The only thing thing that I care to see in .24

[xrayfishx]

The joints are, in fact, too strong (on average. Some parts do need stronger joints, but on average it's the opposite). Even with FAR I've only had large rockets break in half from turning too hard and flipping as a result. However, the joints are too flexible. This allows them to bend, causing loss of control.

This has been said before in this thread, but since it's a third opinion, not agreeing with either side, no one even noticed it. Both sides are blaming each other for making invalid arguments while neither side actually seems to give the other opinion or the arguments supporting them much thought.

Those who support current joints basically say either that "its teh kerbal way moar struts moar boosters" or that wobbly makes you try to make your design more efficient.

The "Kerbal Way" argument is BS. Kerbals are excellent engineers, even though they like explosions a lot.

The more efficient design is a valid argument, however it is very restricting without procedural parts and also bigger parts. This does make 30 tons payload a fairly big one.

Those who support stronger joints basically say that wobbly is unrealistic and more an annoyance than a challenge and ruins aesthetics.

"More an annoyance than challenge" and the aesthetics part is something I have to, at least partially, disagree with. It is annoying, but it's a part of the challenge. You can make aesthetic rockets while being efficient. You can't expect to make replicas, however, unless we go procedural. As for realism, this is a completely valid point, but the fix for it is in the first paragraph. It's not the joint strenght that is bad, it's that it flexes instead of snapping. If that is fixed, it's actually more realistic to use many small tanks rather than one big one because then the rocket can snap in the middle of a tank, instead of at where a joint would realistically be, which is, well, realistic.

Now, notice how I wrote all this without bringing up any new points. Every single thing I wrote has been written before in this thread (except for that snapping in the middle of a tank thing - that was in another flame war thread.

We have a total of 175 posts summed up in this post. Please, make sure not to post the same thing over and over again. Especially you, wobbly rocket supporters, but the other side, too.

But to remind you of the main point of this post and make sure gets through, I will inform you that Bob, Johnbro, Obvey and Seanbro died because the wheels on my experimental SSTO flexed. They didn't break, they flexed.

Edited February 18, 2014 by xrayfishx

[Blackstar]

I'm not sure I'd like to see KJR stock quite yet. Maybe eventually, but right now there are too many issues to be ready for prime time.

As it is, I've installed KJR 3 or 4 times and always ended up uninstalling it after a while. Sure, it makes (most) stack connections more rigid, but it seems to make a lot of other joints much MORE flexible than they should be, especially trusses, I-beams, structural panels, and (ironically) struts. Constructions that are are rock solid for me in stock, even if I have to add some struts here and there, bend like spaghetti with KJR and no amount of struts seem to fix it.

[Ravenchant]

The joints are, in fact, too strong (on average. Some parts do need stronger joints, but on average it's the opposite).

Two of those parts would probably be the Sr. docking ports and the largest RCS tank. Seriously, the things shear off like they were held with duct tape.

[regex]

I'm not sure I'd like to see KJR stock quite yet. Maybe eventually, but right now there are too many issues to be ready for prime time.

As it is, I've installed KJR 3 or 4 times and always ended up uninstalling it after a while. Sure, it makes (most) stack connections more rigid, but it seems to make a lot of other joints much MORE flexible than they should be, especially trusses, I-beams, structural panels, and (ironically) struts. Constructions that are are rock solid for me in stock, even if I have to add some struts here and there, bend like spaghetti with KJR and no amount of struts seem to fix it.

KJR isn't becoming stock; something like KJR will show up in 0.24 according to recent dev posts.

[Mr. Speed]

do people actually realise how heavy some of the parts are? I think the connections aren't that bad. only thing I could think of as a problem is how node connections are just one point connections, instead of connecting the whole touching surface of the two parts together.

[Whackjob]

I was able to launch this:

And what does that mean? ...nothing, really. Other than a good example of the kind of extremes one must go to sometimes to reinforce something. I'm down with stronger connection points on tank ends. I can't wait to see what I can do with it.

[Rosco P. Coltrane]

The current problem with the connections, that will be solved in 0.24, is that from a game engine POV, the connection happens in the center of the pieces instead of the point where they come into contact visually. That's way an engine will dance under an orange tank, because the engine is attached to a point in the center of the tank, which is too far away.

This makes the logic for the joints tricky and that is what is going to disappear in 0.24. The connections are not going to be stronger nor weaker (assuming they don't go that way, which they say they won't) they are just going to be more stable.

[Whackjob]

Works for me! I'll take it.

... put it on my tab.

[Rosco P. Coltrane]

I'm sure you'll find a way to overheat CPUs with another flying monster. Probably replacing X amount of struts by X amount of SRB or something like that.

[Whackjob]

I don't really use the boosters provided in-game. They don't scale well to the large. I just make my own.

[justmeman117]

jeez, so much argument over a simple issue. nobody's offering any real solutions either.

here's an idea that might work as a solution to wobbly tanks and weak joints without making it too easy on the player. instead of simply making the joints stronger, why not make them more realistic?

there's nothing wrong with the physics of the game in context of wobbliness, it simulates the wobbly just fine. what's messed up is how the ENGINEERING is simulated.

at the moment, the joint system seems to be based upon joints that connect at a single small point. this isnt very realistic, you dont weld two parts together in the real world by a single dollop of weld, do you? it wouldnt matter how strong that one dollop of weld is, it'd still be unstable in that the parts would be liable to pivot around that point. even though the parts arent supposed to rotate from that point, they still do a little bit, but a little bit on several parts is enough to generate oscillation. the wobbliness has less to do with joint strength than it does with joint area.

imagine 2 orange tanks snapped together. all the weight of the upper tank is resting on a single point, the snap point. but in real life, all that weight would either rest on the lower tank in a disk, a washer, or a hoop. all of these spread out the weight better and provide support on the outer edges of the parts that keep them from rotating about the center. instead, they have to rotate about the edge to rotate at all, which is harder to do than at the single point, because all around the joint it is solid. the parts push together on the side bending in, and pull apart on the side bending out, so you get this sort of double feedback lever that makes things harder to bend in the first place. this is much less wobbly than a joint of equal strength at a single point. i think the game mechanic finds it difficult to keep parts from bending at the single point joint.

so instead of using these point snaps, why not use disks or washers or hoops instead for snapping parts together? they are more realistic, and provide a larger surface area to spread weights over. parts snapped together would act more realistically, not wobbling about single points, but only about the edges of parts snapped together, like a real rocket.

doing this might be difficult since it might make parts incompatible. so to ensure parts are compatible in the game at all times, i think only the option of disk or point joints be used. you could switch between the two for different purposes.

these would make joints more stable perpendicular to the rockets acceleration, but not make them stronger parallel to it (so engineering skill is still needed).

what i'd like to see to make joints STRONGER is some sort of girder/interstage/payload mount BUILDER (specifically for parts like adaptors, girders, etc, mostly i have the octagonal girder in mind, but of varying shapes, sizes, and strengths), so you could replace very unrealistic struts with you're own custom parts. this builder would calculate the strength of the part based on how you've designed them (so you need engineering skill). this is where struts would come in. you use something like struts to make the girders, or use solid for interstages/adapters (but driving up the strength also drives up the weight). once the part was made, you could save it in-game as a single piece part.

why not even go a step further? i saw one post mention that it's unrealistic to slap tanks of standard size together. this is true, and it drives up part count as well. so why not take a program along the lines of PartGen3.1 and make it part of in-game vanilla KSP? on top of the standard parts, you could design your own in game specified to your exact needs, but at the same time, calculates realistic weights and fuel amounts for your rocket (so you dont go all cheaty with it). it's already an application, why not incorporate it like sub assemblies was (which is bugged, btw, but whatevs).

all these solutions combined would both make KSP engineering more straight forward AND realistic. it would even increase creativity, if partgen were made part of vanilla ksp, people could make not just ships, but basic PARTS, and custom textures.

for people who bicker a lot over a small aspect of the game, all you guys are really saying is "fix it or leave it". it does need to be fixed, but i dont think "make the joints stronger" is in any way HOW it needs to be fixed, nor is it the only alternative to wobbly rockets. i think after 19 pages, there should have been more creative solutions than that.

[justmeman117]

*i forgot to mention the kerbal engineer mod, which also really outa be incorporated into vanilla ksp as well.

[KasperVld]

Fixes to the joint system were introduced in Unity and implemented in the game in 0.23.5, so I think this thread isn't all that relevant anymore

[SRV Ron]

Joint connections are much better in 0.23.5 even with the 3.75 and 5 meter Nova Punch mods. However, the better connections still cannot overcome poor design. And, a rocket with 100 stacked donuts tanks will still whip around like a wet noodle.

[SSSPutnik]

I build big rockets, like, 800t to LKO.

I also use very few struts. For 3.75m I typically use 1 pair in X shape between boosters, a decoupler, a fuel line and then 1 or 2 more to central.

With 0.23.5 the connects are better and stuff flies wonderfully with a well built and symmetrical stack.

[justmeman117]

heh, i guess im a little behind on the news then i think im still using a slightly older version of ksp as a matter of fact.

either way, i still think it would be cool if they incorporated partgen into the game itself, and maybe make joints at least a bit more logical. i dont really have huge issues with wobble, mostly just confusion about why the joints work the way they do.

i think a bigger issue with wobble was the way fins and RCS worked in the previous versions of the game, different controls at the bottom and top of the rocket would continually oscillate the whole rocket back and forth. SAS tended to overdo corrections. the fact that they fixed that in 0.23 fixed most of the issue for me.