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

[xcorps]

Sorry for going off topic but - I believe you will need to buy each part in the tech tree once unlocking it. You can see this by when you install a mod, then you have to click those nodes. It says how much currency it is.

I realize that, but I'm glad you pointed it out for folks reading this that don't.

But it isn't really a solution. It's just more unlocking. My point is that science shouldn't be tied to unlocking stuff. Science should progress your program, but it shouldn't be through parts. I'm really ambivalent about the proposed currency system, because it seems to me to be more of the same, just convoluted for the sake of convolution. Buying rocket parts and all is fine, and being able to convert science points to cash is a step in the right direction, but it's really not anything different than what we have now. Maybe Squad has a nice curve ball included. I'm hoping so.

[regex]

In our criticism of KSP aerodynamics, I believe we shouldn't forget that real aerodynamics don't exactly exclude things that can be called ridiculous.

Those things are still aerodynamic.

In my personal opinion, most KSP planes are much more plane-like than these even though they use stock aerodynamics. If you put parts together so they look like plane, it will probably fly. Maybe not well but that's the matter of moving them around a bit and watching a few simple rules. You can build ridiculous and completely non-plane-like crafts, but nobody forces you to do so.

KSP is dumb in that you can do things like this:

Javascript is disabled. View full album

This thing flew at 176m/s (633km/h, or 396mph), has intakes that barely touch the path of flight and have no reasonable connection to the engine, and has the aerodynamic properties of a brick wall. I brought it in for a very sedate landing and parked it with no problems. Now, I get that this is a wonderful sandbox and everything, and no one is out there "forcing" you to do things, but there comes a point where a game trying to act like it has at least some connection to reality should draw the line.

[Superfluous J]

KSP is dumb in that you can do things like this:

That is one of the coolest planes I've ever seen. You just invalidated your own argument

[BudgetHedgehog]

For anyone unconvinced, please can we not forget that Manley made a boxplane?

[r4pt0r]

my earliest experience in the SPH was trying to make an airfoil out of structural panels. pretty upset when i found out only certain parts generate lift.

[Justin Kerbice]

KSP is dumb in that you can do things like this:

Call it dumb (I mostly agree) if you want, but don't forget KSP is... KSP it is supposed (I guess) to be fun !

Such "plane" is quite a good definition of fun.

People who want FS (or Flightgear or anything else), they can play FS, why willing to get a KSP-FS hybrid for example ?

It may become boring very fast if you have to pay any respect to many physics rules as IRL, without forgetting the real puzzle challenge (I guess) made by assembling a rocket, think of everything like electric/electronic limitations & constraints, heat dissipation/insulation, cryogenic fuels needs for safe storage, pumping, flowing, heating, ... add the need of shield against deadly radiations and heat/cold in deep space, the reentry procedure, ... and the fact that's a single little tiny scratch on a space-suit could means fast death, how could you jump like crazy on Mun or Minmus, bouncing around with your kerbals, now play space soccer, etc ?

[Dodgey]

I don't see what any of the has to do with having a better aerodynamic system that makes sense more than the current one.

[BudgetHedgehog]

<snip>

And then there's such a thing as 'suspension of disbelief'. I'm quite happy with not having to wire the rocket myself because I'm assuming the wiring in the parts is done and connects automagically. Same with the insulation/pumping/heating of fuel - I assume the tanks and engines come with that pre-built into them (the Jumbo 64 is orange because it has the insulation).

I'm perfectly happy with some of the poetic licenses used by KSP such as the above. It doesn't break my immersion nor is it unbelievable. But stock aerodynamics laugh in the face of suspension of disbelief and when I can make a boxplane (like the picture I posted), the immersion isn't just broken, it's completely shattered. It is completely unintuitive and should simply NOT fly. Not on Kerbin nor Earth or anywhere with an atmosphere. And yet, it's apparently, a perfectly valid excuse for a plane in stock KSP.

This isn't about realism or believability, it's about suspension of disbelief and stock aerodynamics break immersion constantly.

[r4pt0r]

automagically

that is a brilliant word.

i just want ksp to know that pointy is good, flat is bad. they dont need to take the time to turn it into microsoft flight sim x here, just improve it. alot.

[Sky_walker]

Such "plane" is quite a good definition of fun.

Nope, such "plane" is a definition of fail.

It may become boring very fast if you have to pay any respect to many physics rules as IRL

It's like saying that Kerbal is boring because you need to obey laws of gravity.

This isn't about realism or believability, it's about suspension of disbelief and stock aerodynamics break immersion constantly.

THIS.

Edited June 21, 2014 by Sky_walker

[regex]

This isn't about realism or believability, it's about suspension of disbelief and stock aerodynamics break immersion constantly.

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.

[rkman]

Which is why we can't have a realistic aero model in Kerbin, reaching orbit with FAR is so ridiculously easy that it breaks the illusion of it being realistic,

There is a fix for that: Isp Difficulty Scaler. Essentially it reduces engine efficiency. So, "yes we can".

Some people seem to have a problem with almost anything that might change about ksp, except maybe for the inclusion of their own pet feature. Some people went so far as to defend the infamous wobble.

But given that the game isn't finished, there's bound to be a lot of changes between now and whenever it is going to be finished, including rebalancing which could deal with game balance problems arising from a more realistic drag model.

[m4v]

There is a fix for that: Isp Difficulty Scaler. Essentially it reduces engine efficiency. So, "yes we can".

I know about KIDS, but I regard it as ugly workaround to a problem caused by FAR insisting in applying Earth's model rigorously and without compromises. If the atmosphere of Earth were compacted to the space that it occupies in Kerbin, then it would more dense, and so the "soup physics" shouldn't be so ridiculous as some people make it out to be.

Whichever if you have unrealistic underpowered engines or unrealistic atmosphere density, the result is the same, and I don't see why the aero model is so holy that you have to take the hacky path of tweaking all the engines (including the ones from mods) instead of just trying to fit the model to Kerbin.

Edited June 21, 2014 by m4v

[Red Iron Crown]

I don't think it would be the end of the world if the dV requirement to make orbit dropped as a result of better aero. Sure, it would require a rethink of how we make orbit, but that's going to happen with a new aero model anyway.

[BudgetHedgehog]

Let me just say I have no problem with stock atmospheric density, it's the fact that boxplanes are possible.

[Jouni]

I know about KIDS, but I regard it as ugly workaround to a problem caused by FAR insisting in applying Earth's model rigorously and without compromises. If the atmosphere of Earth were compacted to the space that it occupies in Kerbin, then it would more dense, and so the "soup physics" shouldn't be so ridiculous as some people make it out to be.

Kerbin's atmosphere is perfectly fine. It's pretty much the same, no matter whether we use FAR or stock aerodynamics.

The real problem is that KSP is currently a game about rocket-powered balloons. Everything we build shrinks as it burns fuel, and then inflates again when we refuel it. Staging splits a big balloon into smaller balloons that are surprisingly denser than the original balloon. When you dock two balloons together, you get a new balloon that's bigger than the originals together. If this still makes some sense, add parts with non-standard drag coefficients to the mix, and the balloons start behaving in strange ways.

Stock aerodynamics just don't make any sense. The model is full of game-specific knowledge that defies intuition and is disconnected from everything new players might already know about rockets, planes, and aerodynamics. The forums are full of newbie questions about why their rockets initially fly fine, but then suddenly flip over at some point of the ascent. These questions often get a lot of nonsense answers, before someone who understands stock aerodynamics explains what's going on.

[Superfluous J]

automagically

that is a brilliant word.

Agreed. I even added it to my dictionary so when I type it, Chrome doesn't complain.

i just want ksp to know that pointy is good, flat is bad. they dont need to take the time to turn it into microsoft flight sim x here, just improve it. alot.

This is the best way I've seen the argument presented. Somewhere between what we have and FAR is the Pointy End First model and that's the model I want

[m4v]

Kerbin's atmosphere is perfectly fine. It's pretty much the same, no matter whether we use FAR or stock aerodynamics.

The real problem is that KSP is currently a game about rocket-powered balloons. Everything we build shrinks as it burns fuel, and then inflates again when we refuel it. Staging splits a big balloon into smaller balloons that are surprisingly denser than the original balloon. When you dock two balloons together, you get a new balloon that's bigger than the originals together. If this still makes some sense, add parts with non-standard drag coefficients to the mix, and the balloons start behaving in strange ways.

Stock aerodynamics just don't make any sense. The model is full of game-specific knowledge that defies intuition and is disconnected from everything new players might already know about rockets, planes, and aerodynamics. The forums are full of newbie questions about why their rockets initially fly fine, but then suddenly flip over at some point of the ascent. These questions often get a lot of nonsense answers, before someone who understands stock aerodynamics explains what's going on.

I'm not arguing against that. My original post was that stock aero should make drag depend of the shape and angle of attack of parts (and not their masses), but that the exaggerated atmospheric friction should stay the same.

[Jouni]

I'm not arguing against that. My original post was that stock aero should make drag depend of the shape and angle of attack of parts (and not their masses), but that the exaggerated atmospheric friction should stay the same.

That would make parachutes almost obsolete. If we want a rocket to lose around 1000 m/s to drag during the ascent, its terminal velocity should be around 100 m/s at sea level. The returning part of the payload probably has similar cross section, bigger drag coefficient (assuming that it descends flat end first), and at least an order of magnitude less mass. That can easily bring the terminal velocity down to 10-20 m/s, which is survivable.

[m4v]

That would make parachutes almost obsolete. If we want a rocket to lose around 1000 m/s to drag during the ascent, its terminal velocity should be around 100 m/s at sea level. The returning part of the payload probably has similar cross section, bigger drag coefficient (assuming that it descends flat end first), and at least an order of magnitude less mass. That can easily bring the terminal velocity down to 10-20 m/s, which is survivable.

Fine then. I'll keep on playing stock, boxplanes don't ruin my fun as much as frictionless atmospheres do.

[Dunbaratu]

I'm not arguing against that. My original post was that stock aero should make drag depend of the shape and angle of attack of parts (and not their masses), but that the exaggerated atmospheric friction should stay the same.

That's actually a much more complex calculation - to use the shape of the object to decide. The simpler fix is to just give each part a coefficient of drag number (like it does now), and hardcode the pointier parts to a smaller drag, and then implement an occlusion calculation into the drag model - part B being *behind* part A in reference to the direction of motion would mean that part B's current coefficient of drag is reduced to being nearer to part A's lesser value. (perhaps with a weighted average of the two, favoring A's drag over B's drag). Then part C, behind part B, gets the same calculation using B's new lesser drag number compared to itself.. etc. That would make it so that putting a lower coefficient of drag part (like a nose cone) atop a stack of higher drag parts would "inherit" the lesser drag value down to parts in the stack, with a falloff in how much it lessens them the farther from the nosecone you get. That would fix the broken model most of the way, without being too expensive of a calculation to make at runtime. You could even precalculate most of it and only recalculate it when the part count changes (i.e. staging or bits breaking), reducing the runtime calculation to a simple ( B + A*C ), where B is the base coefficient of friction you precalculated if the vehicle was oriented straight, A is the angle of how far off the craft is from traveling "straight" and C is a factor you precalculated for how much extra drag you get for not going straight.

It's not as realistic as a full flight-sim but its a far step better than what's there now.

[AlamoVampire]

As much as I like this subject, and would like to discuss about it, it is annoyingly on the WNTS list, so its probably gonna get closed...

just because its a what NOT to suggest item, doesnt mean it will be closed automatically. they may keep this going so they can crowd source our thoughts on the idea, as it doesnt seem like this thread is a SUGGESTION but a DISCUSSION. I for one would LOVE a more realistic aerodynamic model for the game <heat damage as well>. yea i could use DRE and FAR for this, but, I like to keep my mod count smallish. I use 4 mods and I get most of what I want to see from those.

Edited June 21, 2014 by AlamoVampire

[SkyRender]

I did not expect this thread to go on this long... Although the massive discussion that's resulted only reinforces my suspicions that changes to aerodynamics are going to be met with dissent no matter how they're handled. I've observed several different fronts on how things will be done and how they'll turn out, and not a lot of concessions going around on any side. Then again, it is largely a matter of opinion, so what can you do? Perhaps we can redirect the topic to the question I had hoped we'd get into: other placeholder systems that are going to run into some player backlash when they get polished.

[Vaporo]

That's actually a much more complex calculation - to use the shape of the object to decide. The simpler fix is to just give each part a coefficient of drag number (like it does now), and hardcode the pointier parts to a smaller drag, and then implement an occlusion calculation into the drag model - part B being *behind* part A in reference to the direction of motion would mean that part B's current coefficient of drag is reduced to being nearer to part A's lesser value. (perhaps with a weighted average of the two, favoring A's drag over B's drag). Then part C, behind part B, gets the same calculation using B's new lesser drag number compared to itself.. etc. That would make it so that putting a lower coefficient of drag part (like a nose cone) atop a stack of higher drag parts would "inherit" the lesser drag value down to parts in the stack, with a falloff in how much it lessens them the farther from the nosecone you get. That would fix the broken model most of the way, without being too expensive of a calculation to make at runtime. You could even precalculate most of it and only recalculate it when the part count changes (i.e. staging or bits breaking), reducing the runtime calculation to a simple ( B + A*C ), where B is the base coefficient of friction you precalculated if the vehicle was oriented straight, A is the angle of how far off the craft is from traveling "straight" and C is a factor you precalculated for how much extra drag you get for not going straight.

It's not as realistic as a full flight-sim but its a far step better than what's there now.

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

[Superfluous J]

... my suspicions that changes to aerodynamics are going to be met with dissent no matter how they're handled.

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.