Star Citizen Live: All About 3.17.2

No. So basically what the nav mesh does is it allows AI to traverse the planet so they can walk the Earth like Kane.

So back to me, AI planetary nav mesh allows AI to walk along the planets?

Correct. And then we need behavioral changes in order to actually get the AI to use special types of weapons. And then from there, we'll also need different sizes of nav meshes in order to have larger creatures or smaller creatures navigate properly. And then for vehicles, that's a different tech discussion, actually getting them to drive around, and it will be extra tech in order to get them to do that.

As is common with many aspects of Star Citizen, this is the first tier, the first implementation. It allows us to test it out live, because it's always worth reiterating that even our live environment is an alpha test environment.

Most of this patch has been all about content, so we're proud of what we have done there. Right now, nothing's really popping into my mind. There's things that we did that we didn't announce, and I prefer to keep it that way. I'd rather let the players go and explore and find it themselves.

Since it's already discovered, things like the Snake Pit racetrack and stuff. One of the interesting commentaries that we see on ISC sometimes is, why aren't you talking about this? Why aren't you explaining this? Or why did we discover this in the game? And I think it's always worth reminding folks that while we absolutely have our duty and our intention to inform the community and discuss the behind the scenes of the project, we are still making a game, and the game is the primary experience that we're creating here. And I say that as a video person. So when we can leave things for players to discover in the universe, those are always a special treat. And you'll see that more and more as the game continues. As more content goes in, we'll pull little things back from ISC and maybe hint at them or allude to them. But we really do think there are some things that are just best left for players to discover. Like when you log into PTU one day and there's suddenly a Centurion there.

Are they asking about being able to walk the ship?

Yeah, could the dynamic nav mesh just be created to make AI suddenly be able to walk through ships and stuff?

That's the goal, basically, for us to have the ships there and have nav mesh actually inside the ships. And then let's say a ramp opens up, then the nav mesh is dynamically spawned, saying, 'Okay, hey, I can connect from the planetary nav mesh up the ramp, then into the ship.' So right now the team is working through a lot of those connectors, whether those be jumps, ladders, so on and so forth, so that the AI should be able to connect and go through all the different instances that they need to in order to navigate wherever they need to.

And the second half of that question was, dynamically, if I put a cargo box down there, is the AI going to know that they can no longer move there?

Yeah, it should. Obviously, there will be points in time where I guess the players can really mess up the AI, if they basically stack a box or build a baby gate, but we'll need to work out how we can fix that and how we can change the AI's behavior in order to possibly move the boxes out of the way or even find a new way around.

Basically, there's a couple of things that happened. There's the player and AI networking changes that have gone in, and this was roughly about six to eight months' worth of work from Daniel Wilson. Before, it was very dependent on the server frame rate and latency, and now what we did was we actually decoupled it. It really breaks down into two things, interpolation and extrapolation. Interpolation is basically smoothing out the data that we've received from the server. And then extrapolation makes up the information in case we don't get an update from the server in that time frame. So basically what it leads to is less desync, better dealing with lower server frame rates, and better-looking remote movements. Now, it's still a work in progress. They're still tweaking it and will still improve it as soon as we go live, just because the live servers behave differently, because they're more heavily loaded versus the Evocati servers. So we'll go through and we'll adjust things and make sure that how it runs in the future is basically tuned and tweaked based off of it. So from there, what we're also having issues with is that there were some modifications that went in for persistent entity streaming, and now we're looking at the server frame rates and trying to make sure that everything is basically being changed and adjusted, because right now our server frame rates are running under certain loads that the AI do not play well with. So anything that's lower than a certain frame rate, basically we're getting desync, bad server frame rate, behavior updates. Lots of the behaviors have been updated and rebuilt for S42. Once those are in a good state, then we'll move them over to the PU. And then there are still some changes that went in that were throttling the servers even more than normal. So basically, we've been looking at what was implemented and adjusting it from there. So it's kind of a hodgepodge of multiple different factors, but it really boils down to server frame rate, desync, and some of the behavior updates.

The answer to every 'when' question is always the same: it's when we think it's ready. It's an alchemy of players in the PTU now testing and developers working behind the scenes to find and smash bugs and resolve issues, and there's no easy way to predict when everything just kind of coalesces. It's when we think it's ready. It'll move into the next step, whatever that next step is.

Once it's stable enough, with some of the stuff that we're doing.

And going back to that second question about whether all the bugs are being fixed, 'once it's stable enough' is the absolute right word there, because remember, every level of Star Citizen, even live, is a testing environment.

Well, think about it like onions. Onions have layers. So we're just peeling off layers. As soon as we get things stable enough, then it might be revealing other gameplay issues, and then we fix those, and then it might reveal other crashes that we're dealing with. So onions or parfaits, whichever you choose.

Basically, for cargo and mining, I still view them as a squashed mission or squashed game loop, just because, in particular for cargo or long hauls, we're expecting that you're traveling through at least one solar system, if not multiple solar systems, and then the danger that would be included in that should be how we basically play that and pay that out. And then we've got elements like, when you're talking about mining, right now the mining loop is: I go spawn the ship, I go out, I mine, I take it to a refinery, and then I go back and mine, refinery, go back, mine. What we want to get to, once we get cargo working, is that the idea is that you get to eject the pods, and basically you stay there, you mine, you eject those pods, somebody else picks them up, takes them to the refinery or refines them in their ship, and then goes and sells it. So that's the game loop that we're looking for for those two. And we'll go through and we'll adjust it as those things start opening up and start being fixed and working the way that we want them to work. And then we'll do more tuning based off of that.

Yes, there has been. Basically, the way that we break it down internally and think about it is that there's really three different types of derelicts. There's inhabited, which basically will not be dynamically generated or anything like that. Then we've got some of our other stuff that will be dynamically generated, which are what we consider fresh and old. Fresh would be like, it's a smoking wreck. You can go in there, you can get items removed. You can go in there and hull scrape, possibly even munch it. And then there's the old ones, where you probably have limited item removal just because of the age or wear associated with those items. And then we'll need to see how they visually look, but my gut feeling is probably no hull scraping on it. But then you'd be able to munch it up and salvage it and basically get the warped material, or basically aged material. So that's the way we're thinking about it. Even with inhabited, the idea is that it's so worn down that you theoretically can't go in there and hull scrape it. But we'll need to see how much of it we want to salvage, whether it be item removal or something else.

For me, the hull scraping is cool, and it's very visual, and it looks cool, but personally, I'm looking forward to the day when we can go and strip components out, and like, this is a component I can actually put in my ship, and find ultra-rare, super fancy military components or whatever, and like, oh, this is actually an upgrade for what I have in my ship, pulled out of the wreck of an old ship. I'm allowed to have wants.

That's where we're heading. That's what I want.

I can see how that can happen or how it can be perceived. Basically, the way it breaks down is that we have teams that solely focus on S42, then we have teams that support both, but those support teams right now are really focusing on S42, and then we have PU-specific teams. Currently, we have three different feature teams that are PU-specific: US PU, MFT, and EU PU. And those teams are working on cargo, missions, hull scraping, resource management.

Things like the Siege of Orison.

Yeah, exactly. So those are really what the feature teams focus on and work towards. And then we have content teams. So we have three EU content teams and then basically one to two content teams in Montreal. So space stations, outposts, underground facilities, derelicts, building interiors, and other points of interest are really their focus. Each team takes one of those mandates and runs with it. A lot of the EU teams have been focusing on Pyro, and so with the space stations or with the outposts, those are really their main focuses. And then the underground facilities team was kicked off, I would say, about six months ago, and they've been building up what the new underground facilities would be like. And then obviously the Montreal team has been working on derelicts and building interiors, and then, like you said earlier, sneaking in the Snake Pit racetrack.

Yes. Well, it depends on what they're talking about. If they're talking about the mission ones, right now we have a tech request in, because we technically can't spawn loot crates on those ships. In the future, we'll be able to do it, so that when you're going through the ships, you'll be able to pull stuff out and get different loot. But on the inhabited derelict, it should have that already in, and you can grab things there. Now, it's not really different than any of our other crashed derelicts, space station, Luchin, or anything like that. So it would be pulling from the same database that those are pulling from.

Is it because you can't spawn on a spawn kind of thing?

Basically, yeah. It's the dynamic spawn of those that is causing us issues. So it's something that I want to remedy soon, just because we are moving towards a lot of these instances where we want to dynamically spawn the fresh derelicts or the old derelicts that we talked about earlier, as well as any sort of missions or anything like that.

Well, the story is really about selling the age of the location and making sure that things that you would find there actually feel appropriate to the time, what they might have had at that time. So with the derelict outposts that we have right now, we wanted it to feel older and we wanted it to feel different. That's why we introduced some of the new things. And it's basically gearing up towards Pyro as well, just because, again, we want to make sure that we have a collection of age on these things, to sell different time periods. I guess the equivalent would be finding a fossil or finding an old Roman coin or something like that. So that's the feel that we're going for in these locations.

It's up to the people to determine the value and the worth. Somebody will choose to collect a bunch of them. Somebody will choose not to care about them. Somebody will choose to want to build the largest collection of them, and a sort of proto-economy will evolve out of that.

And even then, we want to have some that are in pristine condition, some that are worn down. We have an idea regarding old toys or books or liquor, and being able to go through and find the different bottles, or find the different worn-down books or toys that might be in pristine condition or broken.

Yeah. Right now, with the planetary nav mesh, the AI, I guess if we let them, and before we were seeing this, they would just go and walk. They want to stay around certain things. But yeah, we're going through and we're updating all of the other locations with planetary nav mesh, as well as getting AI around there, including dropships coming in or anything like that. So the goal for us with these is basically: let's get it working in one or two areas first, or one or two scenarios. And what I mean by that is the outposts, to me, are a scenario that we put together. Then from there, we'll take it wider and take it to outside landing zones, or, sorry, the old high-tech outposts, or the bunkers or anything like that. So the idea is that we do it once, then we hammer out a lot of the issues, and then we take it and we go wide with it.

I'm not sure I understand 'a lot of quick releases.' You just mean the patch updates to the PTU and stuff? It's pretty normal when we have new content or new fixes; we just push them out.

Well, part of it is we also watch. What we don't want to do is put out a patch update every day. So we do watch how often we put it out to Evocati, and we do watch how often we put it out to Wave 1 in particular. We want to make sure that there's enough sizable fixes between those in order to make sure that what we're pushing out is quality, and it allows us to basically take the feedback and adjust it and fix it. Now, I think with some of this stuff, what we were trying to do was work through some of the stability issues that we were having. We were getting a lot of crash errors on just one or two things, so we'd get these fixes in that we feel would unlock things, then we would do an update to that and push it out. So that might be what you're seeing.

Well, if there's something that comes up that we need to hotfix, then we do that. Currently, most of the team is actually working on whatever tasks they needed to do. Basically, once we hit the end of the quarter, those teams are allocated to start working on other things. Most of the teams have already started to work towards their 3.18 goals, or basically their Q3 goals. And then from there, if they get a LIMAS fix or they get anything that needs to be looked at, then they jump on that, fix it, and then go back to their current tasks. With 3.18, it works just the same way as we worked in the past. Obviously, with the introduction of persistent streaming, there's a couple other things that we need to [handle]; we're introducing that into the streams right now. And so the teams will jump on it and fix it and then go from there.

Because, again, turning off collision damage on this stuff, I think, will have other knock-on effects that we're not prepared for.

Physics is at the core of so many aspects of Star Citizen's gameplay, and the universe we're building is the systemic playground. You take out a core, fundamental system like collision, and who knows what else happens. Plus, it's hard to go back on those things. If it ends up taking months or longer than that to fix, people get used to no collision, and then when you turn it back on, there's all those people that have joined when it was no collision, and it's like, 'Oh, you suddenly added collision.'

Well, it's also that right now, if the player gets into a position that they shouldn't be in and we need to kill them, or it damages them, we need hazardous areas. We need those types of things to create tension for the players. If we turn off any sort of collision or damage for collision, it might be that we have knock-on effects for fall distances, or players can get into areas that they shouldn't have been able to get into, weasel through walls or anything like that.

Combat logging? I'm assuming that they're talking about logging out?

Yes. Let me tell you. I have chosen to spend my night going out there and putting the hurt on fellow humans of the Stanton system. I am a pro pilot in a pro ship that's pro kitted out, and I am putting my talent and skill to the test against people, and right before they're about to die, right before they're about to explode, right before I'm about to achieve my victory, they log out.

Yeah, I guess for me it's logging out, because logging in development terms means that we're actually tracking something. So for logging out, there are plans for it, very similar to other games that have done it in the past, where if you've been in combat within, let's say, the past five minutes or anything like that, then you've got another five minutes before you can log out safely, or 15 minutes. We need to be able to sit down and tune that. But yeah, we're very aware of it. We want it worked on soon. It's something that CR has called out, something I've called out, player relations, Tony, so on and so forth. So it is in the back of our minds. We are thinking about it, and it will be fixed soon.