A Twitter thread on terrain rendering performance.

[MarcAbaddon]

On 3/1/2023 at 2:07 AM, Chilkoot said:

There is strong evidence to the contrary.  See some discussion here: https://old.reddit.com/r/KerbalSpaceProgram/comments/11b8s6f/i_looked_into_ksp2_code_here_is_what_ive_found/j9wm3dd/

One of the guys that went through the code with DotPeek (see above link) noted that the rigid-body physics engine in KSP 2 appears to be custom from the ground up, and it does not use the Unity physics engine (which KSP 1 did use).  I haven't looked myself, just repeating his conclusions.

There's another post (with admittedly far less upvotes) which states that the custom part is mostly only a wrapper for the Unity engine rigidbody. I am inclined to believe this since the it just behaves so similarly to KSP 1 that it is hard to believe that they wrote a custom implementation which just reproduces the same issues, but I think I may take a look at the code myself this weekend. 

[Tazooka]

2 hours ago, MarcAbaddon said:

There's another post (with admittedly far less upvotes) which states that the custom part is mostly only a wrapper for the Unity engine rigidbody. I am inclined to believe this since the it just behaves so similarly to KSP 1 that it is hard to believe that they wrote a custom implementation which just reproduces the same issues, but I think I may take a look at the code myself this weekend. 

I saw this post. Both this one and the one OP mentioned are interesting to read

[Chilkoot]

3 hours ago, MarcAbaddon said:

There's another post (with admittedly far less upvotes) which states that the custom part is mostly only a wrapper for the Unity engine rigidbody. I am inclined to believe this since the it just behaves so similarly to KSP 1 that it is hard to believe that they wrote a custom implementation which just reproduces the same issues, but I think I may take a look at the code myself this weekend. 

Interesting - I know they did hire some physics modeler guy to design a rigid body system from the ground up, but (like everyone) I was surprised to see the noodle rockets back and worse than ever.  Maybe it's a work in progress, or maybe the oddities are a "convergent evolution" of physics modeling.  Either way, it's a pretty big issue right now - Matt Lowe's frustration during launches is palpable in his latest video.

[Mutex]

Maybe the ground-up redesigned physics system isn't ready, so they're using the default one as a placeholder.

[LameLefty]

6 minutes ago, Chilkoot said:

Matt Lowe's frustration during launches is palpable in his latest video.

Yeah but … his rover rocket video last night was infuriatingly BAD. He has a massive weight inside a giant faring cantilevered off a long and spindly rocket. I’m an aerospace engineer by education and early career - that kind of rocket has NEVER worked well in KSP until recent years’ patches that artificially amplify the strength of joints far beyond what a real rocket structure would tolerate. No a real rocket wouldn’t all apart at the staging joints, but it WOULD crumple like a long skinny tin can, and then blow apart in a conflagration as the propellants escaped the pressurized structure and mixed uncontrollably.

Had Matt spent less time fighting with his long skinny heavy-nosed rockets and built something shorter with side-mounted boosters, he’d have had much better luck getting to orbit.   

[MarcAbaddon]

It was a 60m rocket with a 3.75m diameter. While that's a bit more spindly than I usually build in KSP, that's not a totally unusual ratio if you look at real rockets. Most of the wobbling is at the fairing point anyway, I don't think making it shorter would have helped that much.

[LameLefty]

8 minutes ago, MarcAbaddon said:

It was a 60m rocket with a 3.75m diameter. While that's a bit more spindly than I usually build in KSP, that's not a totally unusual ratio if you look at real rockets. Most of the wobbling is at the fairing point anyway, I don't think making it shorter would have helped that much.

It’s on par with a real-world Falcon 9, but … F9 has one of the worst “fineness ratio” (lenght/diameter) of any operational rocket ever designed; F9 requires pressure stabilization of its structure and will collapse under load if tank pressure control fails; and F9 as a MUCH smaller fairing that Matt’s rocket. That rocket also had a heavy/complicated rover inside mounted horizontally IIRC, and I don’t think any rocket with that much mass hanging off the end inside that large a fairing would ever do well in an atmosphere. 

[Kerbart]

On 2/28/2023 at 2:36 PM, Bej Kerman said:

So their decision making is worse than I suspected...

In what universe would having to do hand calculations benefit the gameplay?

In a universe where the creator of the game decrees that "flying by the seat of your pants" is the intended way to play the game. Which was fine pre 0.13 or thereabouts, but they held on to that for way too long.

[tstein]

35 minutes ago, Kerbart said:

In a universe where the creator of the game decrees that "flying by the seat of your pants" is the intended way to play the game. Which was fine pre 0.13 or thereabouts, but they held on to that for way too long.

I can cite other games, Even Microsoft flight simulator you had to make fuel consumption calculations by HAND on the first few versions. In falcon 4 you had to make the bulls eye translation fo coordinates in your head...   

 

that is not so uncommon in SIMULATION.

[Guest]

On 2/28/2023 at 11:36 AM, Bej Kerman said:

So their decision making is worse than I suspected...

Wasn't helped by the fact that a certain segment of the fanbase took it as a point of pride to do those calculations by hand. "Oh, I just alt-tab to my Excel spreadsheet" as if that wasn't already using a convenience method. I had to get porkchop plots from a web app for the longest time, and even then had to eyeball ejection angles which meant frequently getting it wrong (can't remember when they added the ability to drag the maneuver along the orbit, but that was a godsend).

Edited March 2, 2023 by regex

[MarcAbaddon]

I do agree that KSP 1 should have had it earlier and I think a lot of early (and even late) KSP 1 design decisions sucked. This being said, I feel that going all 'look how bad parts of KSP 1 used to be' isn't really a defense for KSP 2.

The KSP 2 devs were able to see the end of a long road for KSP 1. Learning from this that KSP 2 should have a delta-v calculator  is really the absolute minimum and it's not a sign their decision process is better, just that they had the benefit of hindsight. It's a bit like saying modern fps makers have better decision processes because they are using WASD for moving your character compared to some of the first FPS games out there which still used the arrow keys.

In fact, I am pretty disappointed we are missing clock alarms, TWR per stage, atmospheric dV calculations and especially a transfer window calculator.

Edited March 2, 2023 by MarcAbaddon

[Guest]

6 minutes ago, MarcAbaddon said:

It's a bit like saying modern fps makers have better decision processes because they are using WASD for moving your character compared to some of the first FPS games out there which still used the arrow keys.

lolwut

Squad were told for literally years that the game needed a delta-V calculator. Years. That doesn't require hindsight to do the "bare minimum", it requires listening to your fanbase and adhering to the conventions of the genre (space flight). It's a no-brainer.

[Chilkoot]

1 hour ago, tstein said:

Even Microsoft flight simulator you had to make fuel consumption calculations by HAND on the first few versions.

OK, let's be fair here - that feature was added in the 80's (like 35 years ago) and the game was running on the first 16-bit processor (80286) at the time  

48 minutes ago, regex said:

Wasn't helped by the fact that a certain segment of the fanbase took it as a point of pride to do those calculations by hand. "Oh, I just alt-tab to my Excel spreadsheet" as if that wasn't already using a convenience method. I had to get porkchop plots from a web app for the longest time, and even then had to eyeball ejection angles which meant frequently getting it wrong (can't remember when they added the ability to drag the maneuver along the orbit, but that was a godsend).

There was a point where you could grab some output parameters from a web-based porkchop calculator, and paste them into - I think - "Precise Maneuver" or one of the other early maneuver addons.  That was like heaven lol - eyeballing died for me that day and I never looked back!

Edited March 2, 2023 by Chilkoot

[tstein]

5 minutes ago, Chilkoot said:

OK, let's be fair here - that feature was added in the 80's (like 35 years ago) and the game was running on the first 16-bit processor (80286) at the time  

 

Yes, I am that an old gamer

[SimonTheSkink]

Honestly reading this gives me a big sigh of relief. If what I am reading is correct, the performance of the game should improve rather quickly with some proper coding.

[Bej Kerman]

2 hours ago, Kerbart said:

In a universe where the creator of the game decrees that "flying by the seat of your pants" is the intended way to play the game. Which was fine pre 0.13 or thereabouts, but they held on to that for way too long.

I have a hard time believing Harvester seriously didn't see the problem with telling newbies "here's a game that'll teach you space travel in a really simple way, now do the calculations yourself!". Did he not have a clue who he wanted to aim KSP 1 at? Is the first thing they teach in GCSE Media not to try split yourself between two audiences unless you have a really, really good reason to? Again, before that defense pops up again, just because Squad is a small indie group , doesn't mean each developer there should have been completely devoid of the ability to make decisions that make sense. Scattering important readouts to different parts of the screen to make quicktime flying difficult, hiding the alt-F5 and alt-F9 functions until they tweaked the UI flow for 0.21, etc.

1 hour ago, regex said:

lolwut

Squad were told for literally years that the game needed a delta-V calculator. Years. That doesn't require hindsight to do the "bare minimum", it requires listening to your fanbase and adhering to the conventions of the genre (space flight). It's a no-brainer.

For such a fan-central product where mods and Scott Manley are the things that kept it afloat for long enough to become significant, Squad really didn't want to take their hands off their ears in regards to this 'decision' 

[Kerbart]

54 minutes ago, Bej Kerman said:

I have a hard time believing Harvester seriously didn't see the problem with telling newbies "here's a game that'll teach you space travel in a really simple way, now do the calculations yourself!". Did he not have a clue who he wanted to aim KSP 1 at? Is the first thing they teach in GCSE Media not to try split yourself between two audiences unless you have a really, really good reason to? (…)

The organic way KSP has evolved led to this. There are a lot of elements in the game that never came out of an overarching vision, but just grew that way.

It’s something a lot of the why did development take so long complaints overlook and I suspect Nate was the first one to revisit those foundations  and define them. If that led to setting the timeline back one year so be it, it was needed.

[Guest]

2 minutes ago, Kerbart said:

The organic way KSP has evolved led to this. There are a lot of elements in the game that never came out of an overarching vision, but just grew that way.

I'm still convinced that they never had a roadmap of any sort, or even tried to define one until post 1.0.

[Bej Kerman]

10 minutes ago, Kerbart said:

1 hour ago, Bej Kerman said:

I have a hard time believing Harvester seriously didn't see the problem with telling newbies "here's a game that'll teach you space travel in a really simple way, now do the calculations yourself!". Did he not have a clue who he wanted to aim KSP 1 at? Is the first thing they teach in GCSE Media not to try split yourself between two audiences unless you have a really, really good reason to? (…)

The organic way KSP has evolved led to this. There are a lot of elements in the game that never came out of an overarching vision, but just grew that way.

I don't see how an organic development would inhibit Squad from making decisions so basic like "should we aim the game at newbies or make players handcalculate things?"

Edited March 2, 2023 by Bej Kerman