KSP2 EA Grand Discussion Thread.

[JoeSchmuckatelli]

Any 1:1 comparison of KSP development with KSP2 development at comparable dates after release is inapt on its face. 

Novel concept vs recreation. 

 

[Spicat]

We already knew his opinion 4 days before release (Like advising to wait in the conclusion):

 

[Wheehaw Kerman]

31 minutes ago, JoeSchmuckatelli said:

Any 1:1 comparison of KSP development with KSP2 development at comparable dates after release is inapt on its face. 

Novel concept vs recreation. 

 

Sure, but a rewritten from the ground up and far more ambitious recreation, and the comparison really shows how much farther along KSP2 is and how much faster it’s progressing nonetheless.

I don’t know whether you watched the video I linked, but you should, if only for the warm fuzzies you’ll get (or the what factor if you only started playing recently).  Playing KSP from its humble beginnings over the years was a lot of fun - watching it grow and experimenting with the new additions was the best gaming experience I ever had.  I’m almost sad that we won’t get as long with KSP2…

[JoeSchmuckatelli]

25 minutes ago, Wheehaw Kerman said:

far more ambitious recreation

Is true. 

 

26 minutes ago, Wheehaw Kerman said:

rewritten from the ground up

Is not. 

Look - I was ready and waiting to fangirl all over KSP2 when it released. Despite my disappointment, I'm still hopeful that they can achieve their goals.  That said - any optimism I have is deeply tempered by the comparison of what we were led to expect vs what we got. 

I truly appreciate the current 'tone' of the development updates - and think that had they been this honest from the start that the reputation harm would be far less. 

(You might backtrack through some of my posts from pre to post release to see that I am well informed about the game, its direction and issues) 

[Scarecrow71]

1 hour ago, Wheehaw Kerman said:

wheels were still frustrating in 2016

Wheels are still frustrating to this day.

1 hour ago, Wheehaw Kerman said:

I’m optimistic that we’re going to continue to see KSP2 mature into a much better and bigger game than KSP, much faster.

I am trying to hold out hope that this happens.  I'm not sure if it will be faster now than KSP was back in the day, but I do hope it happens.  The one thing that KSP2 has that KSP didn't - at least in the early days - is an entire team funded by a parent corporation.  If I'm correct, @HarvesteR was the only person coding KSP for the longest time.  So to compare what was vs. what is and speed really isn't apples to oranges.  It's more like apples to crab apples.  But still, I do hope that it happens.

[Wheehaw Kerman]

4 hours ago, JoeSchmuckatelli said:

Is true. 

 

Is not. 

Look - I was ready and waiting to fangirl all over KSP2 when it released. Despite my disappointment, I'm still hopeful that they can achieve their goals.  That said - any optimism I have is deeply tempered by the comparison of what we were led to expect vs what we got. 

I truly appreciate the current 'tone' of the development updates - and think that had they been this honest from the start that the reputation harm would be far less. 

(You might backtrack through some of my posts from pre to post release to see that I am well informed about the game, its direction and issues) 

Well, I am neither an IG insider nor a programmer, but my understanding is that the game doesn’t re-use KSP code.  I’d quite curious to see any credible information to the contrary.

Agreed re the tone, and I do remember seeing you around quite a bit.

[JoeSchmuckatelli]

8 minutes ago, Wheehaw Kerman said:

t my understanding is that the game doesn’t re-use KSP code

Go through the various post release Dev posts.  You will be surprised. 

Easiest is to start with the whole PQS+ system and then keep going.  There is certainly new work, but it's foundationally KSP. 

[Wheehaw Kerman]

3 hours ago, Scarecrow71 said:

Wheels are still frustrating to this day.

I am trying to hold out hope that this happens.  I'm not sure if it will be faster now than KSP was back in the day, but I do hope it happens.  The one thing that KSP2 has that KSP didn't - at least in the early days - is an entire team funded by a parent corporation.  If I'm correct, @HarvesteR was the only person coding KSP for the longest time.  So to compare what was vs. what is and speed really isn't apples to oranges.  It's more like apples to crab apples.  But still, I do hope that it happens.

I got a good chuckle out of your wheel comment.  I remember trying to get a very basic rover up a very gentle slope on Duna back in 2016 or thereabouts…

I’m with you on the hope, but FWIW, don’t see a heckuva lot of cause for despair.  I can’t see the remaining dev work taking anywhere near as long as KSP - if nothing else, IG is much bigger and better funded than Squad, as you note.  The tone around here is depressing enough if you take it seriously, but that’s all speculation and semi-informed-at-best guesswork, much of it rage-fuelled.  Even with their non-stellar comms track record IG are more credible.

I’ll start feeling less optimistic if Science gets delayed more than another six months, or if the intervening patches show little to no improvement.  But right now I look at the delta between 0.1 and 0.1.3, and I can’t help but feel that things are headed in the right direction.

 

6 minutes ago, JoeSchmuckatelli said:

Go through the various post release Dev posts.  You will be surprised. 

Easiest is to start with the whole PQS+ system and then keep going.  There is certainly new work, but it's foundationally KSP. 

Right.  I saw the PQS+ post.  Allow me to amend to “already rewritten to a large extent with at least some of the big legacy bits to be replaced over the course of EA”.  

[Dakitess]

Wheehaw, don't you feel that the actual patches already feel like applying dressings* and dressings* on a fragile basis ? It does not feel like logical and legit optimization + bug hunting, to me, but more like patches that address a specific bug while using dirty workaround that might cause some other instability, be they detected or not detected.

(* dressings ? This is what I found for translation, I mean here "bandage", we use this word in French to say that you apply layers and layers of patches on a fragile basis)

It's pure speculation, I admit : I don't know much about game dev and I'll agree that EA update will mostly be this kind of bug hunting and optimization, if we're not speaking about added content and features.

... But the way it's done, the communication, the wording, and the liability so far, does not make me confident. Most of the worst bugs are still in investigation, 4 months is not nothing especially when we know that theses bugs were very probably (I hope ?) known from QA, of course. Lot of bugs were solved with weird contraption that induce new bugs. Again, it's quite normal to some extent, indeed, but it's just a feeling that KSP2 foundation are very weak and prone to be a bug mess really hard to be solved OR to be solved while inducing a lack of reliability, repeatability, etc : it would be quite something game breaking for a KSP game, to me. You need to feel confident in the physic of the game, otherwise it's does not work.

Open question / debate here, nothing more !

[Wheehaw Kerman]

51 minutes ago, Dakitess said:

Wheehaw, don't you feel that the actual patches already feel like applying dressings* and dressings* on a fragile basis ? It does not feel like logical and legit optimization + bug hunting, to me, but more like patches that address a specific bug while using dirty workaround that might cause some other instability, be they detected or not detected.

(* dressings ? This is what I found for translation, I mean here "bandage", we use this word in French to say that you apply layers and layers of patches on a fragile basis)

It's pure speculation, I admit : I don't know much about game dev and I'll agree that EA update will mostly be this kind of bug hunting and optimization, if we're not speaking about added content and features.

... But the way it's done, the communication, the wording, and the liability so far, does not make me confident. Most of the worst bugs are still in investigation, 4 months is not nothing especially when we know that theses bugs were very probably (I hope ?) known from QA, of course. Lot of bugs were solved with weird contraption that induce new bugs. Again, it's quite normal to some extent, indeed, but it's just a feeling that KSP2 foundation are very weak and prone to be a bug mess really hard to be solved OR to be solved while inducing a lack of reliability, repeatability, etc : it would be quite something game breaking for a KSP game, to me. You need to feel confident in the physic of the game, otherwise it's does not work.

Open question / debate here, nothing more !

Un pansement?  Je ne savais pas que c’était le mot pour “bugfix”.

I’m not sure that I can feel coding, or tell the difference between a glitch caused by a foundational coding error and one caused by a typo.

Could the improvement just be bandaids slopped upon bandaids?  Could it be cleaning up imperfections in an otherwise solid architecture?  No way for me to tell.  I can tell that to me, playing on my machine, 0.1.3 seems faster and less buggy than 0.1.  Not perfect, but better, and that and the improvement in the comms gives me hope.

I could go on a riff about how I wish they’d replicated the uncertainty of early spaceflight using a parts failure mechanic instead of using bugs, but I won’t :).

[JoeSchmuckatelli]

@Wheehaw Kerman, et. al. :

I'd like to point out a problem of communication that this community has been guilty of since the release of EA. 

Specifically, conflating 'performance' in terms of framerates with performance in terms of the underlying engine, physics and basic structure of the game. 

The game has had (from release) major systemic problems with the things it is supposed to do well - mimic atmospheric and spaceflight via user-created air and spacecraft.  This is the one thing that KSP fans hoped would have been 'done right' by a 'rebuilt from the ground up' development philosophy.   It wasn't. 

 

It simultaneously had a very high overhead in terms of playability preventing average machines from getting playable frames.  Those with good machines were able to play and find the foundational problems.  Those with lesser machines thought it was progress that after several patches they could play at all. 

The community regularly crows about improvements resulting in a higher framerate average than the previous iteration - when we should be focused on the foundational issues plaguing flight, orbits, craft construction and predictable, reliable game play. 

 

7 hours ago, Dakitess said:

the actual patches already feel like applying dressings* and dressings* on a fragile basis ? It does not feel like logical and legit optimization + bug hunting, to me, but more like patches that address a specific bug while using dirty workaround that might cause some other instability

This is a very real concern - and I wish there was a way for us to require anyone talking about performance to mean specifically the foundational things the game is supposed to get right... And not anything affecting framerates. 

... At least for now / until those issues are resolved.  It would be lovely to be at the eyecandy phase.  We are not. 

[Wheehaw Kerman]

1 hour ago, JoeSchmuckatelli said:

@Wheehaw Kerman, et. al. :

I'd like to point out a problem of communication that this community has been guilty of since the release of EA. 

Specifically, conflating 'performance' in terms of framerates with performance in terms of the underlying engine, physics and basic structure of the game. 

The game has had (from release) major systemic problems with the things it is supposed to do well - mimic atmospheric and spaceflight via user-created air and spacecraft.  This is the one thing that KSP fans hoped would have been 'done right' by a 'rebuilt from the ground up' development philosophy.   It wasn't. 

 

It simultaneously had a very high overhead in terms of playability preventing average machines from getting playable frames.  Those with good machines were able to play and find the foundational problems.  Those with lesser machines thought it was progress that after several patches they could play at all. 

The community regularly crows about improvements resulting in a higher framerate average than the previous iteration - when we should be focused on the foundational issues plaguing flight, orbits, craft construction and predictable, reliable game play. 

 

This is a very real concern - and I wish there was a way for us to require anyone talking about performance to mean specifically the foundational things the game is supposed to get right... And not anything affecting framerates. 

... At least for now / until those issues are resolved.  It would be lovely to be at the eyecandy phase.  We are not. 

Absolutely valid point, although I really can’t tell one way or the other.  I’m really not qualified to address either the underlying physics or the coding, and I’m one of the people with new machines that can run the game well.

[asmi]

10 hours ago, Dakitess said:

Wheehaw, don't you feel that the actual patches already feel like applying dressings* and dressings* on a fragile basis ? It does not feel like logical and legit optimization + bug hunting, to me, but more like patches that address a specific bug while using dirty workaround that might cause some other instability, be they detected or not detected.

(* dressings ? This is what I found for translation, I mean here "bandage", we use this word in French to say that you apply layers and layers of patches on a fragile basis)

This is what a lot of mature products look like in my experience (20+ years of software development of pretty much everything under the Sun from websites and fintech LOB applications to AGV guidance systems and applications for factories' shop floor). It doesn't neccessarily mean that foundation is fragile, but it just happen to not account for certain edge cases that planners/developers/testers didn't think of when they were working on a system. This is where 80/20 rule comes from - it takes a relatively small amount of time ("20%") to implement a system, and then the bulk of work ("80%") goes into ensuring it works over the entire range of cases which it's supposed to work. The first stage ("20%") is usually referred to as "initial development" (or simply" development"), and the second one ("80%") is "stabilization". There is also a rule of diminishing returns - meaning that you tend to get the most gains in the beginning of stabilization stage ("low hanging fruit"), but as it goes on, it takes progressively more effort to achieve the next step.

Armed with this information, you can actually get a rough idea of where on this journey each sub-system is - if you see a lot of bugfixes in certain sub-system, that means it's just entered a stabilization stage (and it's got a long way to go), but if you see developers wrestling with the same issues for a long time - that means this sub-system is closer towards the end of stabilization. So as much as I hate SOI and trajectory bugs, the fact that they are battling with them for a while gives me some confidence that sub-systems involved a rather mature. This is obviously not a certainty as it makes certain presumptions regarding the team (like that people working it have enough qualifications to actually complete it, and there are no non-technical obstacles), but this is the best we have at this point.

I'm personally more concerned about regressions - while they are inevitable every once in a while, they usually indicate deficiencies in a QA process (not neccessarily with actual QA people, but rather the whole flow from documenting bugs to passing this info to QA, to QA actually understanding info that is passed on them, to them executing testing properly, to passing feedback back to developers).

Edited June 27, 2023 by asmi

[Lyneira]

16 hours ago, JoeSchmuckatelli said:

I'd like to point out a problem of communication that this community has been guilty of since the release of EA. 

Specifically, conflating 'performance' in terms of framerates with performance in terms of the underlying engine, physics and basic structure of the game. 

The game has had (from release) major systemic problems with the things it is supposed to do well - mimic atmospheric and spaceflight via user-created air and spacecraft.  This is the one thing that KSP fans hoped would have been 'done right' by a 'rebuilt from the ground up' development philosophy.   It wasn't. 

 

It simultaneously had a very high overhead in terms of playability preventing average machines from getting playable frames.  Those with good machines were able to play and find the foundational problems.  Those with lesser machines thought it was progress that after several patches they could play at all. 

The community regularly crows about improvements resulting in a higher framerate average than the previous iteration - when we should be focused on the foundational issues plaguing flight, orbits, craft construction and predictable, reliable game play. 

A good point that I've sort of had in the back of my mind as well while making my posts - I like to mention performance and stability as separate issues when I talk about the state of the game. Here's how I personally view these, and my attempt to add a definition for accuracy.

• Performance - Does the game feel smooth to play? This includes framerate, framerate stuttering, scene load time and UI responsiveness. (I'm looking at you parts manager)
• Stability - Does the game function over time without failure, in the full range of situations the player can put it in? Lack of stability includes things like VAB construction bugs, flight bugs leading to sudden loss of vehicle through no fault of the player, wheels hitching on flat terrain for no reason, vehicle falling through terrain, improper physics behaviour, save file size explosions etc.
• Accuracy - Is the game simulation accurate enough with respect to the real world? This includes things like the game's part-based physics system using spring joints (wobbly rockets), part impact resistance, heating, aerodynamics model, wheel function, trajectories and terrain. The current trajectory bug that makes interplanetary transfers difficult is a great example of a lack of accuracy, as well as the orbital decay bug.

Accuracy is where we see the effects of the game's foundation most of all. And like you said, I hoped that the foundations underlying accuracy would have been the ones improved in a new game built from the ground up. The most notable improvement here is the ability to use engines under timewarp, but other foundations were not significantly improved on compared to KSP1 such as spring-based joints and drag cubes.

For example, the part-based physics system is the foundation of KSP's vehicle behaviour, and is primarily responsible for performance quickly getting worse when building vehicles with higher part count and many flight bugs as well as phantom forces (aka, the kraken) also originate here. There have been calls on the forum to get rid of the wobbliness or at least rein it in, as well as enabling scaling to high part count vehicles. Both of these come down to this foundational system.

[Lisias]

On 6/23/2023 at 7:56 PM, shdwlrd said:

None said you should skip it, but they did warn their viewers to expect a less than perfect game.

The only problem I see with this statement is that a less then perfect game is something we already have with KSP¹…

People was expecting way too more from KSP2, and I think this is their only really big problem for now (everything else is just going to be fixed as time goes by).

[Dakitess]

I was clearly expecting something else than what we have right now, and even without the whole bug mess, the performances issues, etc : a "perfectly" working KSP2 as it exists theses days is far far below my expectations. Graphics is very out dated, technically speaking and in addition the aesthetic / visuals aspects are really off to me, including lightning, colors, shaders, etc. UI is a cool novelty I'd say but very suboptimal in many ways. Contents and features are KSP1 clones at best, or missing at worst : there is nothing new, like, nothing at all, no ambition, no creation, there is very nothing that would set apart a KSP2 which has been deved for a few years by a Pro Team, using the MASSIVE feedback and experiences we got thanks to KSP1 already existing.

There is... I mean... It's KSP2 ! It's there to be the new KSP. The one that settle a crazy high expectation for 10 years regarding what is supposed to be a... KSP-Like.

I don't get it. There is so much to do, to improve, to innovate, create, to complete, for a game like KSP ! I don't know, being able to clip part while having them perfectly occluded, drag-wise, or even merged, mass-wise. Being able to make procedural habitats with full procedural IVA : you put 2 habitat module stacked, it merges the whole volume and gives it a predefined look inside. You add one more ? Same, it will adapt, offer choices. It's only a so niche example of what could be done on a very specific subject, and there is aaaaaall others topics, ideas, etc.

Yeah, I don't get why we are heading to a Tree-constrained part structured craft, why we still have weird noodle craft with the same physics, why we have the same weird drag model, the same 50% parts, the same graphics limitation, the same game engine, the same terrain, the same sceneries (lack of, actually), no water flowing rivers, not beautiful tree, no true macro-medium-micro topology, no new rover/wheels physic that would feel realistic, no... Nothing new. The same KSP goind from a 2005 graphics to a 2015-17 graphics at most.

It's not about the bugs killing the game, the communication being so bad (now it's better though !), the performances being atrocious (and getting better as well while about 5-10x not enough to be worth the "optimization" promised), the lack of initial content because it's an EA (it's okay). It's about the goal, the long term goal of a finished game in a 3 years (inchallah !) being... A very very very KSP2.0-ish rather than a proper KSP2, a new game set to be the new KSP for a decade, with the whole opportunity to start it from scratch, to improve so many things, to bring new other things.

Are we really just a bunch feeling that KSP2 has sooooooo low ambition ?

[Newgame space program]

https://steamdb.info/app/954850/charts/

As you can see here after patch 3 and the steam summer sale playercount went back into the triple digits, while there has been a slight decline as of recently thats probably people taking a break waiting for more patches. Its pretty likely the playercount will remain stable at 200 to 300 players till the next patch because the decline is dissimilar to the post launch player decline as its not a steep downwards cliff, I think the reason that this patch was so effective was because unlike the previous two patches, major bugs were patched along with new content being added

[Alexoff]

And how many players have played in KSP2 after the second patch in mid-April?

[Royalswissarmyknife]

Also better performance more people can play with smoother FPS.

Does seem to be declining though.

I think we might see it go back down after some time and then do the same thing but for longer with patch 0.0.4.0

And then finally when Sciences comes out it might actually sustain these high number until Colonies maybe.

Edited June 28, 2023 by Royalswissarmyknife

[shdwlrd]

4 hours ago, Lisias said:

People was expecting way too more from KSP2, and I think this is their only really big problem for now (everything else is just going to be fixed as time goes by).

Very true. I'm included in that camp. Even with the EA announcement, I was expecting a better polished game. Even if it was incomplete and somewhat buggy. It was worse than I expected. 

You're right though, with time it will get better. In this case, patience is a virtue.

[Kelcie]

How does it feel to have taken a sure thing and completely train wrecked it into oblivion?

5k+ hours into KSP1. 

Never even considered buying KSP2 after the release announcement with that sketchy af EA roadmap.

And from a company with the talent and money to have done something good. .

for shame.

[dr.phees]

6 hours ago, Dakitess said:

There is so much to do, to improve, to innovate, create, to complete, for a game like KSP !

Exactly that. The weirdest feeling when I start KSP 2 is that I am bored by its non-creativity. I am never bored in KSP 1. I waste time like it is going out of style.

But KSP 2? It is just a slightly prettier, boring copy. It is the beginning of a remake maybe, but not even close to a successor.

I have gone through old forum entries with suggestions for KSP 1. It is a gold mine for anyone who really wants to do a New KSP. Apparently the team never had a proper look at the forums.

The saddest thing is, that if KSP 2 fails, there might not be a KSP 3, which could have been a proper creative successor to KSP 1.

[Vanamonde]

Redundant threads merged. 

[Alexoff]

I noticed an interesting frame in a video from 2022.

Do I understand correctly that Nate is looking at the same colony that we saw back in 2020? (I don't know where the original source is, but Matt's video from 2020 is easily googled)

I have some guesses as to how this happened, but I'd rather keep them to myself.

[darthgently]

3 hours ago, dr.phees said:

I have gone through old forum entries with suggestions for KSP 1. It is a gold mine for anyone who really wants to do a New KSP.

Exactly this is what has confused me for months.  Simply looking at the currently most used mods for KSP1 should be a huge boon inspirational focus; add all the popular upvoted suggestions over the years, prune to the feasible, and at least a first pass at a high level design is basically done. 

Clearly with colonies, ISRU, and multiuser on the roadmap the project has paid some attention.  So the focus on visuals over the meat and potatoes on the roadmap has been confusing given how much time has passed since KSP2 announcements began.

Edited June 29, 2023 by darthgently