Star Citizen, Roberts Space Industries - Chris Roberts' life support and retirement fund [2012-]

[swaaye]

And of course Star Citizen is the ultimate feature creep nightmare. All game development precedents may be invalid comparisons.

 

[Silent_Buddha]

And of course Star Citizen is the ultimate feature creep nightmare. All game development precedents may be invalid comparisons.

Well, I think Duke Nukem 4-ever is still the poster child of feature creep taking over development. Star Citizen still has a ways to go to match that disaster. Diablo 3 also featured a lot of feature creep, but it turned out relatively well.

Regards,
SB

 

[swaaye]

Yeah those were pretty messy too. Though I would think the investment into Duke 4 was considerably less than what's going on here. There's something special about funding a disjointed mess via gamers who don't necessarily have a clue and who are running on peer influence / hopes & dreams.

 

[hoom]

The flight model is significantly better than that of the competition. While CIG's combat is absolute not n00b friendly its great for those who have enough skills.

To be fair it looks pretty good flying around in space.
But as soon as they get close to something (a spacestation, asteroid, the ground...) it just looks ass.

The combat vid you showed doesn't really look any different to the SC combat I've seen except for lower RoF.
Honestly not that much you can do with 1 v 1 in open space I guess & maybe it feels better when playing it but just doesn't look any good watching.

If you're making a MMO with a huge persistent universe for thousands of simultaneous players shouldn't the first priority be making server/netcode that can handle those kinds of numbers?
Maybe the netcode has improved so 50 works over the internet but it was recent, late 2017 presentations that showed laggy as heck with only 20 on a LAN.
But when do they expect it to work with thousands?
From recollection EVE runs each System on separate servers with ability to dynamically scale processing power up/down depending on player numbers & (infamously?) time scaling to keep up with processing load, I'm having a hard time with the idea of separate 50 player servers.

 

[Davros]

On the bright side in early 2016 Jupiter told us Squadron 42 would be released soon

 

[hoom]

™

 

[Shortbread]

Still waiting for my Constellation Andromeda loaner to be updated fully to my LTI Genesis Starliner. :/

 

[Jupiter]

Still waiting for my Constellation Andromeda loaner to be updated fully to my LTI Genesis Starliner. :/

This will probably still take some time. In the last few weeks CIG showed their production of other ship manufacturers (Anvil, Aegis, MISC, Drake, Origin, RSI, Consolidated Outland, Alien ships) and Genisis is a ship manufacturer which has only the Starliner thats why it does not have the largest priority. They establish a spaceship style on a large spaceship and then they can build spaceships of this manufacturer much faster.

The Orion is now 100% larger and comparable to capital ships. The price was $300 and whoever owns one could soon have a $500 to $700 ship. This will not change the operating costs a lot because the Orion was previously regarded as an end ship for mining.

To be fair it looks pretty good flying around in space.
But as soon as they get close to something (a spacestation, asteroid, the ground...) it just looks ass.

The combat vid you showed doesn't really look any different to the SC combat I've seen except for lower RoF.
Honestly not that much you can do with 1 v 1 in open space I guess & maybe it feels better when playing it but just doesn't look any good watching.

If you're making a MMO with a huge persistent universe for thousands of simultaneous players shouldn't the first priority be making server/netcode that can handle those kinds of numbers?
Maybe the netcode has improved so 50 works over the internet but it was recent, late 2017 presentations that showed laggy as heck with only 20 on a LAN.
But when do they expect it to work with thousands?
From recollection EVE runs each System on separate servers with ability to dynamically scale processing power up/down depending on player numbers & (infamously?) time scaling to keep up with processing load, I'm having a hard time with the idea of separate 50 player servers.

Most of them will look similar in Star Citizen but fighting with very good pilots will look different. They try to keep the smallest possible distance.

Sure, but they also had to create assets, etc. because at least then there is something to show and they also need a lot for Squadron 42. The servers wont be as important there. The network stuff is still causing a lot of problems, but the performance is now better with 50 players than before at 24.

 

[Jupiter]

Finally I know why the performance is not good. Answer from a CIG network programmer.

Why is the graphics pipeline waiting for server updates?
"There are a few myths that seem to get repeated quite a lot - please allow me to dispel some of them for you:
* The graphics pipeline does not wait for server updates.
* Server FPS does not affect client FPS.
* Netcode does not make clients run slowly, and never has.
* Netcode does not make servers run slowly, anymore, even though we've added more clients.
* You get better performance on newer servers because there are fewer players on them so your client has to do less work - like physics, animation, IFCS, and entity updates.
* Players hacking the game to play PU in "offline mode" get better performance than they do online because their clients don't have to deal with all the load generated by 49 other players.

What is so hard about fixing the performance problems is that the game is pushing the engine way beyond what it was designed to handle. Fixing that means fundamentally changing how systems work while simultaneously trying not to break everything in the game that uses them. Big performance gains that require making big changes take time. Sometimes we have to do a lot of restructuring before we can even start working on an optimisation. Making all these changes can introduce a lot of bugs, and fixing those takes even more time. Let's also not forget that performance is not the only goal here - we're also trying to achieve fidelity levels not seen before. Fidelity is often the enemy of performance, so we find ourselves having to optimize even further than we otherwise would have had to."

https://robertsspaceindustries.com/...raphics-pipeline-waiting-for-server-up/773513


Netcode:
"Thanks for the interest, @broco2002. As far as I'm aware there currently aren't any plans to do a netcode special on AtV or any of the other shows. Most of what we do on the network team is hidden from view, so it would be cool to be able to give the community an update and go into detail about what's been done, and where we go from here. If the community team feel there is enough interest in something like this then I'm sure it will happen. Believe me, I can talk about the work we've been doing for hours on end - you might be sorry that you asked  

One thing I'd like to point out real quick though is that the graphics stack is never stalled waiting for server responses. That would be a truly awful way to network a game engine! I have seen similar remarks on social media, but I can assure you that this is not how the engine works, and it never has. What is true is that the server does indeed validate some client actions, but your client will carry out those actions locally while it is waiting to hear back from the server. Only if the server disagrees (possibly due to latency or cheating) will your client have to undo those actions and correct itself based on the authoritative version of events from the server. This is an industry-standard technique known as "client-side prediction" (because your client predicts that the server will validate its actions) and is used to make network gameplay more responsive - it does not stall the graphics stack or otherwise affect frame rate."

https://robertsspaceindustries.com/spectrum/community/SC/forum/50259/thread/net-code-special


Fantrailers of 3.0.

(strange music in the frst part of this video but the camera work is excellent)

I like the use of camera tools for such gameplay videos much more than just for screenshots.

 

[hoom]

What is so hard about fixing the performance problems is that the game is pushing the engine way beyond what it was designed to handle. Fixing that means fundamentally changing how systems work while simultaneously trying not to break everything in the game that uses them. Big performance gains that require making big changes take time. Sometimes we have to do a lot of restructuring before we can even start working on an optimisation. Making all these changes can introduce a lot of bugs, and fixing those takes even more time. Let's also not forget that performance is not the only goal here - we're also trying to achieve fidelity levels not seen before. Fidelity is often the enemy of performance, so we find ourselves having to optimize even further than we otherwise would have had to."

Pretty much confirms my suspicion: All these fancy systems haven't been designed from the start to make sure they will scale to proper MMO player numbers.
Then they're finding that A: they don't & B: its a huge PITA to try to retro-optimise them.

Meanwhile they talk up multi-player ships including ones that need twice as many crew as the server can handle...

Edit: happened on this vid which articulates some of my concerns about gameplay probably being un-fun

And thats mostly just about the basic getting around doing everything manually.
Other concerns are the actual economy with everything wearing out & needing periodic repairs, one of the 3.0 vids I watched showed what looked awfully like

 

[Grall]

Other concerns are the actual economy with everything wearing out & needing periodic repairs

You know, I've only repaired my gear about a million billion billion times in various games. Apart from everything wearing out in real life, it's not even a foreign concept in science fiction as such. "Beam me up, Scotty!", remember? (Words which were never actually uttered on the show... )

Btw, full disclosure time: it is my one big hope I will be able to be a starship engineer in Star Citizen one day and repair a starship, even during space combat.

So yeah, bring on the repair stuff. As long as the players are the ones who can do it. It was almost fun having to carry half a ton of repair hammers back in Morrowind, had they not been so damn heavy. And had the equipment wear system not been so ridiculous too. I mean, the pretty damn awesome mace you got from beating the main questline wore like crazy, and you had to keep repairing it constantly or it would basically do NO DAMAGE to enemies... Crazy!

 

[Jupiter]

Pretty much confirms my suspicion: All these fancy systems haven't been designed from the start to make sure they will scale to proper MMO player numbers.
Then they're finding that A: they don't & B: its a huge PITA to try to retro-optimise them.

Meanwhile they talk up multi-player ships including ones that need twice as many crew as the server can handle...

Edit: happened on this vid which articulates some of my concerns about gameplay probably being un-fun

And thats mostly just about the basic getting around doing everything manually.
Other concerns are the actual economy with everything wearing out & needing periodic repairs, one of the 3.0 vids I watched showed what looked awfully like

I think its the Gamelogic Culling. A system that suppresses the calculation of Gamelogic objects that are of no importance to the local player. For Star Citizen an example would be another spaceship that the player can not be see. It does not change my gaming experience if my local machine knows exactly which ship engine is firing at what power and so on. But classic first person shooters systems are often programmed so that they only work if they constantly calculate the entire Gamelogic as this is much easier which does not work for an MMO.

LevelCap is right when the player should be able to spawn next to a friend if he is in the same star system but the rest sounds too arcady to me.
A big German magazine exclusively played the complete version (but only a few hours of it) of Kingdom Come and they praised that it is much harder and more time-consuming than other AAA games. In this game a wrong behavior in a dialoge has consequences which can be mess up a three-hour long quest (this game does not have quick save etc.). Nothing against fast, arcady and easy AAA games but there are already many of them. Demanding titles like Kingdom Come do not exist that often.

 

[tuna]

Pretty much confirms my suspicion: All these fancy systems haven't been designed from the start to make sure they will scale to proper MMO player numbers.
Then they're finding that A: they don't & B: its a huge PITA to try to retro-optimise them.

"Suspicion"? Haven't this been absolutely clear since day one of development (with CryEngine)?

 

[Jupiter]

CIG wants to do something for which their engine was not intended. It is not impossible but is usually expensive in terms of both the budget and the schedule. Which one has already seen with the Frostbite engine that should suddenly be EAs miracle weapon and thus killed some studios.

But there are already multiple CRYENGINE MMOs so it's not impossible. With a new engine Squadron 42 would take even longer to build and people are already complaining because of shifts.

 

[Deleted member 13524]

In which case it's only starting to get as long as World of Warcraft (4-5 years original development time), Final Fantasy XIV (5 years for original release, PS3 version delayed indefinitely at that time due to it being so bad, and then finally releasing in a playable state on PC and PS3 in 2012, so 7 years for the real version), and many other MMORPGs.

If you want something really long. Diablo 3 originally started development in 2001 and didn't release until 2012...11 years. Star Citizen has a ways to go before they match that.

When talking about the development time of a game, it's good to keep in mind the scope of the game, and how fleshed out they want to make a large world. Comparing Star Citizen to single player games of much smaller scope doesn't make sense, but even in that case, as Diablo 3 shows, you can still run into cases of exceptionally long development times.

Regards,
SB

Who's complaining about MMOs development times?
Not that I think any of the aforementioned comparisons are valid, though.
Duke Nukem Forever was actually shelved for quite some time. No one knows what exactly started in 2001 on Diablo 3, and the game was only announced in 2008.



Problem here is the first backers - the ones who actually allowed the project to kickstart - gave their money for a Wing Commander spiritual sequel, which even after some moderate delays should have come out between late 2015 and early 2016.

And then Chris Roberts shat on all these backers and started feeding off the whales to make a MMO he most probably won't ever be able to launch.

So everyone understands how much time the feature-creeping is taking? I'd say most people do.
The problem here isn't understanding that feature X or Y takes time. It's the ever-changing scope of the game due to constant feature-creeping which has obviously gone completely out of control.
Next thing is they're delaying the game another year so they can implement lootboxes or some shit.


Had I known the game was going to be a space MMO, I'd never given my money to them.
When I get the time, this week I'm probably going to ask for my refund.

 

[swaaye]

Yeah I was trying to remember the original kickstarter stuff. I don't recall it being anything like a MMO. Squadron 42 was at the front and was pushed as a space drama Wing Commander style game. That's what I was interested in and decided I would back.

 

[Malo]

I'm pretty sure from the start it was sold as a persistent universe and the Squadron 42 separate game.

 

[swaaye]

I'm pretty sure from the start it was sold as a persistent universe and the Squadron 42 separate game.

That could be. I don't really follow terminology attached to MMO games so I may not have realized what was being hinted at.

I probably thought of the X series of games and how their universe develops on its own.

 

[Jupiter]

Lets see how this ends.

https://de.scribd.com/document/367101474/Crytek-v-CIG


CIG wanted to show something from Squadron 42 on the 21st of December. Hopefully that will not be suspended because of this. ^^

 

[swaaye]

Interesting. Quite a few breaches claimed.

14.

At significant time and expense, Crytek created demonstrations and proofs-of-concept for Defendants related to Star Citizen, and Defendants used those materials as part of the crowdfunding campaign for Star Citizen.

I guess that explains why the prototype stuff looked awesome.