Backed into a corner: why a broken feature can end up worse than none at all

[Dunbaratu]

The raycast drag proposal, while intuitively correct, is apparently a terrible and unrealistic idea, as is proven by Ferram4's raycast drag experiment (Not sure if it still works).

http://forum.kerbalspaceprogram.com/threads/24655-0-19-Ferram-s-Raycast-Drag-Experiment-v0-1?highlight=raycast+drag

I agree that one implementation of it, written by someone with an express purpose of proving it doesn't work, didn't work. Big surprise.

The connection between what that person implemented and what I talked about is minimal at best. For one thing. their model didn't just alter drag depending on occlusion, but it also altered control-surface effectiveness which was what caused most of the problems and was NOT in my proposal. For another thing, they made occlusion boolean. A part was occluded or not. Nothing in between. That was not my proposal at all, if you read it more carefully.

Edited June 22, 2014 by Steven Mading

[cpast]

Ferram actually said he made it to try to prove himself wrong - he had suspected it didn't work, but thought he might be wrong and made it to check. The issue is that that sort of occlusion has approximately nothing to do with real aerodynamics, until you get into hypersonic speeds. That means results are just as counter-intuitive as KSP's model, with a side bonus of "no aerodynamic stabilization at low speeds" (making planes near-unworkable), and plenty of places where real aircraft should see lift but the occlusion model thinks they don't (e.g. vertical stabilizers get blocked by the airframe at very low angles of attack, when real aircraft obviously don't have that happen). Furthermore, it is not simple for the computer to check these things, so it's not a lightweight aero system.

Any remotely sensible aero system has to implement lift and drag as related quantities; if not, it's a horrible system. If a part isn't exposed to airflow to create drag, it cannot then create lift, and if a control surface isn't creating drag, it's ineffective. Sticking an elevator in the middle of a brick wall should not work if you don't want brick walls to fly. Having non-boolean occlusion wouldn't help the computational intensity the model requires.

Edited June 22, 2014 by cpast

[regex]

Yeah but let's face it, any change to the game is going to cause dissent. HarvesteR could change what tissue paper he uses when he sneezes and a portion of the forums would be up in arms over it.

I should have saved my rep for this post, because it sums up the forums nicely.

[Deathsoul097]

This is exactly it. I don't think anyone's really arguing for a flight sim level of realism, or even that FAR should become stock, but the aerodynamic modeling is clearly broken. At the very least my intakes should have to face the path of flight, drag should be applied to the brick wall, and stacked wings shouldn't generate additive amounts of lift.

I agree on the subject of intakes (To an extent, at lower speeds intakes can still manage to pull in air from their surroundings.), and on the drag, but as for stacking wings, I don't mind that. Not that necessary to "fix", IMHO.

[Vaporo]

I agree that one implementation of it, written by someone with an express purpose of proving it doesn't work, didn't work. Big surprise.

The connection between what that person implemented and what I talked about is minimal at best. For one thing. their model didn't just alter drag depending on occlusion, but it also altered control-surface effectiveness which was what caused most of the problems and was NOT in my proposal. For another thing, they made occlusion boolean. A part was occluded or not. Nothing in between. That was not my proposal at all, if you read it more carefully.

In the comments, Ferram4 talked about including something similar (though still not the same, admittedly) to what you're talking about in his experimental drag model. He said that the experiment already took so much computing power that adding it would be very much unfeasible.

Also (at risk of contradicting myself) FAR does do a relatively realistic job of replicating real-life occlusion, also similar to what you're talking about.

Additionally, Ferram4 created the raycast drag experiment to try and prove himself wrong, and that raycast drag could work realistically. He was unable to.

[Dodgey]

I don't think that intakes need to be changed, I see them in the same sense of part clipping, sure it's highly unrealistic but it's only a minor point that comes down to personal taste. It also makes up for limitations in the way ship building is put together.

The same with fuel flow (slightly off topic) and the yellow fuel pipe. Personally I would prefer it if we could draw a line between a fuel tank and an engine in the VAB and say fuel from that tank go to that engine. Regardless of line of sight or the parts inbetween.

[Sky_walker]

I don't think that intakes need to be changed, I see them in the same sense of part clipping, sure it's highly unrealistic but it's only a minor point that comes down to personal taste.

Edited June 23, 2014 by Sky_walker

[Dodgey]

Let me try that again as I realize I wasn't very clear. When I say personal taste I mean that in the same sense that intake spamming is personal taste. Some people have a problem with it others don't. As KSP isn't a competive multiplayer game it is up to the player wheather or not to exploit it, now let's be honest it doesn't really offer any huge advantage apart from if you are trying to get to orbit on insanely low amounts of fuel.

As to the second part I guess I was in the mindset of the second point that I made.

[Ippo]

Some people went so far as to defend the infamous wobble.

This. Wobbling, struts, space-chairs and all the other absurd things you can do are what makes the beta fun... but they need to go for the final release.

Any Battlefield player here? I was laughing mine off when these sort of things happened before relase...

... but leave them in the final retail version and I'm going to demand my money back.

Edited June 23, 2014 by Ippo
typo

[Dodgey]

With people defending the wobble it then gets down to what you think it is and what do you want it to be. How much of the game is defined by its wackyness and how much is defined by how realistic it is. How much is slapstick comedy and how much is rocket science.

[Ippo]

With people defending the wobble it then gets down to what you think it is and what do you want it to be. How much of the game is defined by its wackyness and how much is defined by how realistic it is. How much is slapstick comedy and how much is rocket science.

I remember my first launches some months ago. I was trying to get to the Mun and because I was terrible, I needed all the boosters I could get. So I had this central 1.25m stack surrounded by 8 solid boosters, because I was dumb.

And that's when I found out that the decouplers are so SOFT that 8 SRBs can literally CRUSH it and destroy my ship. I needed struts not to stop the wobble but to distribute the thrust away from the decoupler, because obviously 30 cms of alloy can be so easily crushed.

It is indeed comical, but seriously: just no. The same reasoning applies to the aerodynamic model: it surely is fun to be able to build boxplanes... but no. Because as they said above, the suspension of disbelief can only go so far.

[rkman]

How much of the game is defined by its wackyness and how much is defined by how realistic it is.

And how much of the wackyness is actually due to bugs or place holder code, in other words: unintended.

[Dodgey]

I don't that that the argument can be applied to aerodynamics, and of course there is a point where it becomes silly and the joints need to be fixed. In moderation it is fine, it can make launches more comical, but have too much of it and it's just silly. As with aerodynamics I think that they need to be fixed to be believable, the difference is that aerodynamics are much more fundamental than wheather or not rockets can wobble if not strutted correctly. Do you see the difference?

Thats a good point rkman, all I did was that I pointed out a reason that people would defend the wobbness.

[Ippo]

I don't that that the argument can be applied to aerodynamics, and of course there is a point where it becomes silly and the joints need to be fixed. In moderation it is fine, it can make launches more comical, but have too much of it and it's just silly. As with aerodynamics I think that they need to be fixed to be believable, the difference is that aerodynamics are much more fundamental than wheather or not rockets can wobble if not strutted correctly. Do you see the difference?

I was trying to draw a parallel here: rockets are not intended to wobble, but it was a glitch due to technical limitations in the Unity Engine. So sure, it was comical, but it was in no way intended to be there and therefore it's basically a bug.

And the same reasoning absolutely applies to aerodynamics: the aerodynamics were not designed in this way because comedy. It is actually just placeholder code from when the game was still much more primitive than it is right now and there were bigger issues to be dealt with.

Sure, it results in comical things (once more, boxplanes), and comicity is good... but it comes a point where you have to draw the line and a broken model must not make it into the final release, no matter how comical its bugs were. Have you seen the article about the QA process, where there is that video of random explosions in the VAB? It surely is fun... but it had to be squashed.

[Sky_walker]

It is actually just placeholder code from when the game was still much more primitive than it is right now and there were bigger issues to be dealt with.

I sure hope so. Cause seeing game released with... something like that.... would be rather.... disappointing.

[Red Iron Crown]

Stock aero is actually on its second iteration, it used to be like a solid wall at the edge of where vacuum meets atmosphere.

Here's hoping the third time's the charm.

[Dodgey]

I was trying to draw a parallel here: rockets are not intended to wobble, but it was a glitch due to technical limitations in the Unity Engine. So sure, it was comical, but it was in no way intended to be there and therefore it's basically a bug.

And the same reasoning absolutely applies to aerodynamics: the aerodynamics were not designed in this way because comedy. It is actually just placeholder code from when the game was still much more primitive than it is right now and there were bigger issues to be dealt with.

Sure, it results in comical things (once more, boxplanes), and comicity is good... but it comes a point where you have to draw the line and a broken model must not make it into the final release, no matter how comical its bugs were. Have you seen the article about the QA process, where there is that video of random explosions in the VAB? It surely is fun... but it had to be squashed.

The difference that I see with the comedic side of it is that wobbly rockets have become somewhat iconic, Kerbal if you will. the aerodynamic system is just terrible. I'm not saying that it is a feature, but I can see why somepeople would latch onto it. I mean sure box planes could be funny but aerodynamics have a much more substantial impact on gameplay than the amount that rockets wobble at this point in time. Since the latest update is it even a problem anymore?

I must say I did not intend to play devils advocit in this argument at all, oh well.

[Red Iron Crown]

Pancake rockets are just as "kerbal" as wobbly rockets; to some degree pancake is a response to wobbliness. Both need to go, IMO, though better joints and aero.

[Northstar1989]

pancake is a response to wobbliness.

PRECISELY. I never even THOUGHT of building a pancake rocket, until I started experiencing how wobbly rockets used to be (they're still more wobbly than I would like, though- the joint strengths still don't scale up enough with larger diameter parts- a 2.5 meter node should be at least 4 times as strong as a 1.25 meter node, due to having 4 times the cross-sectional area).

And, in fact, I have only built two or three pancake rockets EVER in KSP- and those were just test rockets, I never did actually figure out a good pancake design that wouldn't implode on the launchpad.

Instead, I installed NovaPunch2 as my very first mod, and started using tall 5 meter rockets with 4-6 radially attached 2.5 meter SRB's for my super-heavy lifting needs. It works perfectly fine, and it's actually realistic.

I've never relied on pancake rockets- the closest I ever came was a super-large (>12,000 ton on the Launchpad) StretchyTank SSTO Rocket that was both extremely tall *AND* wide, and still taller than wide- and that was more for lolz and scrap metal (I promptly recycled the rocket with a Scrapper Ship- I also was running Extraplanetary Launchpads and Orbital Construction by that point) than for real utility.

If experienced players like me can get missions to Sentar (from Krag's Planet Factory) without ever once having relied on pancake rockets (though plenty of mods), with the CURRENT aerodynamics model, there's no reason to think other experienced players couldn't learn to adapt, or that new players will in any way be phased by a more realistic aerodynamics model... (which would REWARD realistic rocket designs with reduced drag)

In fact, I was phased by the stock aerodynamics model right at the beginning- I've hated it since Day 1. That's why I installed Novapunch within a couple weeks of installing KSP for the first time- I didn't want to build pancake rockets, the very thought sickened me...

Regards,

Northstar

Edited June 24, 2014 by Northstar1989

[Northstar1989]

Stock aero is actually on its second iteration, it used to be like a solid wall at the edge of where vacuum meets atmosphere.

Also a great point

[Deathsoul097]

The difference that I see with the comedic side of it is that wobbly rockets have become somewhat iconic, Kerbal if you will. the aerodynamic system is just terrible. I'm not saying that it is a feature, but I can see why somepeople would latch onto it. I mean sure box planes could be funny but aerodynamics have a much more substantial impact on gameplay than the amount that rockets wobble at this point in time. Since the latest update is it even a problem anymore?

I must say I did not intend to play devils advocit in this argument at all, oh well.

I think that some people need to get off their high horses and accept that, no matter how hard they want it, KSP will not be a full on, realistic Space Sim. Yes, the Atmo needs fixing but some of the things that make KSP KSP are the silly things, the comedic things. I mean hell, if you work hard enough, you could quite probably build a boxplane in FAR. You would just have to have the ends open. Still a box plane, still silly. That said, Air intakes do need to be modelled a bit more realistically, and Jets need to be nerfed significantly if the atmo gets fixed. I don't want a mod like AJE becoming stock. I think it makes the game too complicated. Much more of a numbers game than kerbal is intended to be.

(And this is coming from a guy who avidly uses MJ for info and AP. And no, I can actually fly. I just have a crap laptop, so it lags. A lot. And MJ can fly through lag. Me? Not so much.)

That said, KSP is also realistic to a point. There needs to be a balance, and I think most users, especially those on the forums, are very biased to "#lolsokerbal" or "Its unrealistic! *Table Flip* Fix It!"

[Ippo]

I think that some people need to get off their high horses and accept that, no matter how hard they want it, KSP will not be a full on, realistic Space Sim. Yes, the Atmo needs fixing but some of the things that make KSP KSP are the silly things, the comedic things. I mean hell, if you work hard enough, you could quite probably build a boxplane in FAR. You would just have to have the ends open. Still a box plane, still silly. That said, Air intakes do need to be modelled a bit more realistically, and Jets need to be nerfed significantly if the atmo gets fixed. I don't want a mod like AJE becoming stock. I think it makes the game too complicated. Much more of a numbers game than kerbal is intended to be.

(And this is coming from a guy who avidly uses MJ for info and AP. And no, I can actually fly. I just have a crap laptop, so it lags. A lot. And MJ can fly through lag. Me? Not so much.)

That said, KSP is also realistic to a point. There needs to be a balance, and I think most users, especially those on the forums, are very biased to "#lolsokerbal" or "Its unrealistic! *Table Flip* Fix It!"

Nobody said we need a full on space simulator: but a game about science, technology and space travel needs to be at least remotely believable.

Also, about the box plane: if you take out the front and rear panels, then it's not a box anymore. It's more like a biplane, and biplanes do in fact fly.

[Vaporo]

...you could quite probably build a boxplane in far.

challenge accepted!

[SpaceSmith]

I didn't read all 6 pages because that's two much but I think we are all forgetting the absolute best solution.... Make it a clickable choice in the options menu. Stock dynamics or something realistic like FAR...everyone wins.

[MKI]

Updating the Aerodynamic model needs changing, but other things need changing first.

First things first are the spaceplane parts. All of them are dated and do not fit well with the current games form. Being the primary users of atmospheric flights and features, the part themselves should be updated before we even worry about how they perform.

Next there needs to be MORE parts added to allow a player to even play through any amount of Aerodynamic changes. Namely the addition of some form of faring need to be added.

Finally the actual addition of a new Aerodynamic model should be added. Personally i believe it should stick to being more unrealistic rather than realistic. For two reasons.

1. It prevents awesome and interesting designs that give the game a Kerbal Feel

2. Its a space game, about space ships, in space. Areodynamics are as much a part of the game as Kerbal EVA is. They aren't the focus, rather than a side element.

There will always be mods for ultra realism and extra difficulty. But making the game more difficult for the sake of making it more realistic isn't what KSP is really about. Yes it needs tweaking to make things a little less exploitative (flapping ships that fly?) but it in no way needs to become an engineering task. Especially for those who want nothing to do with it. (Rocket Flyers)