KSP2 EA Grand Discussion Thread.

[pandaman]

Honestly  I think the 'Should have used a different  engine to Unity' argument is very much a 'Grass looks greener over there' issue.

A different engine would be just that 'different' it may not have the same issues, but it would have some unique ones of its own.  The end result would likely be just as problematic, but just with different problems.  Then the argument would be 'Why didn't  they just stick with Unity,  (engine x) is rubbish.

[Periple]

1 minute ago, Manul said:

Commercial developers do have a source access to the engine, that's why using Unity doesn's necessarily mean using PhysX as is. And without major changes to the physics engine the ultimate goal to slay the Kraken is unachievable.

Only on the Enterprise plan and that's stinky expensive... and even then you don't get to change it, just look at it. You have to pay a LOT more to be allowed to modify it, and of course that comes with its own set of problems -- in particular it makes engine upgrades very painful.

(Edit: Doesn't mean you have to use PhysX though, you don't need to go into the engine to change that. Nice thing about Unity is that you can do a LOT without having to touch the engine. Almost everything really.)

Edited March 2, 2023 by Periple

[MechBFP]

3 minutes ago, pandaman said:

Honestly  I think the 'Should have used a different  engine to Unity' argument is very much a 'Grass looks greener over there' issue.

A different engine would be just that 'different' it may not have the same issues, but it would have some unique ones of its own.  The end result would likely be just as problematic, but just with different problems.  Then the argument would be 'Why didn't  they just stick with Unity,  (engine x) is rubbish.

Yup 100% this. Everything sounds great on paper until it is implemented. At least in this case they are dealing with similar issues instead of brand new ones they would be clueless on. 

[SimonTheSkink]

This is just my two cents here, I think the main reason the game is in the state that it's in after so many years is because the team developing the game was focusing on creating everything at once instead of the basics and working from there. For example, in my opinion the best way to develop a game like this would be to create the base physics engine and work out all the bugs with that first. Make the game as bare bones as possible and focus purely on making it run flawlessly. After that has been done you can start expanding from there and adding more content. Here it feels like they were trying to focus on everything at once, ran out of time, and the people higher up said we need a confirmed release by X-date. So the team shelved everything they were working on and focused purely on the core game, didn't have enough time and we got what we see now.

I think we should be ok going forward, but they really should have focused on the core game first instead of everything at once.

[Manul]

5 minutes ago, Periple said:

optimize the assets here, replace a stinky algorithm there, tune this bunch of parameters until they feel right

PhysX is designed for shooters and arcades, it's not designed for physical simulation of complicated structures . There is no external fix to turn inaccurate calculations into accurate.

[Klapaucius]

1 hour ago, hansen said:

Yet, in KSP 1 there was never a real purpose to do anything (although this is maybe an interesting comment on real-life spaceflight ;D). You can land on planetary bodies or put stuff in orbit, but there is almost nothing to explore, there is no purpose in bringing stuff and kerbals there.

I've never needed a purpose. I just created my own.  To me that is what a sandbox is all about--finding challenges, making up reasons/stories and expressing them through the game.  The fascination is the problem solving and what you create in your mind to tie it all together.

It seems like you are wanting KSP to be something like Subnautica (a truly great game), but I don't think that's what it is all about.

[Periple]

11 minutes ago, Manul said:

PhysX is designed for shooters and arcades, it's not designed for physical simulation of complicated structures . There is no external fix to turn inaccurate calculations into accurate.

Sure but the kind of physics KSP2 uses PhysX for don't have to be accurate, just close enough to be fun and you can certainly get them there. 

The orbital mechanics do have to be accurate, but it's certainly possible to firewall that off from the bouncy inaccurate physics, so that only engine impulse gets carried over (plus some pretty, random, believable impulses for when things go kaboom). And you'd probably need to write a bespoke system for wheels instead of making them all physics-based. You can 100% do all of that in Unity without messing with it at the engine level.

Edited March 2, 2023 by Periple

[RayneCloud]

PhysX is perfectly fine for what it's used for and what is was designed for, which is handling rigid body physics, fluid dynamics, and more. 

[AlamoVampire]

@pandaman then why was the mod dark multiplayer elevated to feature and mechjeb wasnt? Both camps were equally vocal/loud and yet here we are. I want desperately to say more but ive said all i can over the years. 
 

132203022023

[Bej Kerman]

3 minutes ago, AlamoVampire said:

then why was the mod dark multiplayer elevated to feature and mechjeb wasnt?

Just because the devs thought it'd be appropriate to carry over flying by the seat of your pants, probably.

[Periple]

Just now, AlamoVampire said:

@pandaman then why was the mod dark multiplayer elevated to feature and mechjeb wasnt? Both camps were equally vocal/loud and yet here we are. I want desperately to say more but ive said all i can over the years. 

Maybe they thought that MechJeb doesn't fit the core gameplay design? KSP1 basically has two core gameplay activities: building craft and flying craft. MechJeb makes one of them optional. That's bound to have huuuuge implications for the gameplay design everywhere, and by the time you've got all of that figured out you might have landed somewhere that doesn't bear much resemblance to KSP at all. Might still be a cool game though!

Multiplayer OTOH fits the gameplay design really well -- you're still doing the same core activities, except now you're doing them together with friends. 

Also multiplayer is a MAJOR selling point. There's a huge public of gamers out there who play multiplayer exclusively or near-exclusively, KSP2 would be a very hard sell to them if it didn't have multiplayer. 

I.e. from where I'm at MP is kind of a no-brainer, but MechJeb would be all kinds of complicated.

[RocketRockington]

35 minutes ago, SimonTheSkink said:

This is just my two cents here, I think the main reason the game is in the state that it's in after so many years is because the team developing the game was focusing on creating everything at once instead of the basics and working from there. For example, in my opinion the best way to develop a game like this would be to create the base physics engine and work out all the bugs with that first. Make the game as bare bones as possible and focus purely on making it run flawlessly. After that has been done you can start expanding from there and adding more content. Here it feels like they were trying to focus on everything at once, ran out of time, and the people higher up said we need a confirmed release by X-date. So the team shelved everything they were working on and focused purely on the core game, didn't have enough time and we got what we see now.

I think we should be ok going forward, but they really should have focused on the core game first instead of everything at once.

Based on the fact they they only had a physics programmers at Intercept from 2021 to 2022, I think maybe even they didn't dmtry to do everything at once, they assumed they could copy KSP1 physics code without a dedicated programmers and focused on other things that we saw...for some reason they worked on the VAB and UI very early?  And the Kerbal animation?  Wish we knew what was really going on, in this case I want to know how the sausage was made.

[Manul]

12 minutes ago, MechBFP said:

At least in this case they are dealing with similar issues instead of brand new ones they would be clueless on. 

Fixing these issues in KSP 1 required the whole game to be remade from ground up. That's how KSP 2 was born. And now when everything was made from the ground up the same problems are back.  It's a really bad sign meaning that the problem is much deeper and unlikely to be fixable without starting over again.  So we'll just have to live with large crafts exploding for no reason, distant crafts sinking into the ground, some parts exceeding lightspeed  because of some near infinite force was applied due to floating point calculations error.

[Monger]

This is one of the topics where you can ask 10 people and get 15 opinions. 

For me, what made KSP so great was this perfect balance between light-hearted goofiness and nail-biting complexity about a topic that is so far from intuition as possible. It opened my mind, and made laugh at the same time.

Can any game repeat that for me? I doubt it. Could a game repeat this for someone else? Possibly.

KSP 2 needs its own message, its own audience.  I wondered for years what this might be. I hope it's something along a positive futurism, a small glimpse into mankind's future. But that's just me. 

[Periple]

3 minutes ago, Manul said:

So we'll just have to live with large crafts exploding for no reason, distant crafts sinking into the ground, some parts exceeding lightspeed  because of some near infinite force was applied due to floating point calculations error.

They will fix all that. 

Eventually.

[RocketRockington]

16 minutes ago, RayneCloud said:

PhysX is perfectly fine for what it's used for and what is was designed for, which is handling rigid body physics, fluid dynamics, and more. 

PhysX doesn't do fluid dynamics.  If you're talking about the Flow SDK that's a fancy particle engine that does some gas-like stuff for a narrow range of applications (though that narrow range is what a lot of game devs want - people like smoke and fire)

And it's rigid body dynamics are best for the subset of requirements that doesn't involve mechanical joint linkages.  KSP could really use a better rigid body silver, especially when it comes time to do robotics.

[MarcAbaddon]

I hope you are right. I personally don't believe it. The game is likely too far along to rework the physics engine to any advanced degree. With some degree of luck we'll get a similar amount of reining in the physics bubs as we did towards the end of KSP. 

[MarcAbaddon]

I think focussing on multiplayer over auto-pilot systems is a good call. MechJeb was and can be fitted added by modders (no way they are not doing to expose the necessary functionality), but multiplayer needs to be integrated very tightly. Personally, I'd wish they would have put it earlier in the Roadmap, not since it's the feature I want the most, but I think it's the hardest one to add late.

[RayneCloud]

6 minutes ago, RocketRockington said:

PhysX doesn't do fluid dynamics.  If you're talking about the Flow SDK that's a fancy particle engine that does some gas-like stuff for a narrow range of applications (though that narrow range is what a lot of game devs want - people like smoke and fire)

And it's rigid body dynamics are best for the subset of requirements that doesn't involve mechanical joint linkages.  KSP could really use a better rigid body silver, especially when it comes time to do robotics.

 

Hmmm.. if you say so.

[Kelcie]

Paying full price for barely half a game doesn't make sense. 
Especially for a company that didn't need the startup cash to make the game.

KSP is hands down the best game I've ever played. 

I've waited years for number 2. 

Vowed to not buy any new games for years until KSP2 came out.

But I won't be buying KSP2 anytime soon.  Maybe ever if it stays a flaming pile o' crap.

Also a game dev,
~Kelcie

[Periple]

6 minutes ago, MarcAbaddon said:

I hope you are right. I personally don't believe it. The game is likely too far along to rework the physics engine to any advanced degree. With some degree of luck we'll get a similar amount of reining in the physics bubs as we did towards the end of KSP. 

They don't need to rework the physics engine much, just isolate it from the orbital mechanics. And maybe implement a separate wheel system. The rest is mostly just tuning parameters.

[Bej Kerman]

4 minutes ago, MarcAbaddon said:

I think focussing on multiplayer over auto-pilot systems is a good call. MechJeb was and can be fitted added by modders (no way they are not doing to expose the necessary functionality), but multiplayer needs to be integrated very tightly. Personally, I'd wish they would have put it earlier in the Roadmap, not since it's the feature I want the most, but I think it's the hardest one to add late.

The devs are playing multiplayer behind closed doors and dataminers have found roadmap features in good state.

[Periple]

3 minutes ago, MarcAbaddon said:

I think it's the hardest one to add late.

It's almost impossible to add late! All the systems have to be designed with multiplayer in mind from the start.

Intercept have said though that that's what they've done and they have working multiplayer in an internal build. It is bound to have scads of bugs but those can be addressed when the rest of the game has settled down if the fundamentals are there. 

[MarcAbaddon]

Just now, Periple said:

They don't need to rework the physics engine much, just isolate it from the orbital mechanics. And maybe implement a separate wheel system. The rest is mostly just tuning parameters.

What do you mean by isolating it from the orbital mechanics? The orbital mechanics are already isolated from the current physics engine.

It's really difficult to get everything working as you would like - there  is a reason why original KSP 1 or the modders never managed to fully tame the physics engine.

 

[MechBFP]

16 minutes ago, Manul said:

Fixing these issues in KSP 1 required the whole game to be remade from ground up. That's how KSP 2 was born. And now when everything was made from the ground up the same problems are back.  It's a really bad sign meaning that the problem is much deeper and unlikely to be fixable without starting over again.  So we'll just have to live with large crafts exploding for no reason, distant crafts sinking into the ground, some parts exceeding lightspeed  because of some near infinite force was applied due to floating point calculations error.

It is certainly a possibility that will be the case. Only time will tell how much of the jank will remain.