow of things." [5m20s+]
- Hasn't had any contact with CIG devs. Just working off own experience here. [41m30s]
On being a space sim fan:
- Backed SC in 2012. Playing semi-regularly since 3.10
- Is playing less of 'another one'. IE Elite Dangerous
- Has multiple SC accounts. Happy to play more villainous ways etc. [32m]
On entity groupings vs older approaches: [9m]
- In the old days bandwidth was the issue. Dial up etc. And even today lots of people don't have the best connections. So devs would send only the information they really needed to synchronise. Great for performance, but restricting. If you wanted to give someone a bag, you needed to work out how many things they could have in the bag. Needed to change all the netcode touching that player, and the bag etc. So anything you wanted to add, got hard.
- By comparison entities are much more "generic", especially when it comes to exchanging them with databases etc. When being moved around, all the nested entities inside don't need to be queried.
- In the old system it was easier to have checks and balances, validation rules, to make sure everything is consistent.
- Part of the reason the inventory UI went up so quickly, he thinks. is that all they're doing is showing you the nested objects inside an entity. (May even have had an internal tool to do that already).
- Recursive imbedding in entity system can become a problem in its own right. [Doesn't say how]
- Entity use far from unique. But scale that SC wishes to deploy them at may be.
On Static Server Meshing: [15m25s]
- A number of other games have, it could be argued, 'Static Server Meshing' in place. If SC was to launch with Static SM, Destiny is already doing the same sort of thing now. (IE 50 people in Stanton, 50 people in Pyro, and the jump point being a loading screen).
On Dynamic Server Meshing: [15m50s]
- Sub-dividing the game world very dynamically, and defining the static meshes much more 'generically' = the more ambitious bit. Games that are based around fixed transition points between servers have to be designed like that from the ground up. Using gameplay elements to hide the seams etc. SC is trying to do that much more dynamically. That's where the real challenge and the ambition is.
On blockading full locations: [16m50s]
- Ninetails blockades (blocking local quantum) would work as a way to prohibit entry to an area temporarily, making player take a longer route during which gameplay unlikely to occur (combat) during which a player can be moved to a different shard / instance etc.
On combat across mesh boundaries: [17m52s]
- "With the best will in the world, I don't think we'll ever see a situation where combat can take place over a mesh boundary.". There are so many edge cases. So it has to be a place that's relatively stable, where you have time for someone playing in Aus, on an EU server, to do the exchange of stuff they need to, from one instance to the next.
On capital ship battles: [18m50s]
- What about large capital ship battles? If they were actively in combat, almost definitely they'll all have to be on one server. If there's no combat going on, the actual carriers themselves could exist in a static 'shard', and have purely interiors existing in dynamic 'shards' [assuming he means server here]. Once you start firing turrets / where you need to start arbitrating decisions between different meshes, that's where there are so many edge cases. Doesn't think they'll try that.
- Lot of load on that single server? Yes. [Heavy sigh]
On being seated lowering demands on server etc: [20m46s]
- So much of SC's combat being seated, means it does limit what you need to communicate to others about you. Where is your turret pointing? Are you firing? What are your health stats. Not a lot else. (Far less, than you walking around. Which has a whole load of other attributes. You could be drunk, you could be leaping, crouching etc). [They joke about exaggerating force reactions to point where nobody can leave seats
- Being seated helps with desync. If you're seated it's much easier for the back end to give authoritative info, compared to all the info related to walking. Moving between meshes, would almost guarantee you'll have to be seated. When seated, you can basically be treated as part of the ship. Handy cheat / short cut. You're no longer responsible for your position. You're not costing the server or the networking anything at that point. (Underlines point about games normally faking a simulation, not actually simulating. It's all smoke and mirrors.)
On everyone in SC all going to one location: [24m45s]
- "Never, ever, ever, ever going to happen that you are all going stand there and see each other."
Possible max player count in one place? [26m40s]
- Guesses: 100, maybe 200. It's not what you can do with server performance per se. Some things have to be metered by the performance of the slowest player. (Min spec computers, bandwidth per region etc).
Possibility of more numbers in combat-free cities? [30m]
- While they're armistice zones, yes. But moment armistice zone goes away, no. (But says has never found an effective crime & punishment system in multiplayer game. Thinks more wallet impact would be an effective stick to guide player behaviour etc. [32m4s])
Base building: [34m50s]
- Still unsure as to exactly how they're going to work. If you can store ships, inventory etc there, he can't see them persisting in an attackable form when you're offline. [Even if it's on the shard that you were on]. What he could see, to tackle 'combat logging' style exploits, that if there were people / hostiles in the area, he could see keeping it around in that shard. But in general has to assume they're no different than bunkers. Even if you drop a bomb on them right now, it may destroy some defensive installations, or local players, but the base itself is indestructible.
Merchantman / Privateer marketplaces: [37m55s]
- Would have the same issues as bases, and as capital ships, on server meshing front. How are they going to exist in the game? Will they be attackable? Will it be that if it's in motion, it'll be attackable. If it's parked, maybe it'll be indestructible? (Similar to ED's Carriers etc).
On large owned items persisting: [38m55s]
- CIG will need to decide whether they allow these things to persist when owner not there etc.
- Items with low friction to ownership can cause issues. Imagine a world where every other person owns a Merchantman. (Not something that he thinks is likely, but something that could happen). Backend guys can't rely on players being sensible, you have to protect against the worst case. IE: What would happen if everyone bought a BMM and decided to park them all together?
On devs versus stakeholders: [41m55s]
- Won't hear more on the above until they've experimented with it and have some results.
- "I know from my own side as a programmer you can sit there and say: This is what the limit's going to be. And there will be stakeholders who go: No, I don't believe that. Until you actually prove to them what happens when those limits get exceeded... Some of the stuff is actually proving to both sides that there is an issue. And then seeing how both sides decide to solve it."
- "It's the battle. And why, you know, people always joke about the likes of Peter Molyneux, and other people who will say one thing... Anytime that I ever hear about any game, those things that they say this game is going to be is their vision.... Every game is an iterative process, and whilst you try to stay as close to that vision as you can, sometimes reality means you can't get that. Sometimes that's, y'know, reality brought on by deadlines."
What will Server Meshing look like next year? [45m35s]
- "If the Static Meshes come in, and work as planned, you will see absolutely nothing different."
- Thinks they're quite brave in making all of this detail public, because so much of this stuff goes on in a lot of games, but it's black box info. This stuff, along with load balancing, is normal behind the scenes stuff.
- Potentially will give a little bit more room to the servers. So you might start seeing more responsive AI, things spawning in when they're supposed to, increased item limits in the world.
- Other than things that currently exist performing a bit better, and potentially opening up Pyro, not much will change.
- Should allow Tony Z to do more fancy things in the Quanta background sim.
- The dangling pipes demoed in Pyro, which will have to be networked, will have more headroom. All those coffee makers, with their cups that you can pick up. You can bet that CIG would love for them to be there, working, but currently they don't have the headroom.
Personal Residence Buildings in Landing Zones: [51m25s]
- That could be one of the most trivial cases for a dynamic mesh. If only you, and/or a selection of friends, or even an org, have access. Lots of games do this for habitations. It's an instance. The tricky thing is faking that entrance into the room. Could imagine them using a very sterile L-shaped entrance hall for the transition etc.
- Can't have exact buildings for every potential player. But could have one entrance way behind which is a 'rotating pod' of options. (Could use rental system to make sure only active players were owning a location at any given time, in which case could have bespoke locations).
---
All fits pretty well with what we know from the official Q&A.
Pretty much what we've speculated.
And if someone can somehow manage to fully crew a Javelin (or whatever the 80 crew ship is), when its finally released, they might find they are the only ship in the shard.
That'd be a worst case scenario. But as he notes, networking engineers have to prepare for the worst case
Like, he notes that in non combat scenarios they could support a lot more players in conjunction. But how do you enforce a non-combat scenario? Crime & Punishment systems are rarely a true deterrent, so they have to assume that voluntary no-fire zones can and will be broken.
Kinda summons up visions of big amnesty zones remaining, where ships can't fire at all, with a couple of whale ships just floating around in catwalk mode
(And everyone else locked out from the area by a Nine Tails style no-quantum zone, to prevent the server getting overloaded
)
If they can't instance with you they can't destroy yer ship <grin>
Great vid...like Space Tomato's stuffDecent chat here with an MMO code monkey working at Zenimax:
Source: https://www.youtube.com/watch?v=C41vT-f5dSc
TLDR:
- Static Server Meshing is comparable to systems used by current games, such as Destiny. (IE 50 people in Stanton, 50 people in Pyro, and the jump point being a loading screen). [15m25s]
- "With the best will in the world, I don't think we'll ever see a situation where combat can take place over a mesh boundary." [17m52s]
- Capital ship battles will all have to be on one server [18m50s]
- Max player count in one place? He guesses: 100, maybe 200. [26m40s]
- "I know from my own side as a programmer you can sit there and say: This is what the limit's going to be. And there will be stakeholders who go: No, I don't believe that. Until you actually prove to them what happens when those limits get exceeded..." [41m55s]
- "If the Static Meshes come in, and work as planned, you will see absolutely nothing different". Other than things that currently exist performing a bit better, and potentially opening up Pyro, not much will change. Tony Z might have more headroom for the background sim. [And Chris for coffee cup fidelity
Notes:
On his dev background:
- "I'm one of the code monkeys, backend, for Zenimax, on their games." [1m19s]
- Wolfenstein etc got him into netcode, and for a long time that was his thing professionally. Not a graphics programmer etc. Very different world to his server programmer world. Summarises his works and interests as: "Network, backend, trying to control the flow of things." [5m20s+]
- Hasn't had any contact with CIG devs. Just working off own experience here. [41m30s]
On being a space sim fan:
- Backed SC in 2012. Playing semi-regularly since 3.10
- Is playing less of 'another one'. IE Elite Dangerous
- Has multiple SC accounts. Happy to play more villainous ways etc. [32m]
On entity groupings vs older approaches: [9m]
- In the old days bandwidth was the issue. Dial up etc. And even today lots of people don't have the best connections. So devs would send only the information they really needed to synchronise. Great for performance, but restricting. If you wanted to give someone a bag, you needed to work out how many things they could have in the bag. Needed to change all the netcode touching that player, and the bag etc. So anything you wanted to add, got hard.
- By comparison entities are much more "generic", especially when it comes to exchanging them with databases etc. When being moved around, all the nested entities inside don't need to be queried.
- In the old system it was easier to have checks and balances, validation rules, to make sure everything is consistent.
- Part of the reason the inventory UI went up so quickly, he thinks. is that all they're doing is showing you the nested objects inside an entity. (May even have had an internal tool to do that already).
- Recursive imbedding in entity system can become a problem in its own right. [Doesn't say how]
- Entity use far from unique. But scale that SC wishes to deploy them at may be.
On Static Server Meshing: [15m25s]
- A number of other games have, it could be argued, 'Static Server Meshing' in place. If SC was to launch with Static SM, Destiny is already doing the same sort of thing now. (IE 50 people in Stanton, 50 people in Pyro, and the jump point being a loading screen).
On Dynamic Server Meshing: [15m50s]
- Sub-dividing the game world very dynamically, and defining the static meshes much more 'generically' = the more ambitious bit. Games that are based around fixed transition points between servers have to be designed like that from the ground up. Using gameplay elements to hide the seams etc. SC is trying to do that much more dynamically. That's where the real challenge and the ambition is.
On blockading full locations: [16m50s]
- Ninetails blockades (blocking local quantum) would work as a way to prohibit entry to an area temporarily, making player take a longer route during which gameplay unlikely to occur (combat) during which a player can be moved to a different shard / instance etc.
On combat across mesh boundaries: [17m52s]
- "With the best will in the world, I don't think we'll ever see a situation where combat can take place over a mesh boundary.". There are so many edge cases. So it has to be a place that's relatively stable, where you have time for someone playing in Aus, on an EU server, to do the exchange of stuff they need to, from one instance to the next.
On capital ship battles: [18m50s]
- What about large capital ship battles? If they were actively in combat, almost definitely they'll all have to be on one server. If there's no combat going on, the actual carriers themselves could exist in a static 'shard', and have purely interiors existing in dynamic 'shards' [assuming he means server here]. Once you start firing turrets / where you need to start arbitrating decisions between different meshes, that's where there are so many edge cases. Doesn't think they'll try that.
- Lot of load on that single server? Yes. [Heavy sigh]
On being seated lowering demands on server etc: [20m46s]
- So much of SC's combat being seated, means it does limit what you need to communicate to others about you. Where is your turret pointing? Are you firing? What are your health stats. Not a lot else. (Far less, than you walking around. Which has a whole load of other attributes. You could be drunk, you could be leaping, crouching etc). [They joke about exaggerating force reactions to point where nobody can leave seats
- Being seated helps with desync. If you're seated it's much easier for the back end to give authoritative info, compared to all the info related to walking. Moving between meshes, would almost guarantee you'll have to be seated. When seated, you can basically be treated as part of the ship. Handy cheat / short cut. You're no longer responsible for your position. You're not costing the server or the networking anything at that point. (Underlines point about games normally faking a simulation, not actually simulating. It's all smoke and mirrors.)
On everyone in SC all going to one location: [24m45s]
- "Never, ever, ever, ever going to happen that you are all going stand there and see each other."
Possible max player count in one place? [26m40s]
- Guesses: 100, maybe 200. It's not what you can do with server performance per se. Some things have to be metered by the performance of the slowest player. (Min spec computers, bandwidth per region etc).
Possibility of more numbers in combat-free cities? [30m]
- While they're armistice zones, yes. But moment armistice zone goes away, no. (But says has never found an effective crime & punishment system in multiplayer game. Thinks more wallet impact would be an effective stick to guide player behaviour etc. [32m4s])
Base building: [34m50s]
- Still unsure as to exactly how they're going to work. If you can store ships, inventory etc there, he can't see them persisting in an attackable form when you're offline. [Even if it's on the shard that you were on]. What he could see, to tackle 'combat logging' style exploits, that if there were people / hostiles in the area, he could see keeping it around in that shard. But in general has to assume they're no different than bunkers. Even if you drop a bomb on them right now, it may destroy some defensive installations, or local players, but the base itself is indestructible.
Merchantman / Privateer marketplaces: [37m55s]
- Would have the same issues as bases, and as capital ships, on server meshing front. How are they going to exist in the game? Will they be attackable? Will it be that if it's in motion, it'll be attackable. If it's parked, maybe it'll be indestructible? (Similar to ED's Carriers etc).
On large owned items persisting: [38m55s]
- CIG will need to decide whether they allow these things to persist when owner not there etc.
- Items with low friction to ownership can cause issues. Imagine a world where every other person owns a Merchantman. (Not something that he thinks is likely, but something that could happen). Backend guys can't rely on players being sensible, you have to protect against the worst case. IE: What would happen if everyone bought a BMM and decided to park them all together?
On devs versus stakeholders: [41m55s]
- Won't hear more on the above until they've experimented with it and have some results.
- "I know from my own side as a programmer you can sit there and say: This is what the limit's going to be. And there will be stakeholders who go: No, I don't believe that. Until you actually prove to them what happens when those limits get exceeded... Some of the stuff is actually proving to both sides that there is an issue. And then seeing how both sides decide to solve it."
- "It's the battle. And why, you know, people always joke about the likes of Peter Molyneux, and other people who will say one thing... Anytime that I ever hear about any game, those things that they say this game is going to be is their vision.... Every game is an iterative process, and whilst you try to stay as close to that vision as you can, sometimes reality means you can't get that. Sometimes that's, y'know, reality brought on by deadlines."
What will Server Meshing look like next year? [45m35s]
- "If the Static Meshes come in, and work as planned, you will see absolutely nothing different."
- Thinks they're quite brave in making all of this detail public, because so much of this stuff goes on in a lot of games, but it's black box info. This stuff, along with load balancing, is normal behind the scenes stuff.
- Potentially will give a little bit more room to the servers. So you might start seeing more responsive AI, things spawning in when they're supposed to, increased item limits in the world.
- Other than things that currently exist performing a bit better, and potentially opening up Pyro, not much will change.
- Should allow Tony Z to do more fancy things in the Quanta background sim.
- The dangling pipes demoed in Pyro, which will have to be networked, will have more headroom. All those coffee makers, with their cups that you can pick up. You can bet that CIG would love for them to be there, working, but currently they don't have the headroom.
Personal Residence Buildings in Landing Zones: [51m25s]
- That could be one of the most trivial cases for a dynamic mesh. If only you, and/or a selection of friends, or even an org, have access. Lots of games do this for habitations. It's an instance. The tricky thing is faking that entrance into the room. Could imagine them using a very sterile L-shaped entrance hall for the transition etc.
- Can't have exact buildings for every potential player. But could have one entrance way behind which is a 'rotating pod' of options. (Could use rental system to make sure only active players were owning a location at any given time, in which case could have bespoke locations).
---
All fits pretty well with what we know from the official Q&A.