nurmie posted:

A planetwide derelict corporation-built pyramid scheme-funded ghost city is an incredibly cool idea for a setting, actually

HAmbONE posted:

Merry Christmas to all citizens of stars and to all a [connection to server lost]

Bumble He posted:

enjoy the holidays everyone, with the cloud imperium christmas album, only 49,99 + 23 for shipping!

"All I Want for Christmas PU"
"Glitchin’ Around the Christmas Tree"
"Last Bugfest"
“Sata Ball is coming to town“
"Jingle Bell Roc"
"White Connie"
"Have Yourself a Merry Little Idris"
“Do they know its pledge time?“
"Hobo the Red-Nosed Reindeer"

Beet Wagon posted:

There's also significant obfuscation in here, I think. For example, CR talks about this being the best year ever in terms of "new player growth" but doesn't specify what that means. If we look at "new web accounts made" this certainly was a good year - better than the last few by a significant amount - but not "the best" as you can see here:



So what does 'new player growth' mean, exactly? We know they added ~437,891 new web accounts, and at a 40% capture rate that suggests they added ~176,382.4948 new paying customers over the year, but again that's just an estimate because they didn't actually tell us how many new people paid them money over the course of the year. That number doesn't seem like "best year ever" material, so it must be something else, but gently caress knows what criteria they used to get there.

Anticheese posted:

A friend just raised this theory: what if sq42 is being remade not because it's Duke Nukem Forever, but because Roberts doesn't want his ex wife in it?

Anticheese posted:

A friend just raised this theory: what if sq42 is being remade not because it's Duke Nukem Forever, but because Roberts doesn't want his ex wife in it?

Pixelate posted:

It’s just occurred to me that Crytek were spot on about SQ42 not being 'ripe'

And that Chris has weirdly never boasted once about their triumphant validation in the court case. You’d think he’d be up for a bit of that

Taintrunner posted:

Sandi’s gone? Wha?

Sarsapariller posted:

Things are getting downright stimperical

https://v.redd.it/9271frmm8j761/DASH_720.mp4

JugbandDude posted:

Maybe they settled and cannot comment on the terms.

ComfyPants posted:

Sarsapariller posted:

Things are getting downright stimperical

https://v.redd.it/9271frmm8j761/DASH_720.mp4

Boywhiz88 posted:

I believe she stopped showing up in videos, and then on a rare CRoberts appearance, his wedding band was missing.

Sarsapariller posted:

Things are getting downright stimperical

https://v.redd.it/9271frmm8j761/DASH_720.mp4

quote:

The roadmap to me has no inherent hype. I think some people see it as this thing that exists on its own. It's just a fancy way of looking at the work of hundreds of developers. That work exists and progresses whether the roadmap exists or not.

I log in with each patch and I'm LITERALLY BLOWN AWAY by the progress. I am not alone. Reddit is littered with the same - the distribution of posts of people seriously excited about what's being delivered grows in proportion to the alternate feelings with every single patch. This is corroborated by the player participation data and the pledge-o-meter performance.

I don't need hype when I can hold quality output in my hands. So no hype - they are on a great trajectory, delivering solid progress, and the roadmap is an unprecedented view into how it all gets done. Hype not required!

The proof in the pudding is in the eating, as the phrase goes. And the eating is currently very good!

Merry Christmas!

quote:

Dude, you need to realize a few things:

1. The amount of hate is actually relatively tiny; the term "vocal minority" exists for a reason, and applies here, 1000%.
   
2. If you want to see the counter balance to the angsty, vitriolic hatred on the forums and here, simply take a gander at that skyrocketing, astronomical, World Record Setting "Pledge O Meter" and remember that people generally don't spend money on things they don't enjoy, and you'll understand that for every 1 hater, there are literally 10's of thousands of supporters.

Then take a deep breath, let it out with a smile and understand you are on the right side of history, as is being borne out in data points that cannot be angrily argued away.

The "negative narrative" has changed little since it began in 2013, but the game has progressed enough that it just keeps looking sillier and sillier.

The inflection point where this game proved it's on the right path and is awesome and continuing to be more awesome with every quarterly release occurred back in the early 3.x release cycle, when persistence started to gel and you can see that reflected in the funding gains and most recently in the amazingly elucidating player engagement metrics that Chris shared in his end of year letter.

Haters gonna hate. That they claim to be done with the game, but like a cringey stalker spend all their time here hating on it, is the real puzzler!

Bottom line: let not your heart be troubled. All is well in SC!

Sarsapariller posted:

Things are getting downright stimperical

https://v.redd.it/9271frmm8j761/DASH_720.mp4

Sarsapariller posted:

6) They have like 5 engineers dedicated to bug fixing out of the whole pool, lol

Taintrunner posted:

Sandi’s gone? Wha?

Taintrunner posted:

Jesus Christ that’s bleak, lmao.

quote:

Ok so this is really long but you asked lots of questions that require more than tweet length for satisfying answers. Grab some water.

quote:

There have been two separate engines for Star Citizen. Neither of them are good choices.

In 2011, there were no good choices available for what Chris Roberts wanted to do. One way or another he'd end up having to dramatically rewrite large chunks of any engine they adopted. UE4 was not ready for prime time until 2014 and UE3 was long in the tooth.

quote:

Why did they choose Lumberyard over Unreal 5, which would actually allow them to do the zero g they wanted?

Because Lumberyard is CryEngine so they did less of an engine swap and more of a license swap when they made the switch? The Crytek vs CIG court case revealed that CIG had licensed the old versions of CryEngine that Amazon bought from Crytek, which meant they could take one of the earliest post-switchover commits and have what was effectively vanilla CryEngine 3.8 but under license from Amazon instead of from Crytek. Prior to adopting Lumberyard, Star Citizen had standardized on 3.7(.2 I think?) so it was a minor effort to make a small version hop.

In the four or so years since, they've integrated mainstream Lumberyard features and code, and a random Redditor once anecdotally claimed to me that they're now running vanilla Lumberyard instead of Amazon-CryEngine as the foundation for their modifications so pinch of salt with that.

If they'd adopted Unreal Engine, on the other hand, they'd have had to throw out everything and start from scratch on a new engine unlike the CE->Ly switch. Also, UE5 isn't even hitting 1.0 until 2021 so if the zero-g features you're referring to are specific to that 5 that's far too long to wait for an engine to do their work for them: SC had working (if buggy) zero-G EVA in 2014 and meaningful gameplay associated with space EVA with the release of alpha 2.0 in 2015.

quote:

Why did their first person and third person studios not communicate?

The only major problem between CIG and one of its outsourced partners that I know of was Illfonic, and I have no idea what led them to leave their brains at home for sheer months with managing that outsourcing relationship. That was a genuine debacle.

quote:

Why did they not stage releases like Elite?

There have been playable alpha builds with new features since early/mid 2014. They've also been taking the time to develop the engine to handle a lot of things fundamental to gameplay. It'd be the height of foolishness to polish gameplay to release quality and go big on the marketing budget for a launch when that gameplay is lacking core critical features like the ability to walk through your ship a unoccupied manned turret while in flight/combat and not clip through the floor and get left behind in space.

Frontier chose, if choose is necessarily the right word for their financial outlook at the time, to release the absolute basic shell of gameplay at launch and add on the go. They didn't even allow you to form Wings until adding it in the first post-release patch. Landing on airless planets came a year later, and five years later Elite's now adding the ability to get out on foot on planets with no or "tenuous" atmospheres and shoot at each other or scan alien plants.

For people who find Elite as it existed at launch and as it exists now fun and engaging, this method of rolling out the basic game and building onto it with updates as they come makes a lot of sense and is great because you get to get started on permanent progression and ride the new things as they come around the corner. I disagree, because I've had Elite since 2015 and to me the experience has consistently felt like a series of disconnected silos of gameplay that have historically felt like placeholders that never got replaced - a symptom of tacking things onto the side one at a time without considering how they could and logically would interact with each other or at least either not taking or having the time to explore those considerations.

It probably won't surprise you to know that I personally don't value the benefits of a "released" game - permanent earning, killer sound design, reasonable/good levels of optimization and bug occurrence rates - if I'm not having fun actually playing the game. And I wish I did have fun on a consistent basis, I wish I didn't have all these things to say so I could just be with you going "yeah, why didn't they? Worked out flawlessly for Frontier, get on it boys!" It'd be a much shorter post to write and read.

quote:

Why are they working on a single player campaign NOW?

Because the project was, from the very beginning starting with the contract signed between CIG and Crytek (also an exhibit in the court case), about both the Persistent Universe, aka Star Citizen, and the single-player Squadron 42 campaign. Both are using the same engine and shared assets so there's a lot of overlap and synergy, and SC is the modernized analogue of Freelancer while SQ42 is Wing Commander but EA probably wouldn't let Chris use the IP.

The original development plan predicted that they'd bring in about $6 million tops from the crowdfunding campaign, they'd go to private investors to top the tank up to $20-25 million, and they'd spend that making SQ42 with an engine scoped to be similar to early Elite - you'd be nailed to your pilot seat except for cutscene transitions in stations for limited on-foot mission NPC interactions. The plan then was, paying customers willing, that they'd take the profits from SQ42 and develop the engine further and build an online component, and they'd keep plugging away at building more ambition into the game as long as profits kept coming.

As everyone knows, the money tap never shut off so they could develop them both in parallel and use the PU side of the project for public playtesting without exposing SQ42 spoilers, preserving the bulk of the story for (eventual) release.

quote:

But I keep feeling a lot of the difficulties are a culmination of not thinking the problem through and breaking it down into proper chunks.

I can't speak for every single aspect of every part of development because that's a very broad range, but there are cases where they have clearly thought the problem through and broke it into proper chunks.

Consider, for example, the most important software engineering task they're working on right now other than adapting everything to use Vulkan as the rendering API: static server meshing. It's well-known that SC's servers perform badly in practice and it's because of very heavy load with 50 players and a star system full of stuff; it's the EVE Online problem, but it is completely unacceptable to engage Time Dilation in a first-person game while EVE can get away with it in a third-person strategy-action-ish game.

The groundwork for static server meshing began around 2015/2016 with the development of the Megamap, a specially-formatted CryEngine level that acts like the holodecks from Star Trek. It's an empty container that loads whatever it needs and streams things in and out on demand. The first thing the Megamap loads, as far as players are concerned, is the main menu scene. Because the Megamap never actually changes itself to another level, it just changes what content is playing inside of it, gameplay is seamless without loading screens once connected to a game server.

With the Megamap in place, CIG next developed Object Containers and local physics grids. In simple terms, an object container is a local frame of reference for whatever's inside of it. A space ship is an object container and if you were in EVA outside my ship while I walked around in the cargo hold with the door open and my friend sat in the pilot seat and rolled the ship around, everything would work like you'd expect instead of the physics god getting angry and eating everyone. Object containers can be nested and be generally any size. The current alpha test system's planets are object containers, with child object containers around each moon inside each planet's container, with everything being inside the star system's top-level container.

After object containers were working, CIG was able to move onto client-side and then server-side Object Container Streaming. Prior to OCS, both the server and all clients had to simulate everything in the entire star system, a good 70,000-100,000+ entities. This kills performance. OCS allows the client to completely ignore everything too far away for it to care (anything outside the largest nested container of significant size such as a moon) and for the server to shut down simulation processing for areas that do not have any players in them so there don't need to be AI actively walking around for no one.

With OCS working, CIG got started on server meshing and multiple aspects of persistence. Server meshing should allow different server instances to simulate individual object containers, breaking the heavy load apart across multiple servers working in a swarm. When a particular object container has no players in it, the entire instance running it can be shut down - after saving the data in the server for persistence so when someone flies back into it later everything is where it should be. Neither of these are implemented yet so we'll see if they can get one or both in-game in the next year.

From the Megamap to server meshing, each of the engine tech developments I listed is a dependency for the thing that comes after it. I left out little details and minor tech enhancements, but here's an example of CIG thinking things through for a continuous chain of engine enhancements over a five-year span.

There's also no good point in this chain of dependency-building leading up to server meshing that was particularly suitable for pushing through to release-quality polish and optimization to pull Elite-style staged releases, because engine fundamentals were missing.

Sarsapariller posted:

Things are getting downright stimperical

https://v.redd.it/9271frmm8j761/DASH_720.mp4

Mirificus posted:

In 2011, there were no good choices available for what Chris Roberts wanted to do. One way or another he'd end up having to dramatically rewrite large chunks of any engine they adopted. UE4 was not ready for prime time until 2014 and UE3 was long in the tooth.


Because Lumberyard is CryEngine so they did less of an engine swap and more of a license swap when they made the switch? The Crytek vs CIG court case revealed that CIG had licensed the old versions of CryEngine that Amazon bought from Crytek, which meant they could take one of the earliest post-switchover commits and have what was effectively vanilla CryEngine 3.8 but under license from Amazon instead of from Crytek. Prior to adopting Lumberyard, Star Citizen had standardized on 3.7(.2 I think?) so it was a minor effort to make a small version hop.

In the four or so years since, they've integrated mainstream Lumberyard features and code, and a random Redditor once anecdotally claimed to me that they're now running vanilla Lumberyard instead of Amazon-CryEngine as the foundation for their modifications so pinch of salt with that.

If they'd adopted Unreal Engine, on the other hand, they'd have had to throw out everything and start from scratch on a new engine unlike the CE->Ly switch. Also, UE5 isn't even hitting 1.0 until 2021 so if the zero-g features you're referring to are specific to that 5 that's far too long to wait for an engine to do their work for them: SC had working (if buggy) zero-G EVA in 2014 and meaningful gameplay associated with space EVA with the release of alpha 2.0 in 2015.


The only major problem between CIG and one of its outsourced partners that I know of was Illfonic, and I have no idea what led them to leave their brains at home for sheer months with managing that outsourcing relationship. That was a genuine debacle.


There have been playable alpha builds with new features since early/mid 2014. They've also been taking the time to develop the engine to handle a lot of things fundamental to gameplay. It'd be the height of foolishness to polish gameplay to release quality and go big on the marketing budget for a launch when that gameplay is lacking core critical features like the ability to walk through your ship a unoccupied manned turret while in flight/combat and not clip through the floor and get left behind in space.

Frontier chose, if choose is necessarily the right word for their financial outlook at the time, to release the absolute basic shell of gameplay at launch and add on the go. They didn't even allow you to form Wings until adding it in the first post-release patch. Landing on airless planets came a year later, and five years later Elite's now adding the ability to get out on foot on planets with no or "tenuous" atmospheres and shoot at each other or scan alien plants.

For people who find Elite as it existed at launch and as it exists now fun and engaging, this method of rolling out the basic game and building onto it with updates as they come makes a lot of sense and is great because you get to get started on permanent progression and ride the new things as they come around the corner. I disagree, because I've had Elite since 2015 and to me the experience has consistently felt like a series of disconnected silos of gameplay that have historically felt like placeholders that never got replaced - a symptom of tacking things onto the side one at a time without considering how they could and logically would interact with each other or at least either not taking or having the time to explore those considerations.

It probably won't surprise you to know that I personally don't value the benefits of a "released" game - permanent earning, killer sound design, reasonable/good levels of optimization and bug occurrence rates - if I'm not having fun actually playing the game. And I wish I did have fun on a consistent basis, I wish I didn't have all these things to say so I could just be with you going "yeah, why didn't they? Worked out flawlessly for Frontier, get on it boys!" It'd be a much shorter post to write and read.


Because the project was, from the very beginning starting with the contract signed between CIG and Crytek (also an exhibit in the court case), about both the Persistent Universe, aka Star Citizen, and the single-player Squadron 42 campaign. Both are using the same engine and shared assets so there's a lot of overlap and synergy, and SC is the modernized analogue of Freelancer while SQ42 is Wing Commander but EA probably wouldn't let Chris use the IP.

The original development plan predicted that they'd bring in about $6 million tops from the crowdfunding campaign, they'd go to private investors to top the tank up to $20-25 million, and they'd spend that making SQ42 with an engine scoped to be similar to early Elite - you'd be nailed to your pilot seat except for cutscene transitions in stations for limited on-foot mission NPC interactions. The plan then was, paying customers willing, that they'd take the profits from SQ42 and develop the engine further and build an online component, and they'd keep plugging away at building more ambition into the game as long as profits kept coming.

As everyone knows, the money tap never shut off so they could develop them both in parallel and use the PU side of the project for public playtesting without exposing SQ42 spoilers, preserving the bulk of the story for (eventual) release.


I can't speak for every single aspect of every part of development because that's a very broad range, but there are cases where they have clearly thought the problem through and broke it into proper chunks.

Consider, for example, the most important software engineering task they're working on right now other than adapting everything to use Vulkan as the rendering API: static server meshing. It's well-known that SC's servers perform badly in practice and it's because of very heavy load with 50 players and a star system full of stuff; it's the EVE Online problem, but it is completely unacceptable to engage Time Dilation in a first-person game while EVE can get away with it in a third-person strategy-action-ish game.

The groundwork for static server meshing began around 2015/2016 with the development of the Megamap, a specially-formatted CryEngine level that acts like the holodecks from Star Trek. It's an empty container that loads whatever it needs and streams things in and out on demand. The first thing the Megamap loads, as far as players are concerned, is the main menu scene. Because the Megamap never actually changes itself to another level, it just changes what content is playing inside of it, gameplay is seamless without loading screens once connected to a game server.

With the Megamap in place, CIG next developed Object Containers and local physics grids. In simple terms, an object container is a local frame of reference for whatever's inside of it. A space ship is an object container and if you were in EVA outside my ship while I walked around in the cargo hold with the door open and my friend sat in the pilot seat and rolled the ship around, everything would work like you'd expect instead of the physics god getting angry and eating everyone. Object containers can be nested and be generally any size. The current alpha test system's planets are object containers, with child object containers around each moon inside each planet's container, with everything being inside the star system's top-level container.

After object containers were working, CIG was able to move onto client-side and then server-side Object Container Streaming. Prior to OCS, both the server and all clients had to simulate everything in the entire star system, a good 70,000-100,000+ entities. This kills performance. OCS allows the client to completely ignore everything too far away for it to care (anything outside the largest nested container of significant size such as a moon) and for the server to shut down simulation processing for areas that do not have any players in them so there don't need to be AI actively walking around for no one.

With OCS working, CIG got started on server meshing and multiple aspects of persistence. Server meshing should allow different server instances to simulate individual object containers, breaking the heavy load apart across multiple servers working in a swarm. When a particular object container has no players in it, the entire instance running it can be shut down - after saving the data in the server for persistence so when someone flies back into it later everything is where it should be. Neither of these are implemented yet so we'll see if they can get one or both in-game in the next year.

From the Megamap to server meshing, each of the engine tech developments I listed is a dependency for the thing that comes after it. I left out little details and minor tech enhancements, but here's an example of CIG thinking things through for a continuous chain of engine enhancements over a five-year span.

There's also no good point in this chain of dependency-building leading up to server meshing that was particularly suitable for pushing through to release-quality polish and optimization to pull Elite-style staged releases, because engine fundamentals were missing.

colonelwest posted:

Uhm excuse me, I own an Idris so I am a MAN not a boy. And, according to the roadmap you’ll soon be moping it.

Pixelate posted:

It’s just occurred to me that Crytek were spot on about SQ42 not being 'ripe'

And that Chris has weirdly never boasted once about their triumphant validation in the court case. You’d think he’d be up for a bit of that

cmdrk posted:

i wonder how many children each woman would need to have on average from now until 2950 or whatever, to fill a city planet. ArcCorp is basically one of those ghost cities in China.

Sarsapariller posted:

Things are getting downright stimperical

https://v.redd.it/9271frmm8j761/DASH_720.mp4

Boywhiz88 posted:

I believe she stopped showing up in videos, and then on a rare CRoberts appearance, his wedding band was missing.

I still can't believe SC is still "living." I wonder the wheels finally come off, what will the be the catalyst? Just too long in its current state? It somehow getting buggier? Less buggy so that it's boring parts becoming overwhelming?

Thoatse posted:

Please go stand by the escalator
https://i.imgur.com/gbvXMnJ.mp4

The Titanic posted:

Is this a bug or did they make a "shot in the dick and I die" animation?

kilus aof posted:

Nut shots are actually pretty common in games. Johnny Cage's low blow in Mortal Kombat, nut shots in Soldier of Fortune, groin shots in Fallout 1 and 2, nut shots(and an achievement for it) in Saints Row 2 and every WWE game has low blows.