Reverse the Verse: February 19th, 2016

Yes, but we can actually have more than one.

Yeah, there's a thread on Reddit: they're going to fire Jared, look at his job listing. First of all, if we were going to fire Jared, we wouldn't put the listing up first. We're smarter than that. There's 1.2 million of you out there and only one of him.

Yes, Ben was the original community manager. Ben, and even Pete Mackay was community manager for a while.

We had James Pugh and Will Lewis. All sorts of people come and go.

And absolutely we need more than one. Pete's now a designer. Ben, you're now Director of Community Engagement, Online Strategy and Spaceships. Oh, and we had Patrick.

Patrick had been our lead moderator, which is what a community manager does at most companies. So it all swirls together.

So basically it's all on us at the moment, so we're looking for some help. If you're local to SoCal, that's a big help because that's where the position is, and if you're interested in helping out, apply online.

We're looking for people who love space games, who get Star Citizen, who are passionate about community building.

It's funny you should ask. So we keep talking about how we want to come in on Saturday and watch. If you guys saw last year, we did all six Star Wars movies in one day. It was late last year, December. It was exhausting. It was fun. Somehow hundreds and hundreds of people watched that and enjoyed it. And people keep saying, will you do it again? So we're going to do it again tomorrow. We talked about doing a trilogy of Star Trek movies. We talked about doing Indiana Jones, all the Starship Troopers movies. And we kind of just settled on, hey, we just kind of want to hang out outside of work, but at work. So we've picked 10 possible options, things that are available on the US Netflix so people can watch along with us. And just before this show started, I put a poll up at the off-topic forum, that's the one at the very, very, very bottom of the forums, asking people to vote. And we will watch the top two. Tomorrow afternoon in California. We're probably going to start about 1 p.m. Pacific time. Do a double feature. We'll watch a movie, have some pizza, watch another movie, and call it a day, I think. It's all in the sci-fi genre or thereabouts. Some are classics and some are classically bad. It's anything ranging from Star Trek and Wing Commander and Starship Troopers to stuff like Space Buddies. We have a couple of options there if you want to troll us.

Have you guys talked about the player hostility stuff?

Only mentioned it. Just that it exists.

Right, okay, so player hostility exists. There'll be a big write-up on it, and I don't want to try to explain it and get it wrong. Mostly it's like a reputation. It's not a very granular reputation system, but it's really meant to...

It's the very first iteration of it.

Right. And again, with Star Citizen, we always want player-generated missions, because that's what gives us the emergent gameplay. So it's things like having a distress beacon and dropping a distress beacon. Now, maybe a player's actually in trouble and he drops a distress beacon. Maybe a player drops a distress beacon because he wants to bring people in. This is what I mean about emergent gameplay. So with the reputation system, if there's somebody griefing other players, they will go up in sort of a red reputation level to a point where there will be a bounty put out on them, I believe. I don't know exactly on the design side what you give the players for killing these guys, but they're basically bright red. They become a mission in themselves, basically, to go get rid of these guys. And it's not to get rid of them, but it's more of a challenge if somebody wants to put themselves up to a high reputation or a high... What's the wording we're using? I don't even know off the top of my head for the design side. But a high hostility level, basically. There'll be a mission and you can go hunt these guys down, you can group up with a couple buddies. Most MMOs have something to this effect, where a PvP player is going to be bright red. I played Lineage 2 for a number of years, and if you ever saw a bright red guy coming your way, you're going to take off and run, or you're going to try to group up with some people and try to go fight them. So that's pretty cool. It's a great system, gives a lot more gameplay, so that's fun.

There's a lot of stuff on the tech side that we fixed up for 2.2. Tons and tons of bugs that we've rectified. Some of the bigger ones are surrounding EVA, getting in and out of the ships, transitioning between different physics grids. These are all kind of unsexy things, but very, very important to be fixed.

One of the other things that we're working on: even just two days ago, we were running a bunch of our assets through a build process. We use something called a resource compiler, and it's kind of a magic black box in the build system that takes all of our source assets and converts them into a target format. So you export a model from Max, it then goes up on Perforce, the build system pulls down from Perforce, compiles the asset, and it may or may not have errors, and then it ends up in the build. Now, what you showed very nicely on your Twitter the other day was a tomato man. So tomato man is actually just replace-me balls, is what they are. It's a really simple sphere that has one texture map on it that literally says 'replace me'.

So something interesting with that is that we have wanted to convert our mesh format for a while for most of our assets. In video games and real-time rendering, we use a certain floating point. On consoles, you kind of have to use a 16-bit floating point, whereas on PC, you can allow some assets to have a 32-bit floating point. When it comes to an asset, what that means is that the position of that vertex is accurate to a 32-bit floating point versus a 16-bit floating point. And as everybody knows, if you've ever converted an analog signal to a digital signal, there's quantization that happens. And quantization means converting 32-bit down to a lower fidelity, basically. What I mean by fidelity is the accuracy of the vert. So what happens to the artists is that they'll run it through 3ds Max, they'll export it as 32-bit floating point, but the build process will then quantize it down to 16-bit. And what that can do is it'll take a mesh like this microphone and it'll crumple it just a little bit. This vert will be just accurate to here. It's like rounding, is what it's doing. So when you round a mesh, it can destroy it. And our artists spend so much time on every single vert being just right, so when they see their asset in the build, they freak out. They're not happy, for obvious reasons: hey, I made this asset, it's really gorgeous, and now you've run it through your build process. You've made it optimized, great, but you also made it ugly, or whatever has happened to it. So basically, Carsten had worked on a new mesh format that we want to support, and this is a flat integer 16. And we're running all the assets through that. But of course, anytime you do that, you end up with errors. So we have about 300... QA was super excited about that yesterday, when I said, put a Jira, put a task for every single one of those issues.

And Jake Ross came in this morning and saw a list of 300-some odd.

Right. And then they tried to tell me, maybe we'll just release with that. I'm like, no. We're not going to release with that. We're going to fix the issues. And here's the thing: they're all super easy fixes. It's not that they take a long time. It's just that QA is not happy about sitting there making 300 bugs. But I timed them. It took them 35 minutes as a group to create all those. So I'm like, you guys are doing fine. Didn't take all day. So we just got to get those assets fixed now. I think the idea is that we might not get through all 300 of them by the time PTU goes live. So there may be a missing asset or two. If there's a wall not there or whatever, then that's what's going to happen if we want to release it with the asset errors now. Will we get them fixed? Absolutely. It's just how long that's going to take, I don't know, and that's up to production.

You're going to know it as soon as you try it. It's drastically different. It's a little bit apples and oranges because the control method is completely different. It's not animation-controlled, it's purely physical, which means it flies a lot more like the ships. And if I had to give a real benefit to it, it's more unified with how the ships move around. This was always kind of frustrating to me: when I jumped out of the ship, all of a sudden my control scheme was drastically different. There needs to be some level of unification when you go from EVA to ship to FPS combat. It's got to be similar. So probably the biggest benefit is that it's more like the ships. There are tons of other little add-ons to it that'll make the visuals of other people in EVA a lot cooler, and that's stuff I've talked about before.

Like the change in the center of mass?

That was an ugly one. When you rotate around the feet, that looks ridiculous, so now you always rotate around the center of mass, regardless of what's on your body, which is pretty cool. So if you have a helmet, it actually adjusts a little bit. If you don't have a helmet, which, okay, that's not going to make a lot of sense out in space, but just to give an example of the weight distribution on your character. And then on top of that, we've got parameterized thrusters. This just means that the amount of thrust, the effect that comes out of the back, is actually driven by the amount of thrust that you're applying. So if you just apply a tiny bit of thrust, it's just a tiny little effect. And if you're sitting there pedal to the metal, then it's...

And the UK audio team have redone all the sounds for the EVA thrusters.

That's right, which are also parameterized. So every time we send some sort of throttle into it, you get a different sound. So it's not the infamous psst, psst, psst. It adjusts based on how much you're throwing in.

I haven't even worked on it that much since then, to be honest with you. Marco Corbetta, Carsten Wenzel, and a guy named Pascal was doing a lot of the art. The main three, anyway. But again, it's a team here. There are bits and pieces every month.

That's an important thing that gets lost lots of times. You may see the same people in front of the camera giving you this information, but this is a team of over 300 people just in our company, and then all the outsourcers that are working on this. It is definitely a group effort. You may be the one to present something, or Matt may be the one to present something, or us.

Right. Doesn't mean we actually made it. And trust me, those are real uncomfortable demos when you've not had anything to do with anything and then you're the one having to show it. Luckily, the procedural one I had a little bit of say in.

I like that we were showing procedural generation for the first time and the thing you were most nervous about was being able to land.

I'm not good at landing. But I would argue it's because our landing system is not good, not because I'm not good. People were watching my video of the landing, like, why didn't you just go decoupled? What you don't know is that the thrust of the ship was 4,000 times what it was meant to be, so I had to be real touchy with it. Some really cool stuff got working on the planets this week. We've shown some physics stuff with it, so we had talked about the fact that we have physics on it now. There's been some really nice work on the gravity for the planets, so that's really cool. But with Chris being out there in Frankfurt with the guys directly, I can imagine that every day he was pushing something that they had to get done. So, some great work this week anyway.

It depends on what animation. We're always updating them. So yes, the locomotion sets have been in there for a while, I think. One of the big refactors that we're doing, and this is super behind the scenes, is that we're changing our animation pipeline. I hate to just throw out the word pipeline because it's so nebulous, but we're changing the tools that we use to actually export these animations. Doing that requires a ton of work. The problem with it is that we've got something on the order of 3,500 animations across the project. That includes ships, characters, anything that moves in the game. So whenever we do a change like this, we have to make sure that we support all the legacy assets. Basically we've got to batch export all these things. It requires a lot of tools work that is, again, a very unsexy thing to be working on, because you've done your job right when you have exactly what you started with, which seems very strange. But you are then on the new pipeline, which is quite nice. And again, this is all about scalability. You get those animations done in the first place; they took you two and a half years to do. The hope is that as you improve your pipeline, it takes you half the time. Then you take another half off that. Then you take another half off that. As long as you can scale a lot going forward, then it's a good thing. And that's the idea with the FPS animation. So we've been converting this over to another pipeline.

And having said that, one of the things that Steven's actually working on today is we have a whole state for the hunch. People have probably noticed this: in the hangars, when you walk under a ship, they wanted to have a whole movement state. But it makes no sense for this thing to be an entire movement state in itself. Most of our animations are split up between sort of a full-body or an upper-body kind of animation, and we author them very differently. So when you go walking around the hangar and you walk under a ship, your guy hunches like this. What's actually happening behind the scenes is you're going from a normal relaxed walk locomotion to the relaxed hunch locomotion. But a locomotion set takes about 96 assets for the full body. And with the hunch, you don't walk that much differently than you do when you're just walking relaxed. So why is the hunch a whole separate motion set, when it could just be an upper-body additive? What I mean by an upper-body additive: you take exactly what your relaxed walk is and then you add one layer on top of that that says, oh, hunch a little bit. So rather than initiating a whole other set of 96 assets that you've got to stream in or keep up to date, we're just going to put in an upper-body additive. So the state still exists, but you're actually reusing what the relaxed no-weapon set would be. So that's one thing. The other thing he's working on is a bunch of the grab animations. Somebody put a placeholder grab animation in at some point in the project and it's just stayed there forever. So anytime you go pick up a weapon, there's something really, really strange happening right now. So he's going to fix that up today with the new pipeline.