Developer Insights #23 - Black Hole Sun

[NexusHelium]

3 minutes ago, herbal space program said:

I've built some pretty big stuff recently, and although it still requires the judicious placement of some struts manually, it's way, way better than it was before.

Makes you wonder how Heretic391 built his Dres Station before FS!, huh? I can say for certain with the construction of my mega structure that this issue has at least been applied a band-aid solution to. And I do remember them saying that they were working on a more long term solution that doesn't involve the need for auto strutting (hopefully prioritizing more critical issues though above this one, of course.)

Edited April 9 by NexusHelium

[Presto200]

On the topic of the wobbly rockets fix, it is a good thing if you need struts for some designs still, that's why they left struts in the game. It makes sense that truly massive ships need some extra help, especially if it is stuff hanging off the side. It's good we don't need excessive amounts of struts just to launch a simple vertical stack rocket. The Devs did a good job on this and we should recognize their goods if we want any right to recognize their bads.

[NexusHelium]

1 hour ago, Presto200 said:

It's good we don't need excessive amounts of struts just to launch a simple vertical stack rocket

I honestly don't need any struts to launch single stack rockets anymore. The fix was incredible in my book. It allowed me to create bigger ships using part counts until I accidentally built one up with over 2000 parts...

Okay... maybe I took advantage of it a little bit  

[cocoscacao]

2 hours ago, Scarecrow71 said:

Is it the best or most elegant fix on the market?  Probably not.  But it currently is not an issue.

Exactly... currently. When you set out to do something more bigger and ambitious, than KSP 1 in this context, you have to plan ahead.

2 hours ago, herbal space program said:

are you sure there is actually a better solution than that using this physics engine?

I won't pretend that I know anything about that physics engine. But, I do have a little knowledge about programming, and as stated above, I'm very much curious how they plan to execute many things. Rocket wobble is just one of them. Continuous long burns, heat transfer, battery draining, etc... computations on all vessels during time warp... those are the meat mechanics of the (new) game.

Thus far, as @PDCWolf pointed out, there was a talk about a short term and long term solution for rocket wobble. They went with middle term solution, involving autostruts. Why? I'm not dissing devs. If you asked me how to implement these things efficiently, I'd say wait for home quantum computers.

The point I'm trying to make is that, we haven't got any indication on how these crucial things will be resolved. After so many years of development apparently,(judging by autostruts, they don't have any indication either. If they do, I'd be very curious to read an update on that. I don't care if I don't get any new parts, or parachute bug resolved within the next few months. I care whether the team is capable of developing the core mechanics needed for this game. Falling back to autostruts doesn't inspire much confidence, and I pray to be proven, oh so very wrong.

Edited April 9 by cocoscacao

[Scarecrow71]

9 minutes ago, cocoscacao said:

Exactly... currently. When you set out to do something more bigger and ambitious, than KSP 1 in this context, you have to plan ahead.

What part of "it's fixed" are you not getting?  It doesn't matter if the fix is current, temporary, or is causing other issues.  IT IS FIXED.  Period.  If you are still experiencing issues, then you need to check your build.  When I have a single stack, I don't see wobble.  When I start going outwards from the center of the rocket, such as when I'm asparagus staging side tanks/engines, or trying to add all the little pieces to the side of a rocket, then yes, it wobbles.  I expect it to wobble if I put something on a radial decoupler.  Why?  Because it isn't part of the central stack and needs support.

Beyond that, I can't help you.  But the issue is fixed.  Full stop.

[The Aziz]

36 minutes ago, cocoscacao said:

They went with middle term solution, involving autostruts. Why?

Because they needed "a" solution fairly quickly. Enough people had severe problems with single stack rockets so to ease with suffering and make the basicest part of the game playable they went for autostruts (which you can turn off by the way, in game settings if you want wobble back). That made people happy (ok, happier than before) so they could focus on other features, while keeping a long term solution in the backlog. If they went for a long term from the start, you'd still have wobbly rockets today and instead of nearly 7000 people playing FS on release day, you'd have 1500, and maybe 90 today, as only masochists would be playing.

[steveman0]

1 hour ago, cocoscacao said:

Exactly... currently. When you set out to do something more bigger and ambitious, than KSP 1 in this context, you have to plan ahead.

Are you of the opinion that if you build some thousand part momstrosity that also shouldn't wobble under extreme load? If so, reality might want to have a word with you.

Reality is that even single stack rockets must be carefully designed to account for bending forces under load in real life. As a KSP rocket designer, the realistic implementation is one where you must consider the potential bending forces when your design becomes extreme either by moving to a more structurally sound core design or by structural reinforcement, i.e. struts.

The wobbly rocket fix is about as close to perfect as I could ask. Realistic looking rockets don't need struts. Move to asparugus staging and add boosters and you might need a few. Move to 1000 part launch craft and you might need to question your design choices. Or just add more struts. There can be no more Kerbal solution.

[herbal space program]

41 minutes ago, steveman0 said:

Move to 1000 part launch craft and you might need to question your design choices.

I dunno, the game does seem to be pushing us in that direction with FS missions like Eve 10 and Big n' Husky, which to me portends even bigger stuff on the horizon. Still, I'm up to around 450 parts on my most complex launches now and I'm not having any wobble problems yet using only a KSP1-typical number of struts, so I'm not too worried about it. The framerate at that part count however is another matter.

[Icegrx]

2 hours ago, cocoscacao said:

 

The point I'm trying to make is that, we haven't got any indication on how these crucial things will be resolved. After so many years of development apparently,(judging by autostruts, they don't have any indication either. If they do, I'd be very curious to read an update on that. I don't care if I don't get any new parts, or parachute bug resolved within the next few months. I care whether the team is capable of developing the core mechanics needed for this game. Falling back to autostruts doesn't inspire much confidence, and I pray to be proven, oh so very wrong.

It’s fixed. The mechanic works extremely well in the context of the issue.

“Falling back on auto struts” - yes they fixed the issue with a simple solution that was already present in game. That’s not falling back on anything. 

There’s definitely criticism to be had of KSP2, but if you beat them up continuously over a past issue that’s been resolved, how are you ever going to look to the future? Conversations like these derail the criticism in an un meaningful way. We should be talking bout current issues and current bugs. 

[Scarecrow71]

15 minutes ago, herbal space program said:

I dunno, the game does seem to be pushing us in that direction with FS missions like Eve 10 and Big n' Husky, which to me portends even bigger stuff on the horizon. Still, I'm up to around 450 parts on my most complex launches now and I'm not having any wobble problems yet using only a KSP1-typical number of struts, so I'm not too worried about it. The framerate at that part count however is another matter.

And until the developers get performance under wraps, stuff like that won't happen on a regular basis.

[cocoscacao]

2 hours ago, Scarecrow71 said:

What part of "it's fixed" are you not getting?  It doesn't matter if the fix is current, temporary, or is causing other issues.  IT IS FIXED.

It very much is important. Pushing problems under the rug, while ignoring things in the long run, in software, leads to failure...

2 hours ago, The Aziz said:

Because they needed "a" solution fairly quickly.

Reader comprehension issues.

20 minutes ago, Icegrx said:

“Falling back on auto struts” - yes they fixed the issue with a simple solution that was already present in game. That’s not falling back on anything. 

There’s definitely criticism to be had of KSP2, but if you beat them up continuously over a past issue that’s been resolved, how are you ever going to look to the future?

It worked well for the first one, that never had gargantuan constructions in orbit with God knows how many parts...

1 hour ago, steveman0 said:

Are you of the opinion that if you build some thousand part momstrosity that also shouldn't wobble under extreme load? If so, reality might want to have a word with you.

Same as the paragraph above...

[The Aziz]

36 minutes ago, herbal space program said:

the game does seem to be pushing us in that direction with FS missions like Eve 10 and Big n' Husky, which to me portends even bigger stuff on the horizon

I'm waiting with those until colonies. I simply cannot be bothered with creating a ship capable of lifting 10 Kerbals off Eve (where I have never done that even with a tiniest probe) and landing it fully fueled and in one piece on the surface. No thanks. I'd rather send down there some colony pieces and build the ship on site.

23 minutes ago, Icegrx said:

We should be talking bout current issues and current bugs. 

Yes but, like, not here maybe? This thread has derailed long ago.

[Scarecrow71]

Just now, cocoscacao said:

It very much is important. Pushing problems under the rug, while ignoring things in the long run, in software, leads to failure...

Not in the context of the fact that it is fixed.  It being fixed is what is important.  I even addressed the fact that it wasn't the best fix.  But it is fixed.  Forget semantics here:  the wobbling is fixed.

If you continue to have issues with craft wobbling, that is on you.  But this issue is fixed.  Period.

[cocoscacao]

1 minute ago, Scarecrow71 said:

If you continue to have issues with craft wobbling, that is on you.  But this issue is fixed.  Period.

So... if tomorrow, you start building your orbital fuel depo like the one in the trailer video, and it starts ripping itself to shreds because autostruts aren't enough to keep it together, and they say, "well whoops, we didn't take into account those part counts and layout...", will it be fixed then?

34 minutes ago, Icegrx said:

but if you beat them up continuously over a past issue that’s been resolved, how are you ever going to look to the future? Conversations like these derail the criticism in an un meaningful way. We should be talking bout current issues and current bugs. 

Just to add to my previous reply. These aren't past issues. These are core features. The pillars of the game. The things that were planned to be done, before development even started.

[herbal space program]

41 minutes ago, The Aziz said:

I simply cannot be bothered with creating a ship capable of lifting 10 Kerbals off Eve (where I have never done that even with a tiniest probe) and landing it fully fueled and in one piece on the surface. No thanks. I'd rather send down there some colony pieces and build the ship on site.

For Eve I consider it an interesting challenge, especially since I don't have the magic inflatable heat shield yet. Landing a 300t ship on Duna however just seems like a boring, pointless grind to me: easy, but also very time-consuming without proving anything.

Edited April 9 by herbal space program

[The Aziz]

7 minutes ago, herbal space program said:

Landing a 300t ship on Duna however just seems like a boring grind to me: easy, but also very time-consuming without proving anything.

Another reason to wait. Those base modules can't be too light, even when they're inflatable. With that, I can get the mission done, and groundwork for a colony. Hell if there was a mission to Ike I'd probably attach a lander to that ship as well. 3in1.

[herbal space program]

3 minutes ago, The Aziz said:

Another reason to wait. Those base modules can't be too light, even when they're inflatable. With that, I can get the mission done, and groundwork for a colony. Hell if there was a mission to Ike I'd probably attach a lander to that ship as well. 3in1.

Indeed, if you're going to land 300 bleepin'  tons of something on Duna, it should actually be something that will ultimately be useful in some way! Nothing we have in our current parts inventory looks very likely to me in that regard. Even my Eve 10 lander, which at least serves the purpose of getting ten Kerbals to the surface and back, only weighs ~175t with its landing legs and brakes/chutes still attached, so 300 tons of what?

[Icegrx]

Starting to see why the team gets more and more quiet as time goes on. 

[Scarecrow71]

45 minutes ago, cocoscacao said:

So... if tomorrow, you start building your orbital fuel depo like the one in the trailer video, and it starts ripping itself to shreds because autostruts aren't enough to keep it together, and they say, "well whoops, we didn't take into account those part counts and layout...", will it be fixed then?

You really are looking for a reason to harp on this, aren't you?  The issue at hand was wobbly rockets, not an orbital station with 1000s of parts.  Because wobbliness isn't a prevalent issue in space, but rather upon launch.  I'm not saying it doesn't exist in space, but the issue that was fixed was related to launch.  We are all aware of what the kraken can and will do to ships with 1000s of parts in space.  That isn't wobbly rockets; that is merely a hungry space demon.  But the issue we are discussing is fixed.

Stop harping.  Please.  You are literally just looking for something to be wrong where nothing is.  The initial issue is resolved.  Period.

[cocoscacao]

9 minutes ago, Scarecrow71 said:

You really are looking for a reason to harp on this, aren't you?  The issue at hand was wobbly rockets, not an orbital station with 1000s of parts.

Re-read my last few posts again

[PDCWolf]

5 hours ago, Scarecrow71 said:

When I have a single stack, I don't see wobble.

2 hours ago, Icegrx said:

It’s fixed. The mechanic works extremely well in the context of the issue.

Big problem with those statements.

It is not fixed, in the sense that the "fix" only accounts for a single configuration. In a game where you build land vehicles, aircraft, spacecraft, space stations, land bases, interstellar ships, and so much more, a solution that only ever works for vertically stacked tanks on the same central axis of the root part is useless. That's not even going into the performance tradeoff (huge) when you build big stuff with autostruts on.

There's such a huge amount of possible cases that only fixing one is probably a single digit percent of the total use cases. It's not fixed, it's a lame bandaid that they themselves said were not going to apply, to hold back a glorified bug that they themselves can't be bothered to patch out; and the community sucked up to said bug to the point of adoring it, and ruining not one but two games for.

Edited April 10 by PDCWolf

[cocoscacao]

1 hour ago, PDCWolf said:

It is not fixed, in the sense that the "fix" only accounts for a single configuration.

One could even put that statement into debate. If I don't put struts on boosters, the entire craft rips itself to pieces on the launchpad.

Workaround is to add struts. But it's still a workaround. Did it make the game more playable? Yes. Did it fix the problem. No. Colonies are next, so we'll see what's been done. Though I understand their decision not to make everything completely stiff... 

[Icegrx]

3 hours ago, PDCWolf said:

 where you build land vehicles, aircraft, spacecraft, space stations, land bases, interstellar ships, and so much more

I’ve built all of these besides interstellar ships, have not had a problem with wobble in any of these situations. 

the one exception I’ve found is with some larger wings when I’m not careful to what it’s mounted on. They can fall off if you aren’t careful. Has yet to stop me from building any particular craft. 

when scaling up the part count, the biggest performance dive for me is when multiple fuel sources are involved. Especially when splitting multiple fuel sources into multiple engines or RCS. 

I don’t see this as an auto strut issue, rather a general performance issue. Once the game gets a few more optimization passes I believe auto struts will be negligible on frame rate. 

[Oak7603]

My apologies, by wobbly rockets, I meant the rockets that start wobbling on the launch pad and then explode and then a relaunch fixes it. @herbal space program in the post above is proof taht this is still happening.

As for @The Aziz needing to sigh, I never meant for that guy to fix it but your stalwart defense of KSP2 does not go unnoticed. To be fair to that guy though, it's great that his team has time to play around and spend time on a pointless, miniscule addition.  Maybe if more of the game was ready, like the colony parts or new parts or terrain or lighting on the supposed new planets etc, They would have some more worthwhile tasks to oversee.

The point I was clearly making was that as a whole, the studio should be working towards fixing the small things that have a big impact, not this waste of time.

[GGG-GoodGuyGreg]

10 minutes ago, Oak7603 said:

The point I was clearly making was that as a whole, the studio should be working towards fixing the small things that have a big impact, not this waste of time.

Agreed. 
To me it sounds like they’re just capitalizing on the recent eclipse and writing a theoretical article just so they check it off the marketing  hype todo list.

The only thing I would see NASA being interested in reaching out for is how they managed to be the first ones to expand time.
With update cadence which went from “in the coming weeks” to updates in the coming years. 

Edited April 10 by GGG-GoodGuyGreg