Star Citizen

A place for all the city-builders, world-creators, transport managers, block-arrangers and intergalactic explorers.
User avatar
stingtwo
Posts: 100
Joined: 05 Nov 2015, 17:47

Re: Star Citizen

Post by stingtwo » 27 Mar 2017, 23:38

boars wrote: There's definitely been some ion-storm esque moments... but most of the money stuff is guesstimation so I take that stuff with a grain of salt.
The helmets in the twitter post I linked come directly from the company that makes them http://facewaretech.com/pricing/#hardwarepricing
Mo-cap isn't cheap, estimated costs are all over the place but the average works out to $25 per second of footage captured and Imagination is where they were doing all their mo-cap, that place is considered the most expensive in the industry.
And the London studio, there is a requirement to file financial reports in the UK, costs of operating it come from them https://beta.companieshouse.gov.uk/comp ... ng-history

The long running joke on somethingawful is the $18,000 coffee machine and the EST. $20,000 entrance door that CIG claimed was made with plywood and a garage door at the LA studios, everything else questionable is a guesstimate I agree but quick Google searches does come up with results on the cost of the items in question, lets not be idiots and go "RARRRHHH now your stating CIG cannot spend any money on office furniture!" No but when something simple as a coffee table costs $800, questions are allowed to be raised into the spending habits.
As for the graphics engine, I took that as more of a dropping support for DX12 rather than changing the rendering engine so to speak. Thought they were supporting both - they'd always promised win7 etc support so not sure how they could have gone solely DX12 if that was the case, since it's win10 only.
I was given shit by a few people on IRC when I made fun of SC announcing the switch to DX12 because it did break their kickstarter promise and why they picked it over OpenGL let alone Vulkan still was a strange decision as cryengine supports OpenGL out of the box and it still takes months of coding to change, even if they said "yeah we will support linux at a later date".
It'll most likely release in some form or another, will it be worth playing, that's the real question :P

Some things like flight times in systems sound somewhat stupidly long.
Reminds me of how annoying flight paths in games such as WOW back in the day could get damned boring after a while and they had things to look at. I'm not sure how exciting a blur in the stars is going to stay interesting, sure you'll be able to chat and such whilst doing so but if you don't want to, piloting your single seat fighter any large distance is going to be dull if it takes too long.

The flipside to that problem is if it's too easy to get everywhere you lose the feeling that space is vast.

So many unknowns...

If 3.0 manages a release in the next few months 2017 should be an ok year. (by slow CIG standards)... for progress in StarCitizen.
If even a slice of singleplayer doesn't appear soon, the Squadron 42 saltyness will probably explode.

I'm willing to wait because I'm still dreaming of the game the Infinity space guys were talking about building back in 2004 or so... I might have to wait another 5 years at this rate :P

One thing that STILL grinds my gears though is the downloader.
For the LOVE OF GOD AND ALL THAT IS HOLY. SDFSDS@#$@#$@ @#$@#$@# Just fix this @#$%. hahaha.
Seriously, I don't do PTU server or patch the main client all that often because it's absurdly time consuming and data intensive.
Thankfully the house I recently got a contract on has NBN... which solves the problem for me but I cannot believe it's still an issue.
We can look at this from a game design perspective, Look at many open world games and why they fail to fill their world with interesting stuff when compared to GTA 5.
GTA 5 you can take the time away from doing the campaigns to look around and see whats going on, can head north and listen to country folks talk about barn yard dancing or follow the only black guy in the town before a cop shows up. drive near downtown and watch gangs dish it out, head into central, steal a rich persons car and get chased by cops for a few minutes, wait till dark with a cop car, find a bunch of hookers, pull up the car that picks one and just watch her piss bolt into the sea. Things are happening in world not even because of your influence. Lets look at something like Wildlands, I am finding it very fun with friends, but once everything in a sector of the map is done, like missions, side missions, collectibles, what else is there in the area,nothing was going on before we killed everyone, nothing happens after, we don't revisit the area again, it's onto the next area, rinse repeat, this isn't new or unique to open world games either.

Lets take all the technical issues Star Citizen has now cause those "can" be fixed aside because my beef isn't it being a buggy game. Whats the design goals they are doing? how many players can roam around in the world? planets? what does landing on those offer? How does one help a transport carrier? It boils down to "This new feature(things like the biodome), what does it offer for players?" We don't get told anything, Chris Roberts says so many things and yet so little about how far into a feature they are done and players like myself can only go on whats actually existing.

Is the offer it's going to be the same as Freelancer like everyone is hoping or Wing Commander: Privateer(which was one of the 1st PC games I played) I had hoped which is why I like so many had put money into this, then pulled out because the goals posts shifted so frequently they are now in a new stadium.

Even if it was all a buggy mess that boiled down to pulling a bunch switches that have to be done in order, it would be more than whats currently in the PU, 2 stations, a single mission that thats about 3-5 minutes total to get to and hit a switch, a big bennys machine you can kick over and clothes shopping.
0 x

User avatar
boars
Posts: 393
Joined: 08 Nov 2015, 11:33
Location: Q. OO-su-to-ra-re-ah
Contact:

Re: Star Citizen

Post by boars » 02 Apr 2017, 02:59

Your're right, Imagination Studios is expensive... so maybe a 30k headset is actually not that bad an option for doing further mocap captures? After a few sessions, you'll have paid itself off by not having to go into a studio that costs ludicrous amounts of money?

The latest ATV featured talk about the mocap footage and basically said it was hard to use so they just cut it up into smaller pieces for purpose, didn't sound like they threw it out. Just wasn't easy to use it for non cut-scene type animations but was still used. Was a fantastic episode actually, looking at how they blend animations together was pretty damned impressive. Talk about time consuming though, crazy. https://www.youtube.com/watch?v=Bbq2Crt ... u.be&t=515

Physics grid refactoring is done... apparently: https://www.youtube.com/watch?v=Bbq2Crt ... u.be&t=423
Looks like something that was definitely needed for 3.0 to go live.

The quest chain is now about 9 or so missions long, that starts at the ICC ScanHub satellite - so probably a little longer than 3-5 mins. Although there's certainly some exciting button pushing to get the black boxes talking after you kill anything about. You also rescue and assorted other pretty standard space game missions. Not overly exciting the 3rd or 4th time thru :lol:

There's also one other minor mission where you go explore an abandoned space station to work out what happened (PI Wanted). The rest, last I checked were just the basic restart satellites called CommArray SCC, bit repetitive those. Yawn :/

All in all there's been very little increase in the amount of missions to do - hopefully that's because they're hard at work on s42 instead.

There are 5 (4 if fuelstops don't count) types of stations:
Port Ollisar, Grim Hex, Kareah, Covalex Shipping Hub and CryAstro Refueling stations - think there's 3 or 4 of these - mind you, you cannot go inside of it... so it's hard to call it a station but eh.

I guess you feel only Ollisar and Hex rate as stations though... :) 'fair nuff.

annnnd 2.6.2 is now live (https://robertsspaceindustries.com/comm ... 62-Is-LIVE) featuring:

- added support for non-traditional screen resolutions like very wide screens using 21:9 aspect ratio
- re-enabled support for TrackIR.
- Multiplayer Mega Map
- Serialized Variables
- Drake Buccaneer ship now flyable

If you don't have a Buccaneer, you probably wont notice anything new... downloading it now because I'm a sucker for wasting my quota - another 27g+ patch, sigh.




Apparently the drake Dragonfly is ready to fly but is awaiting 3.0 or some patch that allows planetside.
Seems the ship pipeline is on track, even if the core game one isn't.
0 x

User avatar
boars
Posts: 393
Joined: 08 Nov 2015, 11:33
Location: Q. OO-su-to-ra-re-ah
Contact:

Re: Star Citizen

Post by boars » 15 Apr 2017, 02:00

Apparently the two primary capital UEE ships in S42 are completed (fully functional), Javelin and Idris.
The Javelin promo video is in the latest ATV - supercut version below:


Also should be getting a 3.0 schedule today... so looks like it might just be the next patch.

Still no word on delta patcher though :wtf:
0 x

User avatar
Otto-matic Reiffel
Posts: 546
Joined: 05 Nov 2015, 23:23

Re: Star Citizen

Post by Otto-matic Reiffel » 15 Apr 2017, 03:45

Javelin seems a little too roomy for a spaceship. Might be the odd perspective from few crew but the corridors and stuff seemed 3+ metres wide and the rooms are huge.
Fun fact: At 345 metres long the Javelin class destroyer is bigger than any current day aircraft carrier.
0 x

User avatar
stingtwo
Posts: 100
Joined: 05 Nov 2015, 17:47

Re: Star Citizen

Post by stingtwo » 15 Apr 2017, 21:50

boars wrote:
15 Apr 2017, 02:00

Also should be getting a 3.0 schedule today... so looks like it might just be the next patch.

Still no word on delta patcher though :wtf:
Well they did put out this and it's an amazing pile of spaghetti on the wall.

https://robertsspaceindustries.com/medi ... _Lines.png
0 x

User avatar
bender
Posts: 105
Joined: 05 Nov 2015, 18:08

Re: Star Citizen

Post by bender » 16 Apr 2017, 00:13

CIG released their schedule report:

https://robertsspaceindustries.com/schedule-report

Looks like 3.0 is scheduled for the end of June or early July. Which is about what I was expecting.
Image
The page has a pretty in depth video explaining/outlining their production and scheduling. The page itself goes into more specifics. I have to say, the amount of detail CIG has provided in this report has really impressed me.
boars wrote:
02 Apr 2017, 02:59
Your're right, Imagination Studios is expensive... so maybe a 30k headset is actually not that bad an option for doing further mocap captures? After a few sessions, you'll have paid itself off by not having to go into a studio that costs ludicrous amounts of money?
I was also thinking that Chris Roberts mindset for this game is for long term considerations. He intends (hopes) the game to go for 10+ years, so maybe CIG bought the headsets with those considerations in mind? If CIG is regularly adding content (short stories/quests etc) to the game using mocap, those headsets will more than pay for themselves.
0 x
Asus P8P67|2600K @4.5GHz-EK WC|16GB DDR3|Crucial MX100 512GB|HD6970 2GB-Aquacomputer WC|MCP35X|Obsidian 800D|X-fi Titanium|AD900

User avatar
boars
Posts: 393
Joined: 08 Nov 2015, 11:33
Location: Q. OO-su-to-ra-re-ah
Contact:

Re: Star Citizen

Post by boars » 16 Apr 2017, 02:45

stingtwo wrote:
15 Apr 2017, 21:50
Well they did put out this and it's an amazing pile of spaghetti on the wall.

https://robertsspaceindustries.com/medi ... _Lines.png
If you watch the video, you'd see that's nothing compared to the sheer amount of tasks they're tracking in Jira and MS Projects. Holy shit. I've worked on some projects I thought were a little bit involved in Jira but nothing of that magnitude. FMD. Getting some form of schedule out of that amount of crazy is quite impressive - even if it is inaccurate all the time. :lol:

Was actually a surprisingly interesting video.
Otto-matic Reiffel wrote:
15 Apr 2017, 03:45
Javelin seems a little too roomy for a spaceship. Might be the odd perspective from few crew but the corridors and stuff seemed 3+ metres wide and the rooms are huge.
Fun fact: At 345 metres long the Javelin class destroyer is bigger than any current day aircraft carrier.
True its too big but I guess you have to take fun into account and also the rule of cool seems to win out a lot in Sci-fi.

I prefer it this way... although I'd love to see a game based on the expanse... whether that'd actually be entertaining to play though, probably not. flying around for hours on end for combat that takes mere seconds. heh.
0 x

User avatar
stingtwo
Posts: 100
Joined: 05 Nov 2015, 17:47

Re: Star Citizen

Post by stingtwo » 16 Apr 2017, 16:57

boars wrote:
16 Apr 2017, 02:45

If you watch the video, you'd see that's nothing compared to the sheer amount of tasks they're tracking in Jira and MS Projects. Holy shit. I've worked on some projects I thought were a little bit involved in Jira but nothing of that magnitude. FMD. Getting some form of schedule out of that amount of crazy is quite impressive - even if it is inaccurate all the time. :lol:

Was actually a surprisingly interesting video.
It's an abortion, look at it, lets say each line represents a group to tackle each thing, why would a network team then go work on graphics, why are ships(the only thing they can manage, going to take longer than working on fixing the utterly broken netcode, why is fuel 2.0 mentioned twice.

But it's all pissing in the middle of the ocean at this point, it's all fantasy, they have been releasing this kind of stupid graph for months and the only thing that changes are the dates. That and locking their forums up :lol:

Acutally, just went ahead and watched the video, cargo is apparently hard to make work in a videogame
0 x

User avatar
boars
Posts: 393
Joined: 08 Nov 2015, 11:33
Location: Q. OO-su-to-ra-re-ah
Contact:

Re: Star Citizen

Post by boars » 16 Apr 2017, 17:23

stingtwo wrote:
16 Apr 2017, 16:57
It's an abortion, look at it, lets say each line represents a group to tackle each thing, why would a network team then go work on graphics, why are ships(the only thing they can manage, going to take longer than working on fixing the utterly broken netcode, why is fuel 2.0 mentioned twice.
Yes, that's what the colour code (key) is for because it's a condensed list and each row clearly doesn't mean anything of the sort. I'm pretty sure you realise this and are just being somewhat flippant though.

Impossible to know why fuel is there twice, in those instances as one is cut short and even if the Jira Epic title was complete it wouldn't necessarily give you enough context. Not to mention fuel is on there more than twice, spotted at least three.

Even ordering it all by date and colour coding, it's just too damned much to consume without spending far too much time or having any real understanding of what each task is or what potential issues lay buried amongst them. Not to mention the knock on effect of missing one or many of said Epics.

http://i.imgur.com/PCvrcYb.jpg
stingtwo wrote:
16 Apr 2017, 16:57
But it's all pissing in the middle of the ocean at this point, it's all fantasy, they have been releasing this kind of stupid graph for months and the only thing that changes are the dates. That and locking their forums up :lol:
They have, the dates have changed, patches have also released. Such is video game development. How many games have delays, practically all of them and those are just the ones we're aware of in the "last mile".

Not saying they haven't been terrible with things like "weeks not months" and the vertical slice of s42 disappearing into vapour before Christmas... they're certainly not faultless. Frustrating but progress is being made, it's just at a seemingly glacial pace at times. A failure of timeliness, no question but this is Chris Roberts, he hasn't exactly been known as Mr On-time in the past, not even prior to the whole kick starter.
stingtwo wrote:
16 Apr 2017, 16:57
Acutally, just went ahead and watched the video, cargo is apparently hard to make work in a videogame
It's not hard when it's just some text on a screen like in EVE or Elite... or probably even in Euro Truck Simulator... but shit me if they haven't bitten off a bit much in the way they're trying to do it here. Might as well be called cargo management simulator. As soon as the cargo becomes physical objects, it becomes a major pain in the arse. The sheer tetris annoyance it could prove to players manually shipping things. I want the crate in the middle of the pile because it'll sell for $2 more here... but then, it'll take me 4 minutes to get to it... :wtf:

I mean they've only got to work out (what the status of each of these is, nobody but they know):
  • How to manually place and lock down cargo
  • How to get AI to automatically unload/load
  • What happens if cargo isn't locked down properly
    • can it damage ship?
    • damage cargo?
    • damage players?
    • does cargo have damage states?
    • can cargo explode?
    • if cargo falls out of the ship does it persist? how long for?
  • ownership of cargo, how does one obtain it, lose it?
  • how to tell if cargo is stolen?
  • how do you salvage ship parts after taking out another ship and put them in your cargo hold without cargo containers? can you?
  • can cargo be shot from other ships rounds penetrating (or internal damage from other sources such as internal explosions or explosive rounds from other ships, how to deal with this?
  • can the system locking containers to floors be damaged in combat or via EMPs?
  • can small arms fire on ships cause damage?
  • can anyone walk onto a ship and pickup cargo and walk out with it? some sort of anti-theft system? how do pirates circumvent this system?
  • is it possible to jettison cargo, if so, how?
  • if you go beyond your G force rating by turning off the system, will this cause cargo to lose the ability to stick to the floor? Additionally do contents have G force limits or ratings?
I mean that's not even an exhaustive list of issues they have to deal with, making it a text only option would have been far far easier :problem:
0 x

User avatar
stingtwo
Posts: 100
Joined: 05 Nov 2015, 17:47

Re: Star Citizen

Post by stingtwo » 17 Apr 2017, 07:04

The cargo thing, no it isn't, they are not inventing some magical new wheel here. Wildlands has this, Call Of Duty has this, Payday 2 has this. Item has value, take item to where game tells you where to go,when game says it's at location, have whatever the value of the item in question was, money/xp/another item. It's a flag in any game with Capture The Flag mode, it's a item bag in payday 2, it's a rebel supply in wildlands. And for the record, it will be text on a screen, they can say otherwise but it's no different to every other game in existence.

Saying "oh they are trying to get cargo in the game" is pointless, whats it's purpose?

Lets do a basic transport mission. One planet to another transporting some Kilrathi slaves. How do you travel? multiple Waypoints in space, earth-single universe then to the destination planet? Who can attack me, Players, AI or both? Can I bring some people to help protect the transporter? How is that done, whats their incentive, do I bring up a menu ask them to join for a share of the earnings or is the mission profit split equally? Are the missions entirely for Squadron42? If you start a transport mission, does a piracy side mission open up for others in the server instance? No hints/mentions on any of this ever.

Everything you posted about the cargo physics is dumb hand waving that can be done at a later point and completely ignores the issue on how do they come into play, side missions or otherwise. But for sake of argument, to get every thing you just posted about the transporting done, the following is required, AI, Advanced flight controls, proper physics, damage, not to mention about 20 other things that need to be designed and finalized 100% before you tell a highly possible underpaid programmer to work on getting cargo to do everything you just asked.

This cargo crap is just a long line of CIG saying they have an idea, show it off in 3dmax and say "we are thinking about this being put in the game", backers eat it up till they slowly notice it's no where to be found in the game because Chris Roberts saw a trailer for a new videogame, and onto the next thing and the cycle continues.
0 x

User avatar
Matty
Posts: 1109
Joined: 05 Nov 2015, 21:41

Re: Star Citizen

Post by Matty » 17 Apr 2017, 17:24

Otto-matic Reiffel wrote:
15 Apr 2017, 03:45
Javelin seems a little too roomy for a spaceship. Might be the odd perspective from few crew but the corridors and stuff seemed 3+ metres wide and the rooms are huge.
Fun fact: At 345 metres long the Javelin class destroyer is bigger than any current day aircraft carrier.
Only just though, by a few meters.
0 x
Matter of fact, I didn't even give you my coat!

User avatar
Nekosan
Posts: 1099
Joined: 09 Nov 2015, 19:39
Location: GoN Refugee Camp

Re: Star Citizen

Post by Nekosan » 17 Apr 2017, 22:27

stingtwo wrote:
17 Apr 2017, 07:04

This cargo crap is just a long line of CIG saying they have an idea, show it off in 3dmax and say "we are thinking about this being put in the game", backers eat it up till they slowly notice it's no where to be found in the game because Chris Roberts saw a trailer for a new videogame, and onto the next thing and the cycle continues.

This is why i no longer follow anything official to do with the game and i just pop in here once every 6 months. The worst thing about game dev cycles like this is the fans who just circlejerk ideas and over hype everything with imaginary features. Watched this happen with the devs of Vanguard a good decade + ago, end well it does not.

The base functions in a game and engine need to work properly before the circlejerk content gets worked on, otherwise it all fails.
0 x
steamcommunity.com/id/krautpants

User avatar
boars
Posts: 393
Joined: 08 Nov 2015, 11:33
Location: Q. OO-su-to-ra-re-ah
Contact:

Re: Star Citizen

Post by boars » 17 Apr 2017, 23:29

stingtwo wrote:
17 Apr 2017, 07:04
The cargo thing, no it isn't, they are not inventing some magical new wheel here. Wildlands has this, Call Of Duty has this, Payday 2 has this. Item has value, take item to where game tells you where to go,when game says it's at location, have whatever the value of the item in question was, money/xp/another item. It's a flag in any game with Capture The Flag mode, it's a item bag in payday 2, it's a rebel supply in wildlands. And for the record, it will be text on a screen, they can say otherwise but it's no different to every other game in existence.
While I'm familiar with CTF I had no idea about some of those games... so I checked out the following videos:
Payday... https://www.youtube.com/watch?v=8RSFSgRU4Xw
Wildlands supplies... https://youtu.be/7dSuqOiGSXI?t=1237

Neither of those games demonstrate anything beyond a basic CTF game, wildlands didn't do even that much really other than mark items as "tagged".
I'm assuming they probably just disappear while you're not there?

If SC was supposedly planning on just having a simple system like the above then sure, prior tech exists and sure, should be no issue to bang that out.
Heck it might even be iteration one, then they may, or may not iterate on top of that.
Maybe they'll run out of money, ELE and all that.

No one has proof that they are not working towards their goals. I don't think saying that 3Ds max demo's makes things non-existent conversely I'm not saying anything necessarily exists - they're just not that open with their development, even when they say they are.
It also neither proves nor disproves their intentions, all we can do is take them at their word or not, as the case may be with you.
And yes their words are somewhat more aspirational than... well planned.

Besides 3dSmax/Maya placement, they have also demo'd rudimentary pickup of items and placement (including rotation) of them (in cryengine/starengine). A feature similar to games like Landmark (is that still a thing - *googles* - as of feb 21 it is no longer a thing) and maybe even games like space engineers - except in those games you don't physically pickup the object, hold it and place it down. That's probably fairly trivial to do but it'd also be inherently annoying in a building game to only hold one item at a time. They've technically already got some of this sort of basic building functionality in Hangars, with nodes you can place things on.

stingtwo wrote:
17 Apr 2017, 07:04
Saying "oh they are trying to get cargo in the game" is pointless, whats it's purpose?

Lets do a basic transport mission. One planet to another transporting some Kilrathi slaves. How do you travel? multiple Waypoints in space, earth-single universe then to the destination planet? Who can attack me, Players, AI or both? Can I bring some people to help protect the transporter? How is that done, whats their incentive, do I bring up a menu ask them to join for a share of the earnings or is the mission profit split equally? Are the missions entirely for Squadron42? If you start a transport mission, does a piracy side mission open up for others in the server instance? No hints/mentions on any of this ever.
There is a dearth of any substantial information - a lot has to be pieced together- you have to trawl forums for dev posts and watch practically every damned video to find out anything anyone has said.
Then you have to wonder, what of this is pipe dreaming, what is reality - who besides CIG actually knows, the answer to that is no one.
There's been a distinct lack of design documents for a long time, or at least publicly accessible ones.
No doubt, not even debating this.

The only logical stance in that regard is to shrug and wait.
stingtwo wrote:
17 Apr 2017, 07:04
Everything you posted about the cargo physics is dumb hand waving that can be done at a later point and completely ignores the issue on how do they come into play, side missions or otherwise. But for sake of argument, to get every thing you just posted about the transporting done, the following is required, AI, Advanced flight controls, proper physics, damage, not to mention about 20 other things that need to be designed and finalized 100% before you tell a highly possible underpaid programmer to work on getting cargo to do everything you just asked.
Problem is with no info, all one can do is guess.
Yes, it leads to people over thinking and believing it'll be what they want rather than what potentially eventuates.
They may have their plans already sorted but whether or not some things work the way they intend or not... best laid plans and all that.

But if you took everything at face value, that'd be a shitload of work... a stupid amount. That's all I'm saying.
If they did (or tried to do) what they said they would in various vague and dreamy statements then it's not so simple.
What ultimately becomes reality may well be EVE/Elite like cargo hauling.
stingtwo wrote:
17 Apr 2017, 07:04
This cargo crap is just a long line of CIG saying they have an idea, show it off in 3dmax and say "we are thinking about this being put in the game", backers eat it up till they slowly notice it's no where to be found in the game because Chris Roberts saw a trailer for a new videogame, and onto the next thing and the cycle continues.
I'm hoping Chris has no free time to look at other games anymore... or maybe he's seen them all...
I don't know, I'm thinking Erin may actually get him in line and s42 might actually see the light of day.
A lot of the big things they need to solve are SC specific and shouldn't (in theory) cause issues with s42... well, that's the dream.

Maybe by the end of July we'll also be able to have a look at SC3.0.
I doubt it'll be end of June, one key thing causes a blocker and boom to that massive schedule, something that big is probably only valid for an hour before it's significantly out of date :wtf:
0 x

User avatar
bender
Posts: 105
Joined: 05 Nov 2015, 18:08

Re: Star Citizen

Post by bender » 18 Apr 2017, 19:15

Nekosan wrote:
17 Apr 2017, 22:27
stingtwo wrote:
17 Apr 2017, 07:04

This cargo crap is just a long line of CIG saying they have an idea, show it off in 3dmax and say "we are thinking about this being put in the game", backers eat it up till they slowly notice it's no where to be found in the game because Chris Roberts saw a trailer for a new videogame, and onto the next thing and the cycle continues.

This is why i no longer follow anything official to do with the game and i just pop in here once every 6 months.
You lie, you made a post about 4 months ago! I bet you secretly view every official SC post/video. /joke
Nekosan wrote:
17 Apr 2017, 22:27
The base functions in a game and engine need to work properly before the circlejerk content gets worked on, otherwise it all fails.
Yeah, this so much. I have been saying it for years. Last I looked on the official forums (over a year ago), they are just full of "its going to be amazing when we can do <insert amazingly overestimated feature here>", or "what are you going to do on your <insert ship here> when the game is released". Every time I mentioned the lack of core tech in the official forums, the white knights came out and shouted me down: "Its alpha, CR knows what he's doing, trust in christ roberts, there are no publishers limiting the games potential, it will be amazing, the game is sticking it to the publishers etc.".

IMHO, 3.0 will be a big indicator as to how well they are doing with their core tech. They are already somewhat behind, but maybe CIG can get themselves on track? I still think the biggest hurdle for SC will be the cross region play (if they even still intend to do it). CIG still haven't given an indication as to how they will manage players from EU/US/AUS playing together without detrimental lag. In a skill based game, latency issues will have a huge impact on the experience and ability to play. Being an MMO, if the general game experience sucks due to network issues and latency, who cares about cargo and other 'unprecedented state of the art' non-core features. The fact that it has taken them this long to start sorting out their network issues doesn't give me much hope they can manage it - at least in any reasonable timeframe.
0 x
Asus P8P67|2600K @4.5GHz-EK WC|16GB DDR3|Crucial MX100 512GB|HD6970 2GB-Aquacomputer WC|MCP35X|Obsidian 800D|X-fi Titanium|AD900

User avatar
stingtwo
Posts: 100
Joined: 05 Nov 2015, 17:47

Re: Star Citizen

Post by stingtwo » 19 Apr 2017, 17:13

boars wrote:
17 Apr 2017, 23:29

While I'm familiar with CTF I had no idea about some of those games... so I checked out the following videos:
Payday... https://www.youtube.com/watch?v=8RSFSgRU4Xw
Wildlands supplies... https://youtu.be/7dSuqOiGSXI?t=1237
Maybe watch 5minutes further than the time-stamp in the wildlands video you linked, not the tagging of items, but taking something from point a to b but maybe a better example would be this https://www.youtube.com/watch?v=F_wchRLMSE4, As for the payday 2 bag...yep, whats different? Item goes to from it's location to where the game tells you where to take it. I see zero difference to what CIG wants to do, adding a vehicle, making the AI do a scripted thing at the checkpoint, having to go through some mundane in-ship forklift system or even bringing up a map to mark the location is not new, using big buzz words to try and explain why it's not working doesn't and won't make me feel sympathy for them, not this damn late in development, not when this was pitched in the kickstarter, it displays incompetence, bad project management and leadership.

I used CTF mode as an example since you decided to say taking items from somewhere to another place is nothing but text when the flag in that mode has been a physical object since quake has been around. To say payday/wildlands transporting of items is "just CTF" is laughable. It's dumber than saying ARMA and battlefield 4 are the same game.
No one has proof that they are not working towards their goals. I don't think saying that 3Ds max demo's makes things non-existent conversely I'm not saying anything necessarily exists - they're just not that open with their development, even when they say they are.
It also neither proves nor disproves their intentions, all we can do is take them at their word or not, as the case may be with you.
The rest, yeah it's all speculation on what they are doing, outside of ship sales to milk a few more dollars out of people but I'll take a moment to comment on this though.

There is a fine line between being ambitious and being deceitful to both paying and potential customers and yeah, I have felt CIG has been doing the latter for a long time but can I and so many others who have decided to curse this game be blamed for thinking that. If you pitch a feature in the game and it's not in it, you are flat out deceiving them. No Man Sky is a recent example of this.

If the community manager says he has played Squadron 42, I'm going to take his word on that. If Chris Roberts shows what planet landings are and players don't get it in whatever new build, it's deceit. It's the Bengal Carrier you posted on page 6 that flat out doesn't work in Cryengine
bender wrote:
18 Apr 2017, 19:15
IMHO, 3.0 will be a big indicator as to how well they are doing with their core tech. They are already somewhat behind, but maybe CIG can get themselves on track? I still think the biggest hurdle for SC will be the cross region play (if they even still intend to do it). CIG still haven't given an indication as to how they will manage players from EU/US/AUS playing together without detrimental lag. In a skill based game, latency issues will have a huge impact on the experience and ability to play. Being an MMO, if the general game experience sucks due to network issues and latency, who cares about cargo and other 'unprecedented state of the art' non-core features. The fact that it has taken them this long to start sorting out their network issues doesn't give me much hope they can manage it - at least in any reasonable timeframe.
You cannot change the distance between nations or beat the speed of light, lag between the US/AU/EU is always going to exists. Putting servers in the 2 nations and having them sync up into one large instance won't fix this, if anything it would make it worse.
0 x

User avatar
bender
Posts: 105
Joined: 05 Nov 2015, 18:08

Re: Star Citizen

Post by bender » 19 Apr 2017, 21:23

stingtwo wrote:
19 Apr 2017, 17:13
bender wrote:
18 Apr 2017, 19:15
IMHO, 3.0 will be a big indicator as to how well they are doing with their core tech. They are already somewhat behind, but maybe CIG can get themselves on track? I still think the biggest hurdle for SC will be the cross region play (if they even still intend to do it). CIG still haven't given an indication as to how they will manage players from EU/US/AUS playing together without detrimental lag. In a skill based game, latency issues will have a huge impact on the experience and ability to play. Being an MMO, if the general game experience sucks due to network issues and latency, who cares about cargo and other 'unprecedented state of the art' non-core features. The fact that it has taken them this long to start sorting out their network issues doesn't give me much hope they can manage it - at least in any reasonable timeframe.
You cannot change the distance between nations or beat the speed of light, lag between the US/AU/EU is always going to exists. Putting servers in the 2 nations and having them sync up into one large instance won't fix this, if anything it would make it worse.
There is a server architecture which can normalise the effect of latency so no player is disadvantaged more than another when playing from different regions - it levels the playing field. I wrote up a more detailed post about it in another forum a bit after 2.0 came out at the start of 2016:
For a long time I have been wondering how cross region play will work? Many organisations contain large numbers of players from different regions. For aussies, 250ms ping in the PU is basically impossible. How will EU/US/AU players play together in an enjoyable manner?

I was thinking about a way to obtain the best of both worlds. Local servers and global play. I came across a multi-server design called Mirrored Server architecture [1,2]. In this design, players connect to their local server in their respective regions. The local servers are then mirrored through a synchronisation process across regions. This means that whilst latency will still be somewhat notable for players across a different region, it wont disadvantage them - or at least both players will be disadvantaged equally.

For example, hits will generally register correctly for both players (from different regions). However, both players may experience scenarios of being shot when behind a wall after just moving there, or shooting through an enemy due to the enemies erratic motion. This would only be the case for players interacting from different regions, players from the same region wont notice any of these effects. The above is for a worst case scenario as there would be many cases where prediction is correct and players wont notice anything. This design also means that the game for all players, regardless of regions, should generally feel responsive. In comparison, currently players joining from a different region are often unable to fly as fast in tight spaces, or navigate tight hallways/doors quickly.

From my understanding, latency in general would also be reduced for cross region players. The local server only needs to wait for data to be sent from the other servers, rather than an entire round trip. Though there is likely some added latency for the more complex synchronisation process.

Code: Select all

Cross region:
[Client AU]--125ms-->[Server US]--125ms-->[Client AU]
total = 250ms

Same region
[Client AU]---20ms-->[Server AU]---20ms-->[Client AU]
total = 40ms

Cross region mirrored server
[Client US]---20ms-->[Server US]---20ms->[Client US]
                          |  100ms
[Client AU]---20ms-->[Server AU]---20ms->[Client AU]

AU <-> US = 140ms
AU <-> AU = 40ms
US <-> US = 40ms
Whilst I think such an architecture would work for star citizen, I have no detailed knowledge of their networking and game server/client design, so who knows. Maybe there's someone here who has much more experience that can comment? I do hope that CIG are at least working on some solution though. I think a lot of people, particularly orgs, will be disappointed on the game release if they cannot play together due to latency. Currently, I think star citizen will work reasonably well for servers in your own region (when they fix the server framerate issues). Outside, the experience will be significantly poorer. I really hope that doesn't remain the case.

[1] http://www.eecg.toronto.edu/~ashvin/cou ... umtr01.pdf
[2] https://www.comp.nus.edu.sg/~bleong/hyd ... icient.pdf
So its possible, though still not ideal. I'm starting to wonder if CIG do not intend to have cross region play and they are just not saying so, as many people wont like it. Based on an (somewhat recent) ATV episode where they talked about their server architecture, it looked like it was designed for a per region mesh instance based setup (no cross region play). They gave no indication that the architecture would be changed to support cross region play. Maybe they will support it eventually, but not upon initial release.
0 x
Asus P8P67|2600K @4.5GHz-EK WC|16GB DDR3|Crucial MX100 512GB|HD6970 2GB-Aquacomputer WC|MCP35X|Obsidian 800D|X-fi Titanium|AD900

User avatar
Otto-matic Reiffel
Posts: 546
Joined: 05 Nov 2015, 23:23

Re: Star Citizen

Post by Otto-matic Reiffel » 20 Apr 2017, 11:26

Have played a game where they used a similar system to slow all game speeds down to that of the slowest PC in multiplayer. Think it was Stronghold Crusader. That was very painful when playing with somebody that I'm not sure met the minimum specs for the game.

Can't imagine it would be good if somebody from Australia, South Africa and Chile were all in the same instance.
0 x

User avatar
stingtwo
Posts: 100
Joined: 05 Nov 2015, 17:47

Re: Star Citizen

Post by stingtwo » 20 Apr 2017, 15:56

bender wrote:
19 Apr 2017, 21:23
There is a server architecture which can normalise the effect of latency so no player is disadvantaged more than another when playing from different regions - it levels the playing field. I wrote up a more detailed post about it in another forum a bit after 2.0 came out at the start of 2016:
For a long time I have been wondering how cross region play will work? Many organisations contain large numbers of players from different regions. For aussies, 250ms ping in the PU is basically impossible. How will EU/US/AU players play together in an enjoyable manner?

I was thinking about a way to obtain the best of both worlds. Local servers and global play. I came across a multi-server design called Mirrored Server architecture [1,2]. In this design, players connect to their local server in their respective regions. The local servers are then mirrored through a synchronisation process across regions. This means that whilst latency will still be somewhat notable for players across a different region, it wont disadvantage them - or at least both players will be disadvantaged equally.

For example, hits will generally register correctly for both players (from different regions). However, both players may experience scenarios of being shot when behind a wall after just moving there, or shooting through an enemy due to the enemies erratic motion. This would only be the case for players interacting from different regions, players from the same region wont notice any of these effects. The above is for a worst case scenario as there would be many cases where prediction is correct and players wont notice anything. This design also means that the game for all players, regardless of regions, should generally feel responsive. In comparison, currently players joining from a different region are often unable to fly as fast in tight spaces, or navigate tight hallways/doors quickly.

From my understanding, latency in general would also be reduced for cross region players. The local server only needs to wait for data to be sent from the other servers, rather than an entire round trip. Though there is likely some added latency for the more complex synchronisation process.

Code: Select all

Cross region:
[Client AU]--125ms-->[Server US]--125ms-->[Client AU]
total = 250ms

Same region
[Client AU]---20ms-->[Server AU]---20ms-->[Client AU]
total = 40ms

Cross region mirrored server
[Client US]---20ms-->[Server US]---20ms->[Client US]
                          |  100ms
[Client AU]---20ms-->[Server AU]---20ms->[Client AU]

AU <-> US = 140ms
AU <-> AU = 40ms
US <-> US = 40ms
Whilst I think such an architecture would work for star citizen, I have no detailed knowledge of their networking and game server/client design, so who knows. Maybe there's someone here who has much more experience that can comment? I do hope that CIG are at least working on some solution though. I think a lot of people, particularly orgs, will be disappointed on the game release if they cannot play together due to latency. Currently, I think star citizen will work reasonably well for servers in your own region (when they fix the server framerate issues). Outside, the experience will be significantly poorer. I really hope that doesn't remain the case.

[1] http://www.eecg.toronto.edu/~ashvin/cou ... umtr01.pdf
[2] https://www.comp.nus.edu.sg/~bleong/hyd ... icient.pdf
So its possible, though still not ideal. I'm starting to wonder if CIG do not intend to have cross region play and they are just not saying so, as many people wont like it. Based on an (somewhat recent) ATV episode where they talked about their server architecture, it looked like it was designed for a per region mesh instance based setup (no cross region play). They gave no indication that the architecture would be changed to support cross region play. Maybe they will support it eventually, but not upon initial release.
In that situation, it's still not possible. the round trip still exists, instead of the players, it's now the servers willing to take the round trip, in Star Marine, a twitch based shooter, players would be reporting pings of the 20-80, unless the players you are going up against are on the same instance, there will is major lag in everything from hit detection, movement, to interaction with objects.

I admit to skimming the concept papers themselves but 2 things come into play, from a design standpoint they read like the servers were put on the same location, maybe even the same box to remove any delay having 2 servers synced up would suffer. It also was written in 2002, 2002 was hell for gaming clans looking for bf1942 servers in the US, so few had the network capacity to handle 64 players because up to that point, they only had to deal with 16 player arena shooters.

But if CIG wanted to handle more players in an instance, just fix up the broken netcode that uses 3 times more data than bf1 does.
Have played a game where they used a similar system to slow all game speeds down to that of the slowest PC in multiplayer. Think it was Stronghold Crusader. That was very painful when playing with somebody that I'm not sure met the minimum specs for the game.

Can't imagine it would be good if somebody from Australia, South Africa and Chile were all in the same instance.
Can you imagine this in any Shooter, let alone twitchy fast paced? You wouldn't hesitate to riot the 1st this happens to you.
0 x

User avatar
boars
Posts: 393
Joined: 08 Nov 2015, 11:33
Location: Q. OO-su-to-ra-re-ah
Contact:

Re: Star Citizen

Post by boars » 20 Apr 2017, 16:58

Time dilation, Eve does that too. Works ~ok~ for Eve*, terrible idea for twitch based combat though.

Short of creating a modem that makes use of quantum entanglement (which last I checked isn't possible) networking is going to be a huge issue. Having said that I've happily played planetside on the US server. It's not ideal but there's no escaping that reality sadly.

This is an area that cryengine is extremely bad at, so it'll be interesting to see what they've done/do with libyojimbo, lumberyard and whatever else they are making/using.

* Allows reinforcements that would normally be too far away to participate in a battle to actually get to the fight causing all sorts of issues for a huge surprise attack.
0 x

User avatar
Otto-matic Reiffel
Posts: 546
Joined: 05 Nov 2015, 23:23

Re: Star Citizen

Post by Otto-matic Reiffel » 21 Apr 2017, 10:19

TBF I used to play Day of Defeat on dial up with ~300-350 ping vs others on ADSL and did reasonably well. It can be done, just that the backend matching needs to make it unnecessary most of the time and where it's unavoidable, make it equally bad for everyone.
0 x

User avatar
bender
Posts: 105
Joined: 05 Nov 2015, 18:08

Re: Star Citizen

Post by bender » 21 Apr 2017, 15:26

stingtwo wrote:
20 Apr 2017, 15:56
bender wrote:
19 Apr 2017, 21:23
There is a server architecture which can normalise the effect of latency so no player is disadvantaged more than another when playing from different regions - it levels the playing field. I wrote up a more detailed post about it in another forum a bit after 2.0 came out at the start of 2016:
For a long time I have been wondering how cross region play will work? Many organisations contain large numbers of players from different regions. For aussies, 250ms ping in the PU is basically impossible. How will EU/US/AU players play together in an enjoyable manner?

I was thinking about a way to obtain the best of both worlds. Local servers and global play. I came across a multi-server design called Mirrored Server architecture [1,2]. In this design, players connect to their local server in their respective regions. The local servers are then mirrored through a synchronisation process across regions. This means that whilst latency will still be somewhat notable for players across a different region, it wont disadvantage them - or at least both players will be disadvantaged equally.

For example, hits will generally register correctly for both players (from different regions). However, both players may experience scenarios of being shot when behind a wall after just moving there, or shooting through an enemy due to the enemies erratic motion. This would only be the case for players interacting from different regions, players from the same region wont notice any of these effects. The above is for a worst case scenario as there would be many cases where prediction is correct and players wont notice anything. This design also means that the game for all players, regardless of regions, should generally feel responsive. In comparison, currently players joining from a different region are often unable to fly as fast in tight spaces, or navigate tight hallways/doors quickly.

From my understanding, latency in general would also be reduced for cross region players. The local server only needs to wait for data to be sent from the other servers, rather than an entire round trip. Though there is likely some added latency for the more complex synchronisation process.

Code: Select all

Cross region:
[Client AU]--125ms-->[Server US]--125ms-->[Client AU]
total = 250ms

Same region
[Client AU]---20ms-->[Server AU]---20ms-->[Client AU]
total = 40ms

Cross region mirrored server
[Client US]---20ms-->[Server US]---20ms->[Client US]
                          |  100ms
[Client AU]---20ms-->[Server AU]---20ms->[Client AU]

AU <-> US = 140ms
AU <-> AU = 40ms
US <-> US = 40ms
Whilst I think such an architecture would work for star citizen, I have no detailed knowledge of their networking and game server/client design, so who knows. Maybe there's someone here who has much more experience that can comment? I do hope that CIG are at least working on some solution though. I think a lot of people, particularly orgs, will be disappointed on the game release if they cannot play together due to latency. Currently, I think star citizen will work reasonably well for servers in your own region (when they fix the server framerate issues). Outside, the experience will be significantly poorer. I really hope that doesn't remain the case.

[1] http://www.eecg.toronto.edu/~ashvin/cou ... umtr01.pdf
[2] https://www.comp.nus.edu.sg/~bleong/hyd ... icient.pdf
So its possible, though still not ideal. I'm starting to wonder if CIG do not intend to have cross region play and they are just not saying so, as many people wont like it. Based on an (somewhat recent) ATV episode where they talked about their server architecture, it looked like it was designed for a per region mesh instance based setup (no cross region play). They gave no indication that the architecture would be changed to support cross region play. Maybe they will support it eventually, but not upon initial release.
In that situation, it's still not possible. the round trip still exists, instead of the players, it's now the servers willing to take the round trip, in Star Marine, a twitch based shooter, players would be reporting pings of the 20-80, unless the players you are going up against are on the same instance, there will is major lag in everything from hit detection, movement, to interaction with objects.
The cool thing with the design is there is actually no round trip between the servers. So not like this:

Code: Select all

           (request state)         (send state)
[AU server]----100ms--->[US server]----100ms--->[AU server]
The servers actually simultaneously broadcast (multicast) the change in their state each frame to each other:

Code: Select all

             (AU state)
           ----100ms--->
[AU server]             [US server]
           <---100ms----
             (US state)
So, when a server receives the state of another server, it rolls back its state to figure out what actually happened, reprocessing anything (hence only 140ms effective latency instead of 250ms). The only time this would need to happen, is if players from different regions needed to interact. This means that us here in AU could play on local servers, run around and interact with game objects (doors, cargo, NPC) and each other with minimal latency (40-80ms round trip). The only time latency would become an issue, is if a US/EU player attempted to interact (shoot, crash/ram, synchronise timing etc) with us or we with them. However, both players will experience the same latency and are therefore equally disadvantaged. So in a PvP scenario, the US player doesn't have the home server advantage.

This is assuming cross region play is a requirement for SC. Currently it seems the plan is to just have EU/AU players connect to US servers (or US to EU/AU etc) when wanting to play with friends/org members, and the experience would be significantly worse for them. Boarding an enemy ship at 250ms latency when a bunch of other players are at 40-80ms would just suck. Boarding an enemy ship at 140ms for only players interacting across different regions isn't great, but it would provide a better and fairer experience overall.
stingtwo wrote:
20 Apr 2017, 15:56
I admit to skimming the concept papers themselves but 2 things come into play, from a design standpoint they read like the servers were put on the same location, maybe even the same box to remove any delay having 2 servers synced up would suffer. It also was written in 2002, 2002 was hell for gaming clans looking for bf1942 servers in the US, so few had the network capacity to handle 64 players because up to that point, they only had to deal with 16 player arena shooters.

But if CIG wanted to handle more players in an instance, just fix up the broken netcode that uses 3 times more data than bf1 does.
I think the only reason this architecture hasn't become the standard for mmos is due to the reasonably high complexity of state synchronisation and consequently high server CPU demand. In 2002 it was a pretty major issue, but in 2017, with an order of magnitude more CPU power its much more doable.

The first paper only used local servers, but the second used a simulated server latency of 50 and 100ms for quake 3 servers. The the second paper provided examples where complex state synchronisation absolute worst case (4 server synchronisation) took 1000ms on 2002 tech. Now it would probably take 50-100ms or less for worst case (assuming current cpus are 10x faster than 2002 cpus, but its probably more like 30x). The average for 2 servers was 90ms, today that would be <10ms. There are also likely much better synchronisation algorithms than the ones used in the papers in 2002, reducing the latency even more. Multi-threading (one thread for each linked server) could also reduce the processing time by 2-4 times.
0 x
Asus P8P67|2600K @4.5GHz-EK WC|16GB DDR3|Crucial MX100 512GB|HD6970 2GB-Aquacomputer WC|MCP35X|Obsidian 800D|X-fi Titanium|AD900

User avatar
Nekosan
Posts: 1099
Joined: 09 Nov 2015, 19:39
Location: GoN Refugee Camp

Re: Star Citizen

Post by Nekosan » 21 Apr 2017, 18:31

bender wrote:
21 Apr 2017, 15:26
I think the only reason this architecture hasn't become the standard for mmos is due to the reasonably high complexity of state synchronisation and consequently high server CPU demand. In 2002 it was a pretty major issue, but in 2017, with an order of magnitude more CPU power its much more doable.

The reason it isn't done is because the majority of players are in the US and nobody wants to slightly shit up their game to cater to 20% of the playerbase. US players will cry about 20ms differences.
0 x
steamcommunity.com/id/krautpants

User avatar
stingtwo
Posts: 100
Joined: 05 Nov 2015, 17:47

Re: Star Citizen

Post by stingtwo » 21 Apr 2017, 19:42

bender wrote:
21 Apr 2017, 15:26
The cool thing with the design is there is actually no round trip between the servers. So not like this:

Code: Select all

           (request state)         (send state)
[AU server]----100ms--->[US server]----100ms--->[AU server]
The servers actually simultaneously broadcast (multicast) the change in their state each frame to each other:

Code: Select all

             (AU state)
           ----100ms--->
[AU server]             [US server]
           <---100ms----
             (US state)
So, when a server receives the state of another server, it rolls back its state to figure out what actually happened, reprocessing anything (hence only 140ms effective latency instead of 250ms). The only time this would need to happen, is if players from different regions needed to interact. This means that us here in AU could play on local servers, run around and interact with game objects (doors, cargo, NPC) and each other with minimal latency (40-80ms round trip). The only time latency would become an issue, is if a US/EU player attempted to interact (shoot, crash/ram, synchronise timing etc) with us or we with them. However, both players will experience the same latency and are therefore equally disadvantaged. So in a PvP scenario, the US player doesn't have the home server advantage.

This is assuming cross region play is a requirement for SC. Currently it seems the plan is to just have EU/AU players connect to US servers (or US to EU/AU etc) when wanting to play with friends/org members, and the experience would be significantly worse for them. Boarding an enemy ship at 250ms latency when a bunch of other players are at 40-80ms would just suck. Boarding an enemy ship at 140ms for only players interacting across different regions isn't great, but it would provide a better and fairer experience overall.
stingtwo wrote:
20 Apr 2017, 15:56
I admit to skimming the concept papers themselves but 2 things come into play, from a design standpoint they read like the servers were put on the same location, maybe even the same box to remove any delay having 2 servers synced up would suffer. It also was written in 2002, 2002 was hell for gaming clans looking for bf1942 servers in the US, so few had the network capacity to handle 64 players because up to that point, they only had to deal with 16 player arena shooters.

But if CIG wanted to handle more players in an instance, just fix up the broken netcode that uses 3 times more data than bf1 does.
I think the only reason this architecture hasn't become the standard for mmos is due to the reasonably high complexity of state synchronisation and consequently high server CPU demand. In 2002 it was a pretty major issue, but in 2017, with an order of magnitude more CPU power its much more doable.

The first paper only used local servers, but the second used a simulated server latency of 50 and 100ms for quake 3 servers. The the second paper provided examples where complex state synchronisation absolute worst case (4 server synchronisation) took 1000ms on 2002 tech. Now it would probably take 50-100ms or less for worst case (assuming current cpus are 10x faster than 2002 cpus, but its probably more like 30x). The average for 2 servers was 90ms, today that would be <10ms. There are also likely much better synchronisation algorithms than the ones used in the papers in 2002, reducing the latency even more. Multi-threading (one thread for each linked server) could also reduce the processing time by 2-4 times.
So how does it beat the ping between the US-AU EU-US or even to AU-EU, wormhole? THERE IS NO BEATING THIS PART, even when Internode offered a VPN to WOW players for better play, it lowered the ping JUST A FRACTION OF THAT 250MS they had to the servers.

Server has to get the info from both players and server(s) then sends what it has back to both players and to the other server(s), think about it, if all that happened was you recieved the data from other servers without sending any back, you just become ghosts in the world, no interaction because they don't know about your existence because nothing was sent by the server back.

Having them be multicast doesn't instantly remove the previous paragraph because now theres a central server, wheres that going to be to give every player an Equal ping to each other...Sierra Leone? Your description describes it as as the that has to calculate everything to confirm everything the servers sent it sends it back to everone to say "ok this happened", but if you went down the route of multicast anyway, why not remove all the middleman shit, at least you would know what your actual ping is.

Image just going off into a corner of BF1 to check how much the delay is for you, got a ping of 80 ping, so simple, between the gun going off and the bullet hitting the wall should be pretty much instantaneous, now it's half a second because the server in another country has to sync with other servers and send the info back to confirm the shit. You would go off your rocker going "WHY THE FUCKING HELL IS THIS A THING"

Game server architect hasn't advanced that much in the past 15 years(most has been is tickrates and bandwidth requirements), UT4 and any game using it is pretty much the same as setting up UT99, no game server code I'm aware of is multi core, there are reasons why this is still the case. but as I mentioned, this was network issues because people ran servers out of a garage than the hardware being maxed out. I run 3 KF2 servers on 10 year old hardware without a blip of spikes outside of HDD usage on booting them up after updates.
0 x

User avatar
bender
Posts: 105
Joined: 05 Nov 2015, 18:08

Re: Star Citizen

Post by bender » 22 Apr 2017, 14:58

stingtwo wrote:
21 Apr 2017, 19:42
So how does it beat the ping between the US-AU EU-US or even to AU-EU, wormhole? THERE IS NO BEATING THIS PART, even when Internode offered a VPN to WOW players for better play, it lowered the ping JUST A FRACTION OF THAT 250MS they had to the servers.
There is always going to be a fundamental latency between regions. As you put it "THERE IS NO BEATING THIS PART". There's nothing that can be done about that, except better prediction to hide it.
stingtwo wrote:
21 Apr 2017, 19:42
Server has to get the info from both players and server(s) then sends what it has back to both players and to the other server(s), think about it, if all that happened was you received the data from other servers without sending any back, you just become ghosts in the world, no interaction because they don't know about your existence because nothing was sent by the server back.
The servers don't wait for each other. Each server sends a multicast packet(s) to all other linked servers every tick. They synchronise their states, fixing any conflicting states due to latency (the CPU intensive part). When there are no conflicting states, there is no need to update the clients. They already have the correct information and can use client side prediction to smooth movement out.
stingtwo wrote:
21 Apr 2017, 19:42
Having them be multicast doesn't instantly remove the previous paragraph because now theres a central server, wheres that going to be to give every player an Equal ping to each other...Sierra Leone? Your description describes it as as the that has to calculate everything to confirm everything the servers sent it sends it back to everone to say "ok this happened", but if you went down the route of multicast anyway, why not remove all the middleman shit, at least you would know what your actual ping is.
There is no central/master server.
stingtwo wrote:
21 Apr 2017, 19:42
Image just going off into a corner of BF1 to check how much the delay is for you, got a ping of 80 ping, so simple, between the gun going off and the bullet hitting the wall should be pretty much instantaneous, now it's half a second because the server in another country has to sync with other servers and send the info back to confirm the shit. You would go off your rocker going "WHY THE FUCKING HELL IS THIS A THING"
You could have a bunch of US and AU players just walking around, but very little recalculation will need to be done most of the time due to motion prediction (as is the case with current game server tech). Most of the time, for all players it would seem like they are playing on a local server (as they in fact are). The only time things would start to get a bit laggy is when, say, both the AU and US player started shooting each other and moving around erratically.

Where this can get a bit confusing is, for example, if all the US players started shooting each other and the AU players started shooting each other, there would be no latency issues. There are no players from different regions interacting. All players would feel like they are playing on local servers (as they in fact are).

This would mean that if you're playing star citizen with your mates, you could interact and play with each other with minimal latency. Even if you join up with a US/EU player, most of the time latency wont be an issue.
0 x
Asus P8P67|2600K @4.5GHz-EK WC|16GB DDR3|Crucial MX100 512GB|HD6970 2GB-Aquacomputer WC|MCP35X|Obsidian 800D|X-fi Titanium|AD900

User avatar
boars
Posts: 393
Joined: 08 Nov 2015, 11:33
Location: Q. OO-su-to-ra-re-ah
Contact:

Re: Star Citizen

Post by boars » 22 Apr 2017, 15:24

Sad that 3.0 wont feature planets (only 2 or 3 really large moons) but this snippet out of the latest ATV show holds a little promise - canyon runs look fun. Definitely can see them setting up some race tracks through areas like that in the future.


An old leak I stumbled upon... not exactly what I had in mind when I thought "rail gun" but ok:
0 x

Post Reply