Well, jeez, that does seem pretty easy. Why am I even complaining? Faith restored!
Game Discussions Star Citizen Discussion Thread v12
- Thread starter Yaffle
- Start date
-
- Tags
- star citizen
This is the approach of Dual Universe, founded by folks with careers in cutting edge server work. And note the DU technical lead/CEO and their promo material makes clear it will not be Real Time Twitch gameplay but typical fantasy rpg/mmo lock and fire mode. Because they know it won't do twitch, at least for another 10 more years of progress and R&D if then, because physics.
I do. That's exactly why I said what I did.
And the key phrase there is “dynamic”, not “server meshing”. In fact, server meshing isn't even a factor in that. The problem is one of dividing instances intelligently, not begging more servers to die under the load of solving unintelligent instancing issues. And none of that is related to the existence (or not) of other systems — you'd have the exact same problems whether you'd have one system or a hundred.
Again, neither is in any way a requirement of the other.
…and the reason they will fail is because server meshing doesn't in any way address the fundamental issue with (and raison d'etre) of instances: that there are limits to how many clients you can have interacting before the general O(n²):ness of it all overcomes the server's capabilities. In fact, server meshing would only make those problems worse. Not only do you still have the exact same O(n²) complexity, but now you also have to manage the overhead of keeping the two instances connected and synced. Linking them may, under optimal circumstances let you filter out some data from the cross-flow, but that tiny gain is massively overshadowed by the simple fact that you have twice as many n:s to deal with, and added overhead on top of that. The complexity doesn't decrease — which is what the whole thing is trying to accomplish — instead it increases.
It's not a dream. It's a technobabble sales pitch for a fake solution to a problem that they have not been able to demonstrate any comprehension of.
Last edited:
What i'm hearing is CIG should outsource all their work to Turbulent and then something might get finished!
Latest patch. Still in Evocati. On July 15th
pastebin.com
star-citizen-alpha-3-10-0l-ptu-5639290-patch-notes - Pastebin.com
Pastebin.com is the number one paste tool since 2002. Pastebin is a website where you can store text online for a set period of time.
Actually here i'm a little more optimistic than you that its achievable, or at least has to be achievable, because if they don't do dynamic instances the servers will just melt once a few dozen gather in the same location (unless CIG can make some good improvements in this area... to date, it doesn't look like they can).
One thing CIG really need to do though is accept they might have to split instances within the same area when things get busy. I mean, there are limits and only two ways to go about it. Split instances or slow down the ticks, and while that works for EvE, its not going to work for a FPS game.
How long CIG will take to invent this magical tech that has clearly never been done before (aside for a few dozen times, but they don't count for reasons) remains to be seen.
"CIG management constantly lies. Please support this project."
"The game is already released and people should be happy with it."
Which is why selling giant capital ships for thousands of dollars for a system that will never handle them is a scam.
https://pastebin.com/4ZXGpz8h
Star Citizen Patch 3.10.0l (release candidate build)
Known Issues
* Player will be unable to control the movement of Super Hornet using mouse Controls after Exiting the Remote Turret using Interaction Mode
- Attempting to perform an orbital QT to Levski from orbit will instead direct the player to Origin at Orbital QT speed.
- Players are not able to join friends PU instance from the main menu
- All 3D icons in personal inventory don't appear correctly when in the PIT
- Landing gear will take multiple presses of the keybind to activate.
- When leaving an Exit Spline in some Large Ships, the Ships will not re-enter the Spline correctly and fall towards the Planet without returning control to the Player
- Security scans won't recognize the drugs on board a players ship during the Drug Production run missions
- The Ship Hologram / 3D Model will take over 20 seconds to load the first time the Player views one in VMA or the Arena Commander Menus
- The Area18 nav marker will appear at the central plaza, not the spaceport.
- Legally landed ships may be impounded
- Multi-Tool in Klescher Prison Commissary does not come with a OreBit mining attachment
- If the player exits to menu/disconnects/crashes during the prison load screen transition their load-out will not be changed
- Player may get stuck on a loading screen when leaving prison or respawning
- Player may get stuck on a prison loading screen if sent back to prison a second time without clearing their crime stat
- When going into ADS and crouching / prone you will be unable to see through the optics scope on the Arrowhead or P6-LR
- Freelook does not always work properly in turrets
- Turrets ignore power setting
- Players may fall out of turrets and clip through geometry when exiting them with the female model (Workaround: Avoid choosing the female model for now)
- Turret projectiles and fixed weapon fire is slightly desynced so hits may appear to not register but do
- Exiting a remote turret from the pilot seat will activate freelook for the pilot. Press Z to deactivate freelook to continue normally
Early days. ITS ALPHA! CIG can take as long as they need to get it right. I've had more fun in this buggy alpha than i've had in any AAA game from a publisher. Everyone knows you leave bug fixing until the end of development. I'm so excited i'm going to pledge more!
Did i miss any talking points?
Give it two more years I'll be able to use Star Citizen to measure my remaining lifespan in convenient units.
I don't listen to what CR say because he is always too vague about details/technical aspect and gives dates that can't be met. It's not the same that lying for me. The vast majority of managers I've worked with since 20 years can't give tech details and give dates that can't be respected. It's not a surprise for me that CR is the same kind of person.
I like the vision of the game he have and I'm not afraid by the way he manages the project. But for the practical aspect, I refer to the devs. Elon Musk has a vision with SpaceX and SpaceX wouldn't have existed without him. But if I need details on how its rocket works, I will not ask or listen him.
What can I say ? That I'm bored to death and I feel cheated when I test the alpha ? I have fun, it's not my fault. And I 'm not the only one to have fun...
We don't know how many players will be in one instance at release. If it's 50, capital ships are a real problem. If it's 200 it can be adapted with private combat zone mechanics and overhaul the way the game works. If it's 500/1000, it's not a problem.
CIG is slow and the game may not be released before 3/4/5/x years. It's a problem but also can be a solution if you think CIG is bad at netcode. SC run on Lumberyard with AWS. Amazon have ambitions on the Game engine market and their services for game will continue to evolve in speed, power and bandwidth. The bandwith of players evolve too, in France atm a LOT of gamers have 300 Mb/s up and down. Years after years the AWS will allow more and more players in one instance for SC. Do not forget that SC is also the sole showcase of MMO for them for Lumberyard with New World (smaller project).
Last edited:
It's really not a matter of bandwidth - but latency and transit.
I just want to deep dive into this bit:
CIG certainly do talk a lot
'Building Blocks' is just their UI system though. It's a replacement for the prior clunky system using Flash. Finally moving away from Flash for your UI in 2020 is... not a revolutionary thing
I can understand how you may have interpreted it as an impressive thing. The mission team themselves did describe it as 'massive' for the mission system. But what they meant by that, it turns out, was that: It lets players interact with mission info on a screen. Rather than via the inner thoughts system.
'Massive.'
What this seems to boil down to is that BB includes an internal tool for designers to run up their own panel interfaces. Not anything that lets them 'create missions' per se (you still have to like, code / design the missions and such for novel functionality there). Handy, sure, and more flexible than the inner thoughts system. But really pretty standard game dev stuff.
They also mentioned some possible gameplay outcomes. Like they've dreamcrafted about being able to alter the info a player hands in for a bounty mission etc. But that all just seems to be theoretical UX/UI for, more importantly, theoretical mission aspects. Not some pipeline that's about to start pumping out the missions itself. It just sounds like... a nice idea. Nothing more at that point.
This just seems like a classic case of CIG enthusing about some very basic internal processes, and some hypotheticals, with some PR spin added by dear Lando, and it being interpreted as some kind of novel innovation. Which presages some kind of impressive future pipeline output...
As you mention, probably worth seeing if they can get the second system working first
CIG certainly do talk a lot
'Building Blocks' is just their UI system though. It's a replacement for the prior clunky system using Flash. Finally moving away from Flash for your UI in 2020 is... not a revolutionary thing
I can understand how you may have interpreted it as an impressive thing. The mission team themselves did describe it as 'massive' for the mission system. But what they meant by that, it turns out, was that: It lets players interact with mission info on a screen. Rather than via the inner thoughts system.
'Massive.'
What this seems to boil down to is that BB includes an internal tool for designers to run up their own panel interfaces. Not anything that lets them 'create missions' per se (you still have to like, code / design the missions and such for novel functionality there
). Handy, sure, and more flexible than the inner thoughts system. But really pretty standard game dev stuff.They also mentioned some possible gameplay outcomes. Like they've dreamcrafted about being able to alter the info a player hands in for a bounty mission etc. But that all just seems to be theoretical UX/UI for, more importantly, theoretical mission aspects. Not some pipeline that's about to start pumping out the missions itself. It just sounds like... a nice idea. Nothing more at that point.
I also love just how perplexed / worried the Senior Live Designer on the left looks throughout some of these sections 
I wonder if it's a coincidence that he's since left the company)
I wonder if it's a coincidence that he's since left the company
)This just seems like a classic case of CIG enthusing about some very basic internal processes, and some hypotheticals, with some PR spin added by dear Lando, and it being interpreted as some kind of novel innovation. Which presages some kind of impressive future pipeline output...
As you mention, probably worth seeing if they can get the second system working first
I know. Add those two words in my sentence. I dont' know for the US but in France the statistical average latency if better year after year.
…and also bandwidth, given that this is SC we're talking about. The sheer amount of garbage they've demonstrated a need to send and resend at every opportunity is telling. And let's not forget the DDoS patcher that their crack network team put together.
Oh, and time will not fix this. What CI¬G needs is already available and has been for decades. Plural. It's just that CI¬G has no clue how to make games, and the only way for time to fix that problem is if, over time, CI¬G goes away an takes SC with it.
Last edited:
You're skipping over the detail that CR raised money based on those promises. And he paid his incompetent wife a nice fat salary with that money.
Well you said they'd get only to 200 or 300. Now you are changing the numbers. Sorry, but 200 per instance means the selling of $3000 dreams is a scam.
Which just shows how little Amazon cares about Lumberyard or SC, surely? Or are we still pretending this is somehow a great "partnership" and Bezos at any point in his day thinks about Chris?
…and also bandwidth, given that this is SC we're talking about. The sheer amount of garbage they've demonstrated a need to send and resend at every opportunity is telling. And let's not forget the DDoS patcher that their crack network team put together.
Well yes, you are right of course, but I was looking on the bright side that CI-G at least stopped using literal megabytes of XML to describe turret states several years ago.
Lumberyard is essentially DOA. Crucible went back into beta after a terrible release, New World has been delayed (again!) to 2021 and the unannounced games are struggling to find footing. It really feels like Amazon doesn't want it anymore, and even developers don't want to use it; Star Citizen doesn't even really use it, but instead uses a fork of CryEngine (3.6.4 according to CI's own filings) that they were able to license for free from Amazon, per the legal docs from the Crytek v CIG lawsuit.
CI argued that there can be no case regarding the GLA, since the Lumberyard agreement also meant Amazon is able to "give out" previous versions of CryEngine at will that are "bundled" with Lumberyard.
The debate continues on just what branch of CryEngine CI continues to use; legal papers filed by CI's own council claim 3.6.4; Ben Parry stated it was a fork from the 3.7.x branch; other CI employees claim it's 3.8+ or Lumberyard itself.
Most open development ever, indeed.
CI argued that there can be no case regarding the GLA, since the Lumberyard agreement also meant Amazon is able to "give out" previous versions of CryEngine at will that are "bundled" with Lumberyard.
The debate continues on just what branch of CryEngine CI continues to use; legal papers filed by CI's own council claim 3.6.4; Ben Parry stated it was a fork from the 3.7.x branch; other CI employees claim it's 3.8+ or Lumberyard itself.
Most open development ever, indeed.