KSP2 EA Grand Discussion Thread.

[The Aziz]

New branch is out, called 0.1.0.0

Currently the same as public branch. My guess is it's going to stay as beta/old to roll back for players, and 0.1.1.0 is going to replace it becoming identical to public branch.

[InterstellarDrifter]

3 hours ago, RocketRockington said:

Nope.  It's rigid body physics and stock Unity (PhysX) physics.  None of the parts deform, which is what soft body physics would entail.  If we saw things like flexing wings (along the span, not just at the root) that would be soft body physics.

Vs reality, completely rigid rockets are a better approximation than the current wet noodle approach.  Rockets deform but not much.  While they are thin walled, they are held up by internal pressure and clever isogrid reinforcement.  Any rocket that bent like a KSP2 rocket would RUD, as the structure would be massively compromised.  Even if it didn't RUD, the structure would remain permanently bent, rather than continuing to flop around.  If you 'bend' a soda can it doesn't snap back into place later.

Ah, coulda sworn I heard Scott Manley state on a video sometime that ksp used SBP.  Anyhoo..

100% understood and agree with all of this, but I was getting at more of a players fault point of view(devil's advocate, if you will). Not every rocket built in KSP1or2 wobble, so why is that? I was just thinking it is pretty much ones with unrealistic aspect ratios and or decoupler sizes/arrangements.  Also, un-optimal atmospheric ascent profiles. 

Edited March 16, 2023 by InterstellarDrifter

[RocketRockington]

3 hours ago, InterstellarDrifter said:

Not every rocket built in KSP1or2 wobble, so why is that? I was just thinking it is pretty much ones with unrealistic aspect ratios and or decoupler sizes/arrangements.

That's not the case though.  A Falcon 9's fineness is about 20 to 1. (3.7m diameter/70 meter length)  Unless you make it out of just one SRB, you can't make a fineness 20 to 1 ratio KSP2 rocket without making it a floppy noodle.    Now, you can argue that real world rockets do not ever hit the angle of attack that KSP rockets do - and that's generally correct (unless you consider something like pegasus) because real world rockets have carefully controlled ascent profiles to minimize drag. 

But I absolutely doubt KSP2 will ever make Mechjeb ascent guidance a stock feature, and therefore - just like with planes - KSP rockets should be more tolerant of aerodynamic forces, not less - and especially not less in such a goofy manner.   

Further much of the time the goofiness can occur in places real rockets actually do fine - like pogo rockets that do weird oscillations when you're hitting 5,6,7g's.  Real world unmanned rockets, especially on the orbit stage, do hit those g figures near burnout.  

[Gargamel]

1 hour ago, RocketRockington said:

But I absolutely doubt KSP2 will ever make Mechjeb ascent guidance a stock feature,

Even with MJ, floppy rockets were an issue.   I would actually argue a standard KSP gravity turn is less stressful to the joints than one controlled by MJ where there’s constant control input.   Edit: Which just proves your point even more. 

Edited March 16, 2023 by Gargamel

[Jacke]

2 hours ago, Gargamel said:

Even with MJ, floppy rockets were an issue.   I would actually argue a standard KSP gravity turn is less stressful to the joints than one controlled by MJ where there’s constant control input.   Edit: Which just proves your point even more. 

Gargamel, MechJeb in KSP v0.23 could fly rockets with that issue's crazy weak joints better than I could by a long shot.  Probably because it kept the control forces just enough to do the corrections needed.  At the time, I managed to fly a simple rocket without struts to orbit successfully once and was completely exhausted.  With MechJeb's Ascent Guidance in control, it did it over and over again.  Only in later versions did joints get strong enough--with no visible wobble--that I could easily fly rockets to orbit manually.

From what I remember and what I've seen, I'd say KSP2 is more like KSP1 v0.23 in the strength of its joints, if not worse.

Edited March 16, 2023 by Jacke

[Manul]

6 hours ago, InterstellarDrifter said:

I was just thinking it is pretty much ones with unrealistic aspect ratios and or decoupler sizes/arrangements.

And with unrealistic joint model that doesn't create multiple joints. Let me explain why the "sausage tree" model is ridiculously stupid for everything except rockets (for rockets it works not so bad). Let's build a spaceshuttle.    IRL the Big Orange Tank is attached to the shuttle in multiple points and is perfectly rigid. In KSP with it's tree joint structure it can be attached only in one point and wobbles like hell. This can be fixed by struts and still not a big issue. So let's build something like Cocnorde: IRL the wing is attached to fuselage all over the length of it's it's root chorde and is perfectly stable at 2 axis out of 3. In KSP it's attached only in 1 point somewhere in the middle and has all 3 axis of freedom. IRL these loooong engine nacelles are attached to the wing all over their length and in KSP they have only 1 attachment point and hang from the wing like dead snakes.  Crafts inherited their tree physical structure from being a tree as a data structure. My biggest hope for KSP 2 was that it will change and we'll get a possibility to create closed rigid structures like a triangle without struts or docking ports to fix the loose vertex.

[Gargamel]

54 minutes ago, Jacke said:

Gargamel, MechJeb in KSP v0.23 could fly rockets with that issue's crazy weak joints better than I could by a long shot.  Probably because it kept the control forces just enough to do the corrections needed.  At the time, I managed to fly a simple rocket without struts to orbit successfully once and was completely exhausted.  With MechJeb's Ascent Guidance in control, it did it over and over again.  Only in later versions did joints get strong enough--with no visible wobble--that I could easily fly rockets to orbit manually.

From what I remember and what I've seen, I'd say KSP2 is more like KSP1 v0.23 in the strength of its joints, if not worse.

I wonder if the forces were worse with MJ and floppy rockets, but because it would react faster, the rocket flew better.    I remember having to use MJ back then mainly because of frame rate more than any other reason.  

[Dakitess]

The simple fact that Struts exist and are possible, even with the tree-structure, means that it would techically be possible to have closed-loops structural attachments, nah ?

Like, as you said, a wing being attached by its extremity, at least two linear points rather than a single central point. It would already eliminate a massive wobble. For a tank stacked on an other one, the linkage would be something like 3 points to be isostatic ?

Actually, 3 points attachments would allow for a planar definition of the attachment, being iso static and probably enough for anything.

It's already something that we are doing ourselves with struts, back in the day or with this crappy KSP2 physic : struts from the tank below the decoupler, to the tank above the decoupler, to bypass its wobbliness and create a structural rigid linkage made of 4 additionnal points, somehow.

Is it something that's not really possible to define in KSP, multiple attachment points ? Of course, the stacking would still be Single Green Node for the user experience, when he's designing his craft, it's all in the background.

Or do we know if there is a way to "bake" the physic of a craft ? Like, something costing 3-5 seconds to propagate parts linkage that should be rigid ? The idea would be to  weld 3 tanks in top of each other, as it actually define a long tank (since we don't have procedural that would solve that alltogether...), keep decouplers as physic independant linkage (but reinforced), etc. It would even count on Clipping as a welded joint, which would be so useful for airplanes and many others cases. So yeah, something that would treat the physic as a global envelope, while keeping some joints physically independant to allow for decoupling, actions (deployments, etc).

That would be a massive change, something really worth a KSP2, along other things.

[Manul]

49 minutes ago, Dakitess said:

The simple fact that Struts exist and are possible, even with the tree-structure, means that it would techically be possible to have closed-loops structural attachments, nah ?

This is what I do with docking ports. They allow to create loops. If I make a heavy ramp lifted by 2 pistons I dock them to the ramp. I even made a (mostly) rigid scissor-link structure for retractable floats  to make a hypersonic seaplane, but this plane turned into a shapeless mess of parts due to Scene Load Kracken.

 

49 minutes ago, Dakitess said:

Is it something that's not really possible to define in KSP, multiple attachment points ? Of course, the stacking would still be Single Green Node for the user experience, when he's designing his craft, it's all in the background.

This was already done by Kerbal Joint Reinforcement mod. These attachments are still wobbly-stretchy  UnityJoints but by using more of them things get more rigid. Recoupling Bycouplers mod does a similar thing.  This is not a fix for Kracken attacks caused by inaccurate physical calculations but it fixes sausageness.

Edited March 16, 2023 by Manul

[RocketRockington]

@DakitessYes you can parent rigid bodies such that there is no physics joint, it's just rigid bodies who's mass and inertial tensors are calculated from the combination of parented parts.  It's not even difficult in Unity.  You do get some weirdness where, for instance, every part of that combined structure will be jolted equally by an impact - what's nice about physics joints is that if you slam into a light piece of tail section, that part could fold first, for instance.  But you could write a workaround for even that, a bespoke joint stress calculation and detach parts when a joint stresses exceeds a threshold.

 

But doing that would have required more forethought than just copying what KSP1 did, and might have forced them to spend more money on programming earlier, so there would be less money for trailers, movies of developers, and  animated cartoon tutorials.  KSP2 had to prioritize thier budget and allocate it to more key features than improved rocket physics - after all KSP 0.23 rocket physics was good enough to sell Kerbal 1.

Edited March 16, 2023 by RocketRockington

[Periple]

8 hours ago, Jacke said:

From what I remember and what I've seen, I'd say KSP2 is more like KSP1 v0.23 in the strength of its joints, if not worse.

It’s not quite that bad. Engine plates are bugged and very wobbly but otherwise it’s manageable for up to medium sized rockets. But it does need improvement.

You can make them very noodly if you want to though!

[FleshJeb]

8 hours ago, RocketRockington said:

after all KSP 0.23 rocket physics was good enough to sell Kerbal 1.

To be fair, 0.23.5 remains one of the best versions of KSP. There's very little you can do in 1.12.x that wasn't already done better by mods in 0.23.

[cocoscacao]

Are we gonna see some parts that were present in the expansions (KV-1,2,3 pods, Bobcat engine, etc...), or have these been removed?

[RocketRockington]

14 hours ago, cocoscacao said:

Are we gonna see some parts that were present in the expansions (KV-1,2,3 pods, Bobcat engine, etc...), or have these been removed?

I think they're unlikely to bring those back, as that's work for them, it's not just dropping them in game.  It's already likely to be kind of silly that the player will have so many current-tech engines when, most likely, once they start founding colonies they will have access to a smaller list of future-tech engines that make everything in the current engine part list completely obsolete.   KSP mostly tried not to completely obsolete engines but I think KSP2 is clearly going in the direction of 'no, all these engines are pointless now, use the 'questionably plausible' metallic hydrogen torch drives now.

[The Aziz]

27 minutes ago, RocketRockington said:

metallic hydrogen and torch drives

Ftfy. Those are two different things. But also

28 minutes ago, RocketRockington said:

KSP2 is clearly going in the direction of 'no, all these engines are pointless now,

Not really if you consider the cost. Resources not being available at chosen site, overall amount required for production...

[JoeSchmuckatelli]

7 hours ago, The Aziz said:

Ftfy. Those are two different things. But also

Not really if you consider the cost. Resources not being available at chosen site, overall amount required for production...

This - and the progression system is going to contribute.  Unless all you do is play Sandbox, you are likely to have to use rockets similar to what we have now and then unlock the new stuff as you explore outward 

[Sea_Kerman]

Agh this orbit decaying bug is really annoying

[The Aziz]

Having a PE above 25km ASL seems to solve it so far. I know it makes no sense to orbit at 25km above, let's say, Minmus, but it works for the time being.

[cocoscacao]

Docking and undocking bug fix, less rocket noodling, a wee bit more performance optimizations, and we're pretty close to KSP 1 experience. Not bad for early part of an early access.

[Tazooka]

I'm at the point where I'm now getting worried about the future of the game.
KSP 2 has been in development for several years and it feel like this game was only made a few months ago in a rushed state. Obviously we don't know the real reason for this, but it would be nice for us to get some context so maybe we can relate and understand, instead of us speculating and being disappointed.

I have played 15.4 hours total on Steam and that is from 3 saves; 1 that corrupted, 1 from the first patch and 1 patch two as I had issues with the first patch save. I want to play more, but I really can't be bothered with all the bugs and frustrating glitches. I am seeing that more people recently are also sharing similar feelings.

I understand the game has files that contain features and parts for future updates, but they're not actually available and ready. I find that questions to the dev team (past and AMA) are answered in a vague way that, while are reassuring don't actually offer much value.
So far we have a buggy, incomplete, laggy KSP 1 that looks better, with the hinting of features in the game that will either be after 1.0 (robotics), not added (life support) or vaguely described like how it was in the original game (science), unimproved.

I have also felt like the developers and community managers have gone quiet since the first patch and the communication has got worse, especially on this forum. Discord got a lot more attention when the game was released, but even that has been quiet with dev updates and replies recently.
A simple like on a forum post will be enough because it lets us know you're read and acknowledge the issue.

I don't feel like this game is doomed as I am optimistic about the future. I am just concerned with how this release has gone so far.

Lastly, I have to agree with one of the quotes from the AMA, that is also incomplete.

Quote

One unexpected thing is that sometimes we hard focus on one feature/fix that is low priority for the community. [WILL ADD MORE]

 

TDLR:

• The game does not feel like it has been developed for years
• Don't want to play the game much because of its poor state
• Be less vague about the future KSP 2 and its updates
• Communicate more and not just on Discord
• AMA quote is right about prioritising the wrong stuff

[Kerbart]

Writing code is hard, especially the “90% is here, only need to finish it” part. KSP1 wasn’t any different, although with a much smallerr scope patches were rolled out much quicker in the beginning. But not after , say, 0.18. Important features were missing there as well and despite the community asking sometimes not added for years.

It’s an ambitious project with a lot of moving parts that, last year, quite suddenly was confronted with “go live on February 24.” That  means that things like the new terrain system had to be shelved and replaced with the old system, and everything else had to be adopted for it, I assume, based on what we have that most of the work done after the Feb 24 announcement was spent on piecing a working version together from many parts, old and new, to get something to publish.

It’s far from perfect, and disappointing. The lather based on high expectations (set by marketing, not in the least the pricing) and that’s unfortunate as the game deserves better, Take it for what it is: early access, and a work in progress. Don’t go out and replicate the most  complex KSP1 missions. You have KSP1 for that. Instead, pretend it’s a puppy. Watch it grow, learn its strength and weaknesses and remember that KSP 1 took many years to be where it is now.

[Fullmetal Analyst]

1 hour ago, Tazooka said:

community managers have gone quiet since the first patch and the communication has got worse

i can only agree with this, the community managers seem to have no idea what they are doing, im starting to think they are just random dudes with no actual background or experience in community management, aside from maybe "managing" (banning everyone who speaks up) some small discord communities 

so far i have not seen a single poll or anything from them, all they really did recently was gathering some AMA questions, and they actually didnt even forward the most  interesting questions

[Tazooka]

1 minute ago, Fullmetal Analyst said:

i can only agree with this, the community managers seem to have no idea what they are doing, im starting to think they are just random dudes with no actual background or experience in community management, aside from maybe "managing" (banning everyone who speaks up) some small discord communities 

They do seem to have experience, especially @Dakota. But I don't understand why nothing is being said, it seems common with larger companies which is annoying.

No news is good news?
Even if their job is to communicate with the community.

Just now, Fullmetal Analyst said:

all they really did recently was gathering some AMA questions, and they actually didnt even forward the most  interesting questions

I can't believe they chose a question about cereal when there was so many good questions about the game. 'Fun' or not, at least make it relatable. I don't care how Nate has his cereal.

[Fullmetal Analyst]

1 minute ago, Tazooka said:

I can't believe they chose a question about cereal

something is going terribly wrong here

2 minutes ago, Tazooka said:

especially @Dakota.

well i didnt want to call names...

[Tazooka]

4 minutes ago, Fullmetal Analyst said:

something is going terribly wrong here

Trigger warning? What'd I miss?

4 minutes ago, Fullmetal Analyst said:

well i didnt want to call names...

Well there are only 2 CM's 

Dakota has a crazy good history being a CM.
If I remember correctly they've worked with AAA titles like God of War

Edited March 28, 2023 by Tazooka