KSA | The KSP Replacement from RocketWerkz | Seamless Movement and Terrain

[whatsEJstandfor]

I'm so so so so hopeful

[PDCWolf]

8 hours ago, Kerbart said:

Nate mentioned a lot of things too, and had plenty of animations to go with it,  Don't get me wrong, I think it's great that the design team signals they're aware of performance challenges and want to build something up from the ground. But until there's an actual product to show, that's a story we've been told before. So it's great that this is happening but nothing to get too hyped up about.

4 hours ago, Skorj said:

This stuff gives me hope.  If you can't skip time arbitrarily forward, you can't scale to huge numbers of ships in orbit.

We've all been hurt before.  No hype, but hope.

Don't take it as hype, take it as a signal that they're indeed thinking of the long term and not just building another KSP1. In fact, since it's so early and there's no roadmap, it's good to know they're thinking about this stuff in concrete terms.

5 hours ago, Lisias said:

I would suggest taking a very close look on Orbiter's source code - and, Life isn't interesting?, we would have gone full cycle!

Funny. Rather than full circle, I'd say it's an ascending helix: KSP was posted to Orbiter's forums, now KSA is posted to KSP's forums. It's a circle but we're also apparently going up.

[PDCWolf]

I've been digging some (not much) and here's some quotes from Dean about KSA. Of course, don't take anything as set in stone:

On using kittens:

Quote

That’s actually what I really like about using kittens. In my original pitch for KSP2 one of the things I really wanted was for your manned launches to really matter. I like the progression shown in tech mods like probes before crew

On scale and patched conics vs NBody

Quote

Scale and N-body versus patched conics are probably biggest difficulty factors I can think of. I think we are aiming for a base of roughly the same as KSP and then modders can change it.

Scale

At this stage our current thinking is basically do do somewhere between current KSP and x2-2.5 current KSP size for both the bodies and their orbits. In other words, we are aiming to replicate the same feeling, commitment, and challenge of existing KSP. We feel like base KSP is a great compromise between many factors when it comes to scale, and so we are not trying to reinvent that - instead focused on solid datastructures and ease of development for modders to fill any gaps.

Patched Conics v N-Body

The core focus initially is to provide patchec conics, almost identical to how KSP does it. However, it is possible that if the studio has the right talent (and a team member has the desire) for N-Body to be added as an option. Regardless, the game is being built so a modder could develop a C# mod and add this. Care is being taken to ensure the game is being structured so that if we can't add N-Body physics, someone else could add it.

JPLRepo has our patched conics currently implemented and we are refining that.

----

For now we are using patched conics. But we are making this so it could be patched out by modders to provide n-body. There is a small chance the studio could develop an n-body option internally as we are recruiting a PhD maths/physics who may be able to do that.

In terms of performance, that is really my key focus and the whole development of our BRUTAL framework. As we work directly with Vulkan and therefore your GPU, the scale is tremendous. But it does make working with it... rather brutal.

On getting something playable (not specified if public or not)

Quote

Hoping that we can have people poking around initially next year

On whether the orbital sim is moddable:

Quote

At the very least we are constructing the orbital simulation so someone can replace the orbital simulation. The studio also has been advertising for someone with a PhD in mathematics and physics. If we get that role filled, maybe n-body could be an option. Just for myself... my maths and physics is nowhere near good enough

On axial tilt:

Quote

Already in yep!

On the possibility of off-focus thrust:

Quote

Yes. We instantiate Jolt Physics SDK (the current front runner) for complex physics when a camera is near. Otherwise we use varying levels of simulation depending on what is required. This allows us to multithread it all.

Another one about operating craft off-focus, for simultaneous launches or landing boosters:

Quote

Yep that’s the idea of our approach. We can instantiate what is needed separately

On a bigger solar system and life support:

Quote

Very close to KSP in scale and feel, but we are running real solar system data for now for ease of testing.

Survival very much important, so life support I consider essential (but togglable). No details as yet, but I think quite detailed as much as the popular mods. We will make it easy for it to be replaced entirely by mods though.

On multiplayer:

Quote

Yes. Multiplayer will be in. It is our studios specialty in my opinion.

We curate our own approach (called RocketNet) which is already in use in the game Stationeers and allowed us to get tremendous amounts of data through very efficiently, although it is complex to work with. We are using that for KSA as it is built into BRUTAL. It is effectively RakNet.

The FAQ on the discord says its just a shared timeline that can be managed by the host or by vote if settings allow for vote.

On a native vs modded implementation of something like kOS:

Quote

kOS is one of my favorite mods. Absolutely essential mod for me. I would love us to include something similar BUT I also don't want to destroy the potential of mods. So even if we don't add our own language, or even if we do, there would be support for loading modders versions of whatever they want to achieve automation and programming of vessels.

On why the hell would they use XML and C#:

Quote

We use C++ or libraries for anything we need to. Where it makes sense we use C#, which it turns out is most places. Individual projects can use whatever data structures they like. for "human read" data I prefer to use XML personally, as it natively becomes C# classes and is easy to read when well formatted. For computer to computer I personally use JSON. For manifest files for mods we use TOML.

On when does Dean think modders can start touching his baby:

Quote

I see modders getting behind-the-scenes access late this year or early next to initial builds for feedback. Once we have confirmed basic data structures (these are a huge pain to modders to change), then we will do similar but with open access to everyone for free for initial versions.

I hope maybe mid next year people can try out where we are at.

Career + Science absolutely I see as critical (although toggable). Probably both together.

Another confirmation for life support, and some about the building of craft:

Quote

Lifesupport *absolutely*. In terms of building at a base level the parts will feel similar to KSP/KSP2. But I want to expand that making the parts from "subparts". This means we make fairings, small tanks, pipes, connections - that get assembled into a "part". Default will feel much the same as KSP/KSP2. But a super user could add multiple connections to a part, allowing it to "mate" with similar parts above it. This could, for example, allow you to run two separate resource lines across one part connection. Probably not doing a good job to explain it. This system is great for asset development but also (and especially) rendering, as we can draw those mesh parts as a single draw call no matter how many instances there are.

On why the screen looks so busy on the orbit sim video:

Quote

The orbit lines are a special line render technique we made. Not just for debugging, but for use in the game as well. You can turn them on and off very easy or just have certain bodies up. I left all the ones we have on so people could see the simulation running for I think approximately 50 bodies currently.

 

More on parts and the technical direction of the BRUTAL  framework:

Quote

Most of our design is very oriented towards technical aspects of wanting to achieve scale, seamlessness, performance, and asset pipeline (i.e. modability). The innovations I think are coolest are really all those, with our aim being to recreate KSP but with these in mind.

Parts are intended to be (but don't require) to be made out of smaller sub parts. This allows us to batch render them and makes creating stuff easier. It also means you could make custom parts, run multiple fuel line connectors, etc... Parts are then really just a collection of smaller parts. It means someone can just play the game more like traditional KSP and place blocks - or they can make their own standards and have multiple lines connecting different fuels. Hopefully I explained that well enough. Think about default blocks having a central core of fuel connectors, but you could make your own parts from the subparts that allow you to have two separate fuel lines in the same block that would connect if a fellow connection matches in the part above.

On some of the pillars for their design:

Quote

Our core aim is to:

• Make it seamless to move between screens to allow better multicraft missions. This means no loading screens at all

• Make the "kittens" matter, similar to progression given by "probes before crew" etc...

• Provide extremely moddable data structures, so that modders can add to the game even better than in KSP.

Off-world base building absolutely yes. That is why we are using Jolt physics API so we can instantiate it. We need to be somewhat restrained what we do with the base game because it has to work really well. If we provide a good framework, modders can take that much further.

 

 

[Bej Kerman]

9 hours ago, Fizzlebop Smith said:

Hey Bej has found a game he doesn't hate, let 'em have a moment to sniff the Hopium.  Its being shared here bc they don't have official forums.

Why do I detect a hint of malice...?

[kerbiloid]

On 11/2/2024 at 8:08 AM, Skorj said:

Goblin

Koblin. It's Kerbal world, after all.

Bold koblin = kobold.

Thin koblin = kothin.

***

About the kittens.

Crashing the kittened crafts or playing BDArmory would be a problem with them.

Maybe the kittens should be the KSC staff.

[PDCWolf]

27 minutes ago, kerbiloid said:

Crashing the kittened crafts or playing BDArmory would be a problem with them.

I think that's the point: They're not death-happy disposable green humanoids. You're supposed to take care of them.

It's a good change in tone I think, maybe one of the things they learned is that taking "lol so kerbal" too seriously can actually ruin the direction of your game. Plus judging by what I've played of Stationeers, Dean does like to take it somewhat seriously.

[kerbiloid]

2 hours ago, PDCWolf said:

I think that's the point: They're not death-happy disposable green humanoids. You're supposed to take care of them.

Yes, but what about the BDArmory?

Two kittened fighters, one of which is to be shot down.

[Bej Kerman]

7 minutes ago, kerbiloid said:

Yes, but what about the BDArmory?

Two kittened fighters, one of which is to be shot down.

Making the player feel bad sounds like the plan.

[Johnster_Space_Program]

Yeah, I just saw this, and the graphics definitely look impressive so far! Reminds me a lot of what Space Engine looks like in regards to the graphics quality.

I have a new PC build I'm working on with a Ryzen 7 7700 CPU and a RTX 3060 GPU for gaming and my game development projects, so once there's a playable version out, I'll definitely test it out to see how well it runs on my build with low, medium, high settings, etc, compared to, say, KSP 1 and 2.

Also, I found out they have a Discord server you can join here (https://discord.gg/kittenspaceagency), and they also posted this development video there (https://www.youtube.com/watch?v=PFZGFBjE3v8), which appears to be the same one you posted in the OP, but a YouTube video version.

I actually first heard about this game a few days ago from ShadowZone (https://www.youtube.com/post/UgkxSzCWeZA_5UvtD6_fUWk5-fq3Wy98AA3d), who I'm sure some of you might remember for his KSP 2 development history videos. I really hope he'll do similar videos to that about this game's development, and in the community post he did, he already said he's going to do a video on the game, so hopefully more will follow after that as more is shared regarding KSA's development.

Overall, I'm excited to see how this game will turn out, and I'm confident based on what the developers have shared that it won't have the same fate as Kerbal Space Program 2. I'm also excited about the modding capabilities the game will offer (since in addition to being a game developer myself, I also like creating mods like I've done for KSP 1 and 2), and I hope there will be a dedicated site or page (like on SpaceDock) where players like myself can easily share mods for the game.

Although if an in-game mod browser is going to be included, maybe that won't be necessary. But it will be interesting to see where this game goes regardless of how the modding support is integrated, and the multiplayer will be nice to check out as well.

Edited Sunday at 04:50 PM by Johnster_Space_Program

[AVaughan]

1 hour ago, kerbiloid said:

Two kittened fighters

Kzin

[MechBFP]

7 hours ago, Bej Kerman said:

Why do I detect a hint of malice...?

Reminds me of this:
Buzz: You're mocking me, aren't you?

Alisha: Yeah, but in a supportive way!

lol

Anyways, good luck to everyone involved, I know there is gonna be ALOT of salt to wade through from a community perspective so don’t let that get you down. 

Edited Sunday at 06:55 PM by MechBFP

[PDCWolf]

39 minutes ago, MechBFP said:

Reminds me of this:
Buzz: You're mocking me, aren't you?

Alisha: Yeah, but in a supportive way!

lol

Anyways, good luck to everyone involved, I know there is gonna be ALOT of salt to wade through from a community perspective so don’t let that get you down. 

Since I've been quoting Dean, I'll quote him some more:

Q: is this real?

Quote

Hah wait and see till you try it out! Good to keep being skeptical and critically review things, give us solid feedback when we have that first version out

Q: whenever I see a project like this, I'm always very skeptical.

Quote

I think it is critical for people to be skeptical. It is a complex project. And the community play an important role in us getting this right.

I think we're gonna see a very noticeable difference between how they conduct themselves towards a critical community vs what the "professional studio" did, at least in the handling feedback department. The mature thing is understanding salt is feedback, and that shielding developers from criticism is bad for everyone, even the game itself. That's why the KSP1 devs ran to 4chan for, to get actual criticism rather than a hugbox madhouse, you can't grow in those.

3 hours ago, kerbiloid said:

Yes, but what about the BDArmory?

Two kittened fighters, one of which is to be shot down.

Just like in real life, it's a loss, it's probably a worthy life being lost. You should be attached, you should be invested, it's what makes you try to be better at building safer and more complete craft. Still, depending on how they handle death, it might not even matter.

[LunarMetis]

I like what I see so far. Will be following this thread closely for updates. 

[Bej Kerman]

14 minutes ago, PDCWolf said:

I think we're gonna see a very noticeable difference between how they conduct themselves towards a critical community vs what the "professional studio" did, at least in the handling feedback department. The mature thing is understanding salt is feedback, and that shielding developers from criticism is bad for everyone, even the game itself. That's why the KSP1 devs ran to 4chan for, to get actual criticism rather than a hugbox madhouse, you can't grow in those.

I think what MechBFP means by "salt" isn't people who have valid criticisms and points of skepticism, it's people with nuclear grade trust issues who think others shouldn't be allowed to watch this project with anticipation, just because they were hurt by a triple-A title in the same genre.

[kerbiloid]

10 hours ago, AVaughan said:

Kzin

Kzin vs Kilrathi.

P.S.
Interstring fact: none of these sounds could be pronounced by cats.

Maybe "r" only.

[James Kerman]

Some content has been removed. Please avoid personal attacks when commenting as this adds nothing to the discussion besides bad feelings.

[BA-Forums]

Just saw this, who here is gonna create the Kittens to Kerbals mod?

[Empiro]

On 11/1/2024 at 8:36 PM, Skorj said:

Oh, I have.  Even complex games are "mid" compared to large business projects.  But I do wonder what they mean when they say they created their own engine.  These days studios usually mean "a custom framework on top of Unreal", since re-inventing that particular wheel would be ... a bold plan.  C# would be a great choice for a scalable framework built of top of letting Unreal do the crunchy bits, but they seem to be saying they started from scratch?  I'm less hopeful for this project if that's the case.  I've only ever seen studios crater after thinking they should crate a new engine (anyone remember Flagship?).

Still, first time for everything.  I'm cheering for them, either way.

I think the issue is that both Unreal and Unity make certain assumptions about the game: namely at its base it's a scene with some sort of geometry and entities that move and interact within the scene. This works great for 99% of all games but not KSP/KSA because of the scales involved. I think KSP's challenges with Unity wouldn't go away at all if you used Unreal. Somethings might work better, but the core incompatibility will still be there.

At the same time, I think the nature of KSA is that you don't actually need an extremely complex engine. At its core, it's a physics simulation, and for the small stuff, its sounds like they're already using a dedicated physics library, while for the big stuff, it's all Newtonian mechanics. (That's not to say it's easy, but it's at least known, and it's not like any game engine supports Newtonian mechanics for celestial bodies out of the box anyway)

On 11/3/2024 at 5:12 AM, PDCWolf said:

I think that's the point: They're not death-happy disposable green humanoids. You're supposed to take care of them.

It's a good change in tone I think, maybe one of the things they learned is that taking "lol so kerbal" too seriously can actually ruin the direction of your game. Plus judging by what I've played of Stationeers, Dean does like to take it somewhat seriously.

For the default difficult, I don't think anyone would mind if the cats just came back after getting blown up. After all, they do have 9 lives.

[theJesuit]

1 hour ago, Empiro said:

For the default difficult, I don't think anyone would mind if the cats just came back after getting blown up. After all, they do have 9 lives.

I, for one, would like to welcome our new kitty overlords.

I have high hopes.  The right peoplem the lack of hypethe getting fundamentals right first up.

With IVA, EVA stuff i hope that the characters are blue (yellow is taken by a certain movie franchise, and green invites lawsuits.  A different body shape.  Maybe with fur still? 

[magnemoe]

1 hour ago, Empiro said:

I think the issue is that both Unreal and Unity make certain assumptions about the game: namely at its base it's a scene with some sort of geometry and entities that move and interact within the scene. This works great for 99% of all games but not KSP/KSA because of the scales involved. I think KSP's challenges with Unity wouldn't go away at all if you used Unreal. Somethings might work better, but the core incompatibility will still be there.

At the same time, I think the nature of KSA is that you don't actually need an extremely complex engine. At its core, it's a physics simulation, and for the small stuff, its sounds like they're already using a dedicated physics library, while for the big stuff, it's all Newtonian mechanics. (That's not to say it's easy, but it's at least known, and it's not like any game engine supports Newtonian mechanics for celestial bodies out of the box anyway)

For the default difficult, I don't think anyone would mind if the cats just came back after getting blown up. After all, they do have 9 lives.

Now you will run into running rovers on surface and base building who can include landing an base module with mining and ISRU on Laythe and a space plane to explore and getting back to orbit. 
Or you can go a bit larger 

One of my Minmus bases featuring extra planetary launchpad. From right, an future base to say Duna. Extra planetary launchpad main hub, two habitation and food production modules. Small utility lander for use around Minmus, Fuel production. An node for making smaller parts. Resource storage and nuclear reactor and material processing. 

[PDCWolf]

2 hours ago, Empiro said:

For the default difficult, I don't think anyone would mind if the cats just came back after getting blown up. After all, they do have 9 lives.

Nah, I think it's important to set the tone right, and if that tone is that this is a game with consequences and that you have to take seriously, I couldn't be happier. It's one of the biggest holes in KSP1, where not taking itself seriously suddenly infected everything so much that serious, deep mechanics pretty much fell out of scope for a "silly green men exploding simulator".

Edited 12 hours ago by PDCWolf

[HebaruSan]

I guess I've been on record as a pessimist since May. I'll need to see overwhelmingly positive reviews to get past my experiences attempting to play this studio's previous work.

[TheSaint]

2 hours ago, HebaruSan said:

I guess I've been on record as a pessimist since May. I'll need to see overwhelmingly positive reviews to get past my experiences attempting to play this studio's previous work.

Everything is vaporware until, well, it isn't vaporware anymore. I thought that was understood by now?