Assumption based Concerns regarding the Unity Engine or How I Learned To Worry Myself To Death and Fear Things I Do Not Understand

[ArtemisAZ]

Given the sheer scope of KSP2 and all the systems they're going to be adding, would the Unity engine hold the game back in some way? It's a great engine don't get me wrong but I feel like compared to other contemporary engines available Unity just seems like an odd choice and that it lacks the "Under The Hood" capabilities of others. Of course my knowledge on the actual deeper workings of these engines are as non-existent at best so I'm probably making a huge generalization. I just don't want to go another sleepless week worrying about KSP2 not living up my level of internalized Hyper-Hype because of engine issues. 

[GoldForest]

5 minutes ago, Krulliam said:

Given the sheer scope of KSP2 and all the systems they're going to be adding, would the Unity engine hold the game back in some way? It's a great engine don't get me wrong but I feel like compared to other contemporary engines available Unity just seems like an odd choice and that it lacks the "Under The Hood" capabilities of others. Of course my knowledge on the actual deeper workings of these engines are as non-existent at best so I'm probably making a huge generalization. I just don't want to go another sleepless week worrying about KSP2 not living up my level of internalized Hyper-Hype because of engine issues. 

Unity is actually very powerful if used correctly. What holds KSP 1 back is the poor coding done under the hood, and the use of old API's like DX9 which modern machines don't really support any more. Yes, DX9 is still compatible with DX11 and DX12, but compatible and supported are two different monsters. 

KSP 2 will run probably at least 3 times better than KSP 1, and up to 10 times better if everything works well. 

There are other topics on the KSP 2 forum which talk about Unity and KSP 2. I suggest searching around and looking at the arguments that have already been made. 

To prove how powerful Unity is, look at this tech demo. Everything was done in Unity. 

 

Spoiler

 

 

Edited August 25, 2019 by GoldForest

[Guest]

All it proves is that it supports pretty graphics. Under the hood, it's a low-budget piece of junk it always was. It works when you're not trying to do anything unusual with it (like modeling an entire solar system). Yes, sooner or later Unity will likely hold the game back. It'd be much better if they took time and money to write their own engine, or at least used something more advanced.

The only positive thing about it is that modders will have an easier time transitioning. Converting their models and workflows to a new engine would be harder than going between Unity versions.

[ArtemisAZ]

2 minutes ago, Dragon01 said:

All it proves is that it supports pretty graphics. Under the hood, it's a low-budget piece of junk it always was. It works when you're not trying to do anything unusual with it (like modeling an entire solar system). Yes, sooner or later Unity will likely hold the game back. It'd be much better if they took time and money to write their own engine, or at least used something more advanced.

The only positive thing about it is that modders will have an easier time transitioning. Converting their models and workflows to a new engine would be harder than going between Unity versions.

my steady sense of unease is now becoming a significantly more apparent sense of unease

[Yakuzi]

5 minutes ago, Krulliam said:

Given the sheer scope of KSP2 and all the systems they're going to be adding, would the Unity engine hold the game back in some way? It's a great engine don't get me wrong but I feel like compared to other contemporary engines available Unity just seems like an odd choice and that it lacks the "Under The Hood" capabilities of others. Of course my knowledge on the actual deeper workings of these engines are as non-existent at best so I'm probably making a huge generalization. I just don't want to go another sleepless week worrying about KSP2 not living up my level of internalized Hyper-Hype because of engine issues. 

I share your concerns. In the end, I would imagine that that the producer would prefer a short return of investment of the intellectual property they bought, so building upon the same framework as the original with a little bit more features and better aesthetics is the quickest time to market. Will that ultimately lead to substantially better gameplay than building a dedicated game engine from the ground up? I'm skeptical, but love to be proven wrong.

In the meantime, I'm looking forward to the plenitude of unfounded assumptions about this matter and everything else concerning KSP2.

Love the title btw!

[GoldForest]

11 minutes ago, Dragon01 said:

All it proves is that it supports pretty graphics. Under the hood, it's a low-budget piece of junk it always was. It works when you're not trying to do anything unusual with it (like modeling an entire solar system). Yes, sooner or later Unity will likely hold the game back. It'd be much better if they took time and money to write their own engine, or at least used something more advanced.

The only positive thing about it is that modders will have an easier time transitioning. Converting their models and workflows to a new engine would be harder than going between Unity versions.

Unity is very a powerful when used correctly. It's not a piece of junk. Have you seen some of the games to come out of Unity? Unity is powerful. It's the coders who are the problem, that and your machine. 

7 minutes ago, Krulliam said:

my steady sense of unease is now becoming a significantly more apparent sense of unease

Like I said, there's no need to worry. 

[radonek]

Limit factor for KSP engine is not fancy graphics, but physics. Working with multiple moving reference frames aint easy, that's why "better" games/engines tend to have some kind of global speed limit. Only exception I know of is the new Elite. It has _two_ reference frames, one of which have rather severe speed limit and other lacks physical interactions completely.  Working around this will be major thing in whatever engine squad may choose for KSP2.

[nikokespprfan]

so physics breaks without a speed limit, nothing unrealistic there

 

(he says as he realises how much work it is to put that into practice)

[sgt_flyer]

Havok demoed a unity plugin at GDC this year for their physics engine

There's 5000 entities dropped into a planet gravity well at around 18 mn mark - ksp would be reduced to a crawl with that   - i hope the devs will at least test that plugin  

 

Edited August 25, 2019 by sgt_flyer
Time

[Kerbart]

18 minutes ago, sgt_flyer said:

Havok demoed a unity plugin at GDC this year for their physics engine

There's 5000 entities dropped into a planet gravity well at around 18 mn mark - ksp would be reduced to a crawl with that   - i hope the devs will at least test that plugin

Very nice. However, the devs of KSP2, in interviews, very clearly state that the main issues is rigid bodies, and although the newer version of Unity that is used performs better, they're working on novel approaches to boast performance even further. The demo at 18 minutes is not a single rigid body though;it's 5000 point masses. Still impressive, and nice if it works that way... but that alone will do nothing to prevent a 500 part ship to be sluggish, unless they're covering rigid body physics as well.