Sean Tracy

Hi @SeriousFun, great to hear from you and glad you watched! Yeah, I would have loved to have a few more hours to talk deeper on some of the tool chains and other aspects of my work since we covered a very small percentage. I don't usually get on the forums here, but as you were so kind to me in person, happy to share what I can.

@SpaceTomato I watched the section of the video linked by @SeriousFun and happy to give some info.

RASTAR:
Rastar indeed was the stepping stone to what has become Starchitect.
Rastar suffered a bit of a problem on the onset, which was an small identity crisis!
It wasn't created by a tools team to start, although it started as a development tool. It quickly shifted to potentially being player facing, for things in the future.
Development-facing tools have totally different intentions than a player-facing tool would, imo.
Although this is just my opinion, that was part of why it slowed down. Some lessons learned from it helped though, and much of the tech has made its way into Starchitect.
So without diving too deep into it, Rastar has been used to get us as far as we are and is used to place some locations, but most if not all of the development-facing side has moved to Starchitect, which incidentally has been made explicitly with the updated Genesis model in mind and is more fit for purpose.

Houdini/MightyBridge:
Still works! Great work from the MTL Tools group, honestly. It's quite sophisticated and still used in some regards to import Mighty Bridge Tools and live-connect the Editor to Houdini.
Some tools/cases include:

Cover Tools: A tool to place cover objects around a selection of mesh in the Editor.
Ledge Tools: A tool to automatically place ledges at the correct spots on each object you have selected.
Cable Generator: Uses Houdini Vellum simulation to generate cables, using splines to create the base and objects in scene or planet surface as collision.
Tarp Generator: Uses simulation to generate tarps, using splines to create the base cloth and objects in scene or planet surface as collision.
Pile Generator: Allows you to create simulated piles of objects that are affected by gravity and ground collisions from any selection in the Editor. It can also scatter and duplicate them along a spline or in a specific direction.
SplineInstancerTool: A tool to place different objects on a spline with different modifiers.

I love telling you guys about how we are putting this stuff together and its changing more than you'd think it does based on technology, expectations and sometimes crazy ideas so totally understand the questions and hope that sheds a bit of light on the areas you wanted more information on.

-Sean "Criminal" Tracy

All right. So it's a great question. And all of the topics that they brought up there have been brought up by Chris. So the short answer is, yeah. When? Wow. There's a lot of pieces to that. So one of the things that is coming sooner than it is later is scars. And the feature and the technology of it is blood, sweat and tears is what we call it, right? So the characters can bleed, so they can sweat, so that they can cry. And what that does bring you is some decal support for the characters. And that's what's actually really critical to have scars and things like that. So the tech enables this, but that starts getting into design discussions like, okay, when does a scar get made, right? Does a scar last forever? How long does it last? How does Death of a Spaceman work into scars, blood, sweat, tears, these kind of things? So I think there's a lot of great conversations to come for that, but it will be completely enabled. So we're going to be able to put a character in, just put the scar on them, looks great. But as far as that being a dynamic thing that happens when a player is playing, again, we've got to have design discussions to see exactly how that would work out, but the tech would absolutely allow it. So that's just on scars.

The body question is a bit of a bigger question. It's already been a mandate that Chris has put forward that he wants to be able to have custom bodies. Now, I'm not going to talk about the height adjustment because the height adjustment is a different problem than the, I call it thin-fat variations. So thin-fat variations have been around for ages and ages. And the old way of doing it is you literally have a thin one, you have a fat one and you kind of blend in between all of them, right? As you guys know, we have a really, really sophisticated system in DNA within the face. What I'd like to develop and what we are beginning to develop now is DNA for the body. So DNA for the body would give us basically all the same abilities that we have within the face, giving a massive amount of ranges of body types that you could blend in between depending on different controllers or whatever you'd want to do. So that's where Chris wants to take this. Now, there's so many challenges, guys, because it's not just about doing the body. I think a lot of people think about the customization, they see a guy standing there in an outfit, and they're like, cool, that should just work. I agree, but that amalgamation of a character is like 12 assets, all independent, right? If I were to restart this game again, I would make every body the exact same topology, every shirt on top of it the exact same topology so that we could do those blending kind of things. But that's why we've got to utilize something like DNA and a wrap deformer on top of it to be able to do that. So you guys are aware, when we mark up a hair or we mark up sunglasses, toques, or anything around the face, it takes some time. It's a lot like skinning. We have to mark up the vertices to say how much or how little do you move, how do you move, whether it's elastic, whether it's rigid. It takes some time. We would have to do that then with every single body asset. And just to give you some idea, we are nearing 1,200 body type assets within the PU. It might even be higher. That's the last I checked. 1,200 assets. So we're talking a lot of stuff over many, many years. And they're all in different states of, where's the mesh? LODs, face orders, it's tricky. So the things that go on the face to make DNA work, we know the vert order of every single one of those faces, and we know the vert order of the attachment so we can move it. But if we have all these assets on top of a body, we don't know that. So basically, we have to make a proto body, like we have the proto head, which is the mass average. Then we have to mark up every single asset that would go on top of it to move appropriately within all those ranges. So it's just a lot of work. So I think the first step, but it's a question a little bit later, first step is gonna be something like StarWear, right? And then we need to be able to look at the body adjustment.

The other place that we are able to use this is on new types of bodies. And what do I mean by new type of body? Well, I mean things like a quasi grazer. I mean bodies like a worm. I mean bodies that don't have many attachments yet, we can start doing some stuff with those. So having, we like to call them DNA creatures, makes a lot of sense. But there's a ton of challenges. Ivo Herzeg has taken a huge look at it on our animation programming tech. He's already looked into this and we've started the investigations into it. So that's body.

And I realize it's a very long-winded answer, but that's a lot of questions in one. So we've gone over scars, we've gone over bodies. What was the other? Oh, cybernetics. Cybernetics is much easier. We actually have characters that use cybernetics already. There's something at Bataglia within the PU, and there may or may not be some within Squadron as well that we've used that. To do it on a character is pretty straightforward for us. It actually relies on the underpinnings of StarWear to work really well. Because StarWear will be able to adapt to different arms, different lengths of arms. There's a whole bunch of stuff that will come with those attachments. So I think that's shortly to come. There's no tech stopping us now from the cybernetics of it. But the question then begets, can you take it off? What happens to FPS gameplay when you take your right arm off? Right? There's some questions to be answered within...

It could be just cosmetic, but that would not be a very Star Citizen approach.

Yeah, but how ridiculous would characters look running around? Like, okay, yes, you can put a gun in your left hand, but right now, all of our animations very much expect that you're holding that weapon like this, so they're just running around like this? Like, come on. That's not really going to work out. Could we do it with NPCs? 100%, right? No problem. And I think you'll see a lot of that in the coming year for NPCs actually using that kind of stuff because it's there and we do a lot of it. So that is cybernetics.

I love this conversation because anybody that's followed the project for a long time has known that I'm really into the weirdy tech, like the FOIP tech or trying to scan things. As they might be aware, I did all the scanning systems for Squadron 42 for all our head DNA. So this is near and dear to my heart because, of course, when we go sit down on a scanner, as you have, right? So when we scan a head, we need to scan the teeth because we do give you the actual teeth of the character, because if you put somebody else's teeth in somebody else's mouth, it doesn't look right. So that's all fine and dandy, but we can't have people do that at home. In terms of scanning, nobody would want to do that. So there are ways to get the image projected out onto a character, but people have tried this before in games, and I reference a very specific project, NBA 2K15. Go ahead and Google it now. NBA 2K15 face scan, and you'll see the results. And they're very unfortunate, those results. So if you're gonna do it, it needs to work super frickin' well. Because again, you're gonna contort your own players' faces, and as funny as it is, it's not something you want running around in the 'verse. So the only time we would do it is if we could guarantee those results. Now DNA kind of takes us that direction. So what I think is much more interesting is taking a FOIP capture or a webcam capture or even just an image capture and driving DNA as close to that position as possible. And then bringing in whatever texture makes sense for that. You get the closest you can. So it's not a true scan. It's more of a DNA approximation of you rather than a full scan. I personally think that's probably a safer way to go. I think that's a very reasonable way that we could do this. Is it going to work? I don't know. I haven't tried it, but I hope to get there.

I think the Faceware thing, the FOIP thing, is probably the right thing, because you can get all the points in the structure from that, and then alter the structure, and then just find the texture, and you can get people half the way to making their face right there.

Yeah, and I think we've got enough textures, especially now that we've done the new scan set, and we've got like 60 or 70 of the new materials and all that in, plus the skin tinting, you can actually approximate somebody's skin tone pretty close. And there's a whole bunch of algorithms you can do to do that. I think it's worth an effort for sure, but I can't guarantee it because, again, I work a lot on the technology that enables the developers.

Ah, yes, so I mentioned it a little bit, the player heights. The reason we have different heights between male and female is we have different rigs. We have identical human rigs, but they are different heights and they've got different controllers, different offsets, all these different things. So that means that their assets actually get skinned independently. We also have different geometry in some cases. We take parity pretty seriously. We've made sure that when an armor goes out, it's got both versions. And we don't go over the top with, okay, the female version's massively different than the male version. It's darn close. I hope that most of the time you can still tell that there is a difference, but there is and it gets skinned separately. So it's dealt with as a separate asset. And commonly what happens is that somebody just messes up one of them or they're in a rush or we're trying to get a whole boatload of assets done at one time. And either the male one gets missed and the female one goes in or the female one gets missed and the male one goes in. So it happens. But we are really trying to keep parity in place as much as possible. So the reason those break down is just because they're two separate assets, completely different. In terms of design and the data on the design side, they're identical. So we have one, it's called a record, a data record. We have one record, and then depending on what skeleton that record goes onto, it will go down a different path of geometry in that record. So for example, we just set a tag, it says female, and we have other tags like hanging. If it's on a hanger, it looks like the hanging version. If it's in a locker, it looks like the locker version. So it goes through this little tree of subgeometries, we call them. That's why it happens. But we do take it pretty seriously, so most of the time that something like that comes up, we absolutely try to do it. Another big reason is some legacy assets, for sure, that just weren't created for a female in the first place before we actually started taking this seriously. Because if people remember, we didn't always have a female player, right?

The team has been pretty good about this when it gets pointed out and brought up. You mentioned how many assets there are in the game to track. Every single time one has come across my path, I've been able to message like Armando and been like, dude, this shirt's only worked on the male model for the last like six years. Can you help me out here? And it's not a high priority thing, but it doesn't make it any less important. He's always been really good. Shout out to Armando and everybody on the tech art team. They always find a place to squeeze it in between tasks whenever they can. You can't make the priority call above almost anything else, but there's always going to be those downtime moments between tasks and whatever. Whenever they can, they always squeeze it in.

Yeah. And not to use the dirty word, but tech debt, right? That is what it is. That's exactly it. That's unfortunately what you're seeing there, a bit of debt that's accrued because of decisions or priorities.

Well, honestly, if you count how many tools I've made, it's not nearly as many as how many CTG has made, right? And we're easily hundreds and hundreds of different tools, scripts, because it's like, what's a tool? Is it an executable? Is it a Python script? Is it a Python script within Maya? Is it a MEL script in Maya? Is it a Max script? We've got so many. Tech Art has a whole bunch. We've got CTG tools. We've got engineering tools. We've got designer tools. We've got production tools. I've got my Frame.io to Jira tool recently. So there's honestly probably upwards of a thousand. But it all depends on what the function of that given tool is, because what a tool here is designed to be is to force multiply somebody doing something, whatever that may be. And a lot of times it's to speed things up. That is probably how I spend most of my day, trying to figure out how much faster a person can go if I made a process easier for them, if I put something in front of them that made a lot of sense. So a lot.

And we've got entire teams dedicated to even different areas of the tools. Like just alone on the editor, right? We've got a team for the editor. It's not a gigantic team, but it's a very, very busy one. Again, there's thousands of tools just in the editor. Maintaining the Python implementation in the editor, maintaining all the plugins within the editor, whether that's a character tool, whether that's a Mannequin tool, whether that's the Universe Outliner, whether that's level design tools. So there's a whole bunch that happens there. So that's one team. Then we've got another team that works on kind of global tools. It's CTG Tools UK. Yeah, I know, right? It's super uninventive. The cheekier names get confusing, though. That's the only problem, right? So they work on kind of the global tools for all the developers. So this is anything outside the editor they tend to work on. So a good example is CopyBuild. When I got to this company, there was literally no way to download a build to your computer. You had to open up our Perforce, and I'm not sure everybody knows Perforce, but it's our source control repository management, and you would sync everything. Everything, everything. And you would launch it. And I was like, crazy. So they then got a build system. Great. Okay, so now the thing runs through a build and it makes a little package, an artifact. And the best thing they had was just copy-paste from network onto a local drive. Okay. So we have evolved, obviously, to, honestly, just as good as any other company that's out there in terms of our backend build infrastructure and our distribution, which is what CopyBuild does. So CopyBuild does the patching, does all this kind of work. So basically, almost everybody that is going to be working on the game runs CopyBuild. So they work on that. They also work on a very, very big tool called DataForge. And a lot of people will come on here and talk about DataCore or DataForge records or entity records or whatever. That is the tool that manages those. It's huge. There's a lot to it. There's a lot of data. And there's a lot of interfacing in that data that people do. I'm just rattling off a whole bunch of them, but that is what those guys work on, generally overall.

Then we have a newer team, a little bit newer, Montreal Tools. And Montreal Tools works on StarChitect, first of all. That's been discussed quite a bit, I think, especially with the new planets on the way. And that's going to be the way to propagate the locations out in a less manual way than we've been doing it. Because it's the only way that we're going to be able to scale, but that is when things are going to ramp up like crazy location-wise. That's going to get a lot faster.

We just delivered two star systems in the same calendar year. This is going to take us to the next level.

Exactly. And it's all part of Genesis, which is great. So they're working on that. They also work on some density profilers within the engine. We're also working on a system called Tiny Machine, which is a smart asset system. I don't want to get too into details on it, but smart assets are cool. It's basically, as you move something around the environment, it actually has contextual awareness of what it's doing. Just a super simple example is if I put a ladder up to a wall, it'll adjust to that wall depending on what that height is. That way, the level designers, it's not like, oh, I need the 10 meter ladder, I need the 15 meter ladder, I need the six. They just put a ladder there and it's tall enough and it works, right? That's kind of the idea with Tiny Machine, but it'll go much, much broader than that. So I think that'll append to StarChitect, but that's all kind of within its own realm. And then even beyond that, we have the online services tool team, the OST, and the Live Tools team, which works on a tool in our backend called Bootstrapper, which is basically the way that we run our environment without having to send it up to AWS. So you can kind of run the environment locally now. It's getting harder and harder to do that because our universe is bigger and bigger. So trying to run those servers on just a local PC can take a lot of memory. So if you're a developer and you're running Max, Maya, Editor, the client and Bootstrapper, it can get a little heavy. But there's all sorts of things going on with that. So there's a bunch of different tool teams that are helping the developers make this game happen, basically.

When you put the self-adjusting ladder and stuff in, will they still yeet people out of their ship and across the galaxy?

I hope not. But you know what that comes down to? And it is a thing that this fixes, is that comes down to different implementations. So like, okay, somebody made a ladder entity at some point...

Anyways, yeah, so that's a lot of tools. A lot of tools get made. And again, a tool is just as much a process as it is a tool. Sometimes the best tools are just rejiggering somebody's process a little bit. And it can work things out.

Yeah, we have a super rigid schedule now, and actually I got involved just a couple weeks ago just to check in on it, because I'll be super open about what happened. We made all the content, tech art kind of did everything, but it was lacking a bit of engineering support and a bit of design support to finish it off. Now, that's all been addressed. There's a strike team all back together. They've got a really, really good schedule. It's not long. I'm not going to say exactly when, as that will be a roadmap or a Rich thing to talk about. He knows when it is, but it's within the year to get that out to the players. And a big part of it getting prioritized was the community asking for it. So gotta thank the community for that. That was great, that they got this invisible undersuit by accident is what happened. And this invisible undersuit kind of gave them some of what StarWear is meant to do. And people loved it. And of course they loved it. I was seeing jackets on armors and armors on T-shirts and like, awesome, cool, super cool. But I see people saying, oh, just give us back the undersuit. We don't want to half-ass the feature, right? Let's not just half-ass it. Let's get the full StarWear out there to you guys. So design's come in. Engineering support's already in. So they're already working together. A small change happened once we got the designers involved. It was literally some really simple stuff, like splitting up the cuffs of the gloves, because the undersuit is like one suit and we want to chop it into gloves now. So there's a couple pieces of it that required a little more content work, but it was only a month or two of content work, so we're not talking the end of the world. Again, I'm not giving a month or two estimate on when it's coming out. It could come out before the end of the year, but the team's all back together and we've got the right amount of support that needed to happen for StarWear.

Well, I hate fun because it would give a false representation of what you could do. And if something that they could do with the invisible undersuit as a bug doesn't work in StarWear, they're going to then want that back within StarWear, whatever that may be. And that might be a design-breaking thing. So I don't want to give a false representation, I guess, towards it. I think we've talked about it. Armando had gone over it a little bit. So it does really go quite a long way in the customization of the thing. And one other quick little mention about it that's going to be really nice for players as well as for developers is once we get those assets all separated in the StarWear fashion, and a lot of this has to do with that DataForge record alignment and structure, once we get all that, we can mark up those assets or set information on those assets so that our procedural systems can start using it properly. So for example, if I know a t-shirt makes sense in a warm environment, I know sunglasses make sense in a warm environment, I know shorts make sense in a warm environment, if I'm spawning NPCs procedurally, they should probably have t-shirts or shorts or sunglasses or whatever. Or if it's sunny out, then I can just do something. It's these kind of pieces of markup that'll get really interesting.

Every NPC has a manual loadout.

Kind of. It's a set of manual. It's not truly. It's like we could have an NPC that has 10 different, they're called loadout bundles. It has 10 different outfits within the loadout bundles and they will all kind of work together. But yeah, somebody's got to make those, which is crazy. We don't have to do that [with this markup].

So I don't know about creatures this year necessarily, because right now, from the CTG side, I work for both projects. So sometimes I don't even know which one is for what. So I don't know of new creatures for SC this year. There's just always a creature on the way in somewhere within the pipeline. There's been a lot of the Yormandi. So there's always some creature in the group. So creature tech-wise, we're in this state where it just works like a character. They're very similar. But as I mentioned before, we're looking into all the scalable skeletons as well as creature DNA so that we can give some massive amounts of variation to these guys, because it's one thing to pre-can a couple assets and to spawn them randomly or so, but it's a way different thing to have the range of a DNA system or something like that. Even material systems need to be able to scale in the same kind of way. So that's where we're going. That's where we're heading, definitely that route. Now, we're also going to be adding a bunch more boids, like a bunch more of the more simplistic creature, not necessarily a full AI, but a critter that moves around that you can shoot and pick up, whatever. So that's kind of the state of creature.

And I want to tack on to that. When Sean says he doesn't know, he's not being coy like, oh, I don't know if we're going to make it. He's saying he doesn't know. He's not part of the scheduled content. He doesn't make the content plan for the year. He just works on the tech behind it. When other people determine it's ready and all the other aspects beyond his stuff make it ready, that's when it goes in. You often come around going, oh, I didn't realize that went out.

100%.

It happens a bunch. So he's not making a statement of cautious reservation. He's being quite literal. He doesn't know.

So, no, you can't right now. Obviously, because people have probably driven their cars into the trees and stuff like that. Small ones, I think, just disappear sometimes. But most of them you just run into. So the original Crysis, so it's clear, was super ahead of its time. It used a cutter object. It was really cool. And anybody that does 3D modeling will understand a Boolean subtraction. And Boolean subtraction is a technique. You put one mesh kind of over top of another mesh and you subtract the first mesh from the operand, from the next mesh, right? And it creates a hole. And what's cool when you do that, you can project the UVs and get the material from the cutter object anyways. We don't do that anymore at all. That whole system is long gone. What we do have is in Maelstrom, which is not quite rolled out, I think, to the Persistent Universe, but is very in use for Squadron 42. We have 2D fracture, procedural fracturable planes that come out of Maelstrom. We've got fracturable 3D meshes that can come out without the Boolean subtraction. You don't need to do that. It's much, much less heavy in the end. But Maelstrom will be the system we likely rely on for all of that. Maelstrom is a heavily, heavily engineered system for all kinds of different destruction. And I'm sure the community's heard a bunch about it. Physics talks about it and works on it all the time. And Chris is particularly involved and interested. In Maelstrom, if you say the word nearby, he will give you the feature list and the documentation and everything that goes with it. For example, I'm working on another area of Maelstrom right now for a given character, for armor destructibility in some sense. Now, this isn't a big feature that I'm about to roll out. There's a very specific character that has to have some armor break off. That's all. So we need to make that work. And we didn't have characters working with Maelstrom. So that's something that's starting to happen. So I expect that we will be able to break trees. I expect that the level of immersion we've always wanted out of this game will let us do that. But how that works with the planets and Maelstrom, I cannot say. That's going to be physics and planet tech, and they're going to have to hash it out.

I would say it's part of the long-term vision, but how we get there is a whole conversation happening right now. As the community is aware, VR is super new. Super new for us. It's just gone out in 4.5. Sylvan's done a phenomenal job on his own to get that working, OpenXR, awesome, right? But no controller support, right? And it's like the first thing you want to do, because already it's incredible when you go into it. I'll just talk about sitting in a cockpit, which is cool enough, right? You're sitting in the cockpit with your mouse and keyboard like this, right? But you see the character inside the cockpit is holding the joystick, holding the throttle. And you lean and stuff with the VR and it sees you moving and your character's moving inside the room. And all of a sudden you're moving the mouse and the keyboard and it doesn't feel right. You want to have controllers or you want to have your own actual joysticks or whatnot. But the immersion it gives you, and the things you notice in VR that you don't notice when you're just playing on a screen, it begs the need to grab it, to interact with it, to see hands, to move it around. Now, technology-wise, it's not an impossibly hard problem, and it's kind of solved in some cases, in some games. There's a great VR game, it's not that obscure, a lot of people that are into VR know it: Boneworks. Boneworks has completely done this. Their characters can look crazy weird sometimes, especially in multiplayer, when you are doing these kind of things. So a lot of these things, just like FOIP, will come down to not our capacity, but our tolerance for wackiness, right? Because again, as soon as you're driving the hands, because you could absolutely do this now, you could drive the wrists, the joints anyways that you would need to of the given hands, and you could drive them around. Think of what that would look like right now, right? A guy's running through the 'verse like... So there's a lot of stuff to solve, right? To make this so that it makes sense. Because we are totally unified in first person and third person, it's all the same thing. So we can't just show you one thing in first person, like your hand tracking, and it works and it looks amazing. That's kind of how Half-Life: Alyx works, which is an incredible VR experience. That is a bar to achieve just by itself. People will see you in third person doing these kind of things. So that's going to be a tricky thing. So it is something that we want Sylvan to be involved with. We want to start talking to design about it, try to work out how this is gonna work, because it's also not just about supporting one set of controllers. It's about supporting all the sets of controllers that are gonna do it. We've already had a tricky enough time supporting the gamepad, let alone these extraneous devices on top of it. So I think the interaction, it's impossible to say no, because you want it so bad when you are doing VR.

It's one of those things where, and I promise everybody, there will be a time where we'll have Sylvan in this chair and we'll do a whole thing about VR. We're going to let it continue to mature a little bit more and know a bit more about where we're going before we subject him to such things. But if you've spent any serious amount of time with Chris Roberts, you said our tolerance for silliness and whatnot, it's never going to be just okay. It's never going to be, oh, that's enough and we can just sit there. We're gonna push, and we're gonna keep pushing. Occasionally, chat, there'll always be somebody in chat talking about how are they still figuring this thing out in the year of our Lord 2026 and whatnot. And that's because this project doesn't end. If we all have our way, we are making this game and continuing this game well beyond 1.0, well beyond 2.0, 3.0, just going on. It's like any other MMO. WoW is still going 20-some-odd years later. That's our aspiration. So when we talk about these things, the end goal is not 1.0 and you stop. And that's why sometimes the tense that we speak in can seem like we're already looking at and planning the thing that comes next, and the thing that comes next, and the thing that comes next. So at this point, it would be foolish to say no, it would never get to a point where you can grab things and interact with things.

But think about it. We have diegetic screens, we have diegetic UI, we have the cockpit. Of course, of course we need to do it at some point. But when, who, what? We don't have a VR team, right? So the timings and details, I just can't give. But to your point, which is so good, we have to look that far ahead because we have to make decisions now that we don't regret later. It's happened a lot, right, where we go, no, no, no, 1.0 is just gonna be this, so we're gonna do this, we're gonna do this, it's gonna be great. And then that changes, and we're like, shit, now I'm wrong. I say it this way internally sometimes to people: you will eventually be made to be wrong at some point. Just try to project yourself as far as you can so you're not.

It's the number of times we've hit something because of a decision that was made two years, three years, four years, five years ago. And it was made back then with the best of intentions, with all of the knowledge and experience and foresight that we could have had back then. And then you just go, oh, well, we could either accept that, or we can rework the inventory for the third time or the fourth time. And we don't accept it. If we can make something better, we're going to make something better.

And we're not going to sell it short. That's what Chris says. I'm with him. I get it. Does it suck to eat that sometimes? Yeah, it does. But in the end, it does work out. And you've got to trust that.

Yes, totally. We have a facial idle animation, but I agree. It's not as animated as it should be. I think it's a little too idle. Not meaning it needs to talk and do all that, but blinks, a little bit of saccades. A lot of times you do those separately. You almost just put them in as procedural. So you've got eye saccades going on, you've got blinks going on. I think that makes a lot of sense, and actually it's something for facial animation. So it's clear, I know facial animation is spending the lion's share of their time right now wrapping up Squadron. But I'm pretty sure we can get something out of them. I'll mention it to Steve. I'm sure he'll have a look at some point. What also happens, though, unfortunately, is that sometimes that facial scope of the Mannequin animation breaks in certain contexts. So it might break after wearing this helmet and changing this armor and doing this thing, and something breaks that context. And now their face is just dead the whole time. So if they're seeing that, that's a problem. There should be an idle all the time, on everybody. And if they're not seeing it, that's a culling problem, which is something else. So yeah, we absolutely can add that.

No, it's intended for players, absolutely. And the work on it is way more gory than I would have expected. It was actually so uncomfortable. Story time. We were doing a little bit of work on that system before. I had shown what a face looks like. Okay, so I made a little video with Dexter music, and I had a knife in the game, and I sliced the face, and I thought it was super funny. Right, I thought it was great. Forrest thought it was funny, too. He was like, check mark. But for some, it was gory, right? So in the end, I had to blur out a couple of the screenshots. And I know it sounds stupid because it's like, oh, it's a video game. But it looked so frickin' real that it was kind of weird. It was around his eye and stuff. It was all very gory. So there's some questions there, right? Of how far that goes or how far that needs to go. But we're going to have all those abilities. And right now, it's a big deal for wrapping up on Squadron because, of course, there's humans and there's combat.

Yeah, but I don't know that the community is going to care that much. So we talk about tools, right? So when we did StarCloth, we didn't really make a lot of tools for it. We just made it work. So I marked it up in Maya. We could export it. It was super easy. But to really start polishing some of the issues that we had with different sections of it, especially within Squadron, but also within the PU, to really polish some of the outfits that we need to start making, like ponchos over armor and things like that, we needed some more markup ability. So what they've done, it was really cool, Vineet, the physics guys, we have Abhinav in Montreal, they've all worked together to make a physics plug-in for Maya that replicates StarCloth in Maya, and the markup that we do within Maya, the deformation that we do with cages and things like this, so they can see it real time. So what this does is just amp up the quality of the data that we actually get out of it, and it's so much better. Because yes, we would go in the engine, but I never did show the tools that we were using to paint in the engine. They were very basic. We have a system called ImGui that has this very, very debug GUI, and the physics guys had made a quick tool so that we could just kind of broad-stroke it, but you couldn't do anything sophisticated with it. Now that it's in Maya, super sophisticated. You still see it in engine in real time and stuff like that back and forth, so that's fine, but it's just more control over it. Oh, one nice... I get so excited over these dumb things. This sounds simple, but it's not. We can now put rigid objects, so for example, buttons. Buttons on the medical gown, there's an example. There's buttons that go down, or Bishop's trench coat. They're rigid. And they're on top of cloth. Before, they were just like cloth. They would move like cloth, and it's hard to see it. But now that we have these rigid bits, they actually affect the simulation too, just those rigid areas. It looks so much better. So there's a trench coat that Bishop has as well that has these, and it's totally not expected, but it has a belt that hangs down, a big leather belt of a trench coat. And the belt hangs down, and there's these little grommets within the belt, so we made all those grommets rigid, and it totally changed how this thing looked, and it was way more real, way more what you would expect from that. So that's a nice little piece of tech that came out of that little work. I know it sounds small, but again, I'm your developers' favorite developer. They think it's great, so I think it's great.

Imagine working for these guys, these developers, right? You meet them all the time. No, the biggest thing I want them to appreciate is that there are guys that are working for their favorite developers to do that job. We don't have middleware. It's the thing I'm really so proud of. Yes, we started out of Lumberyard. Yes, we started out of CryEngine. I've had a whole history with that. But we have become something so unique and so big that sometimes we're so busy on what we're doing, we look up, and we'll even do some competitive analysis. I'll go into Unreal, I'll go into Unity, I'll look at other games, projects or whatever, and realize how at the top of the game we are. We are progressing amazingly, and those tools people are part of the reason why. There are huge areas of this game that aren't getting the tooling they need, because I'm sure developers are watching this going, my tool's total garbage, right? Because I hear that stuff all the time too, every day. But it really depends on the area. It really depends on the process they're in. And it really depends on what they're working on. One of the things I like to say a lot of the time is that people like to complain about editor stability, is what I hear a lot. I'm like, hey, do you play the client? They're like, no, I'm always running in the editor. I'm like, well, listen, the client's running in the editor. So the editor is no more stable than the game. So please make the game stable and that editor gets more stable. It's things like that. It's just the reality of the work that we do that the tool that it comes into gets the blame. CopyBuild, the one I talked about at the onset. Exact same thing: CopyBuild's slow today, CopyBuild's not working, CopyBuild this, CopyBuild that, and in the end, it's never CopyBuild. It's always a network problem, build store issue, patcher bug. So, honestly, it's a tough job to take that kind of stuff, but it's also incredibly rewarding, because I was just doing some work today, and I even said to the guy, just tell me it works, I need that dopamine hit. And realizing that just satisfying sometimes the request or the need of a developer honestly gives me a sense of pride and joy and reward. So it's a great job to have, but it also can be tough on the same token.

Oh, yeah. Well, and I mean, private lobbies is something that really hurt my heart from the standpoint of I was running the team when we had to remove them. And it wasn't something that we actually removed. It was we had changed over some details that Duncan will probably get into, some details with the lobby system so that we couldn't continue to support the private matches. Now, it was always the thing was next quarter, we'll bring it back. Next quarter, we'll bring it back. Next quarter, we'll bring it back. And it was always something got, the resource got taken, something got missed or whatever. So finally, with a little bit of collaboration with our new friends at Turbulent, we've been able to bring back online private lobbies. So Duncan, do you want to talk any more about that? Maybe some details over why it went away?

Yeah, so it went away essentially because we did that big refactor where we got the friends and party systems to the new system that we've got that you're using in the PU now. And that just simply wasn't compatible with what we had in Arena Commander. So the option at the time was we delay this or we shut down private lobbies for now and we revisit it for the next thing. And as Sean says, priorities got shifted, people get moved on to different things and it just kept getting pushed out. Now, thankfully, we're working on it. And we didn't want to just bring them back. We want to do better with them. We want to give more. So private lobbies aren't called private lobbies anymore. They're called custom matches. And custom matches leans into you'll be able to customize more things about it. And we've got so much more planned for it. For now, we're getting what was old back with private lobbies. But we've got so many ideas, so much requests from the community for what we can add there as well.

Yeah, so this leans into the whole "we want to do better with the private matches." So the almost obvious thing is letting you arrange your teams beforehand. So we've got the idea of you've got a party leader, you've got moderators that can help switch people between teams and whatever. And then you've obviously got the members that join. And then for right now, we've got that team organization. And we're hoping to get some custom settings in and expand upon that. But yeah, bringing back org content to private matches is such a big thing for us because we knew how much orgs used it. We saw the weekly posts on Spectrum like, hey, we're running an org event. Who wants to join? And as soon as the private lobbies went down, that went away. And I'd get pinged constantly about, hey, when are we getting them back for that specific use? And we really want to start accommodating that again.

Yeah, because basically the competitive scene basically shut down without private matches, without being able to customize the match parameters. And it's something that we always want them to be able to have within AC, because ultimately the PU is a multiplayer experience. AC is played best in multiplayer with friends, and we want that to be as easy as possible. And honestly, being able to set up your own little scenarios, I think that's really cool.

So there's actually a lot to this question. Now, right now, behind the scenes, we do have, there is a ranking. And behind the scenes, it works here and there, but it's really not as simple as just a single value. We've come a long way in matchmakers in the game industry overall, where it's got to be region-dependent, it's got to be skill-dependent, it's got to be time-dependent. There's so many different angles of matchmaking. Sort of with that, Duncan, do you want to talk a little bit about what we're going to do for these guys?

Yeah, so as you guys know, we have the leaderboards that are currently online and they haven't really been representative of, well, leaderboards for a while now. The algorithm behind them broke. People have found many exploits behind them. So we want to completely rework those at the same time as doing this proper matchmaking. And that matchmaking is the thing that will have to come later because it needs more help from Turbulent on the online side. But the competitive community in Arena Commander and Star Marine is so passionate, I'll say, about leaderboards and that matchmaking behind it. And we really want to hone in on that and especially push Squadron Battle and Duel to be our primary competitive modes because we think those better reflect those as well.

Yeah, so we have a lot of ideas. We had a brainstorm session where we had so many ideas of different ways to reward people. And right now, we're going with the simple wins. So we're keeping it simple with Spectrum badges because they're fairly easy to get. But we've got some more, I would say, challenges coming up, which we're hoping will have better rewards. It's more than just a Spectrum badge. We're talking with PU teams, the Spectrum teams, and seeing what more we can give that's got a little bit more flair to it. And even within Arena Commander itself, we want to reward people with stuff in-house to Arena Commander. So Arena Commander actually supports skins and all this stuff. So we're hoping to get those as rewards eventually.

Yeah, so to piggyback off that, the questions in regards to some of the matchmaking, ELO ranking, as well as this are all related. What this has to do with is some level of progression within Arena Commander, a reason to play this. And a lot of us, this generation of gaming or before, we kind of need some progression. You need some level of a little bit of persistence and a little bit of progression within each one of your game modes, so that you kind of feel like there's the reason to keep logging in and to keep playing it. And we want to have all of those things within Arena Commander. And it's nice because we have an entire game of content to pull from. And a lot of these are the same conversations that we had during times of when Firesprite was working on Arena Commander. It was all the same sort of progression that we'd like to still put in place here. But we've got to get all the fundamentals, just like we talked about, in place first, which is the badges, the experimental modes, and private matches. And now that we've got that kind of stuff, now we start making more reasons for the players to be there and a little bit more driving the engagement, if that makes some sense.

I think it would be a pretty obvious choice to have a special paint or whatnot that says "I kicked ass in Arena Commander," and you get it through earning that.

Well, it's even further, because what Chris wants, and definitely something we've talked about at length before, is the connection between Arena Commander and the Persistent Universe. Because for a long time, we've had the sim pods; the sim pods were in the hangars. We want them in the Persistent Universe. We want it so that they can go from the Persistent Universe, jump into a sim pod, be in Arena Commander. You're in this simulation that we talk about all the time. But we don't want that to be meaningless. We don't want you just to leave and go to a different sort of game. Why not be progressing Arena Commander reputation? Why not be progressing some level of persistency within the PU so that all the stuff you're doing in Arena Commander is actually benefiting you in the PU or vice versa? What about things happening in the PU benefiting you in Arena Commander? What about ships you own benefiting you in Arena Commander? So there's so many layers of connection that can be made once that progression and persistence, I guess, and I hate using the word persistence, but once some score persistence is in place, because the persistence I'm talking about is not the Persistent Universe. This is just keeping a score.

Yes. I want to make that very clear that yes, we do want spectator modes. And I'm going to go off on a little bit of a diatribe as I do, because spectators to me in Star Citizen and in Arena Commander itself are more like newscasters is what I always saw this as. I always saw this as a newscaster role, as a reputation or something you could do where there was this idea of always being able to record things that were happening in the Persistent Universe. And I think you've heard this pitch from me before, Jared. Where you would go in as a newscaster, you'd want to put "press" on the side of your ship and all these kinds of things. But it'd be dangerous to be capturing battles and things like this that were going on. So I like the idea of spectators actually being a real thing within Star Citizen. But within Arena Commander, I think the little more classic game sort of standpoint makes a lot of sense. Just to be able to, especially during the race, like during qualifying or whatever, you want to be able to spectate and see what's going on. Especially now, oh man, with these new racetracks and watching some of the Evocati people play them. Yeah, you want to be able to spectate these guys because they're freaking good. Really good.