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

[Dunbaratu]

I don't think there'd be a big learning curve if there was a more realistic aerodynamics model because the current model is so oddball that it defies intuition. I think newbies would actually have an *easier* time adapting to the change than experienced players because it would be closer to what intuition is telling you.

Every time I show the game to a new person and they watch me doing a liftoff the first thing they comment on is "isn't a big ugly blunt nosed thing like that going to have huge drag? Why didn't you put nosecones on those boosters? I saw them in the parts bin when you were building the thing…"

The current state, where adding nosecones *decreases performance* is certainly not intuitive.

[r4pt0r]

I don't think there'd be a big learning curve if there was a more realistic aerodynamics model because the current model is so oddball that it defies intuition. I think newbies would actually have an *easier* time adapting to the change than experienced players because it would be closer to what intuition is telling you.

Every time I show the game to a new person and they watch me doing a liftoff the first thing they comment on is "isn't a big ugly blunt nosed thing like that going to have huge drag? Why didn't you put nosecones on those boosters? I saw them in the parts bin when you were building the thing…"

The current state, where adding nosecones *decreases performance* is certainly not intuitive.

that sums it up. a newcomer will know that rockets are long and pointy to pierce the atmosphere.

[Capt Snuggler]

that sums it up. a newcomer will know that rockets are long and pointy to pierce the atmosphere.

yup. good point well made.

[LethalDose]

Oops, double post.

[technicalfool]

I'll mention that physics (and parts) have already been given some tweaks along the way. Anybody who has gotten used to the 0.13 demo will probably be able to verify that getting flung into the 0.18+ versions of KSP is a rather different experience.

[UmbralRaptor]

Oh, hai. I'm one of the groganards who's been playing more or less continuously since 2011 that you're worried about. I'm actually completely fine with a switch to a more realistic aerodynamic model. My main concerns would be things we can expect Squad (or at worst the modder community) to address:

1) Will the VAB (and/or inflight instrumentation) give me the tools to understand the aerodynamic effects?

2) Will I have the parts to properly deal with it? (I think this pretty much means something like pFairings)

3) Will sufficient consideration be given for how this alters the difficulty? (Currently this means more care is required in ascent paths and structural strength, but mass ratios are pushed from ~3-5 to below e.)

I like building stupid asparagus rockets. I like the idea of building "real" rockets. I'm down for either.

I just want to build rockets

This is pretty much my attitude. I'm not worried about losing pancake rockets (and expect a more realistic model to leave some parallel staging as perfectly viable. If it works for the various R-7s, Atlases, Deltas, Titans, STS, and Energia...)

[robopilot99]

I don't care about a perfect aerodynamic model. I think a good middle ground could be struck with a simple aerodynamic model that's also believable.

[Sky_walker]

I don't think there'd be a big learning curve if there was a more realistic aerodynamics model because the current model is so oddball that it defies intuition. I think newbies would actually have an *easier* time adapting to the change than experienced players because it would be closer to what intuition is telling you.

Every time I show the game to a new person and they watch me doing a liftoff the first thing they comment on is "isn't a big ugly blunt nosed thing like that going to have huge drag? Why didn't you put nosecones on those boosters? I saw them in the parts bin when you were building the thing…"

The current state, where adding nosecones *decreases performance* is certainly not intuitive.

An excellent post. Agreed with every single sentence!

I don't care about a perfect aerodynamic model. I think a good middle ground could be struck with a simple aerodynamic model that's also believable.

Noone, not even a single person in this entire thread, asked for a "perfect areodynamic model". Just one that makes at least a little bit of sense.

[Jouni]

I don't care about a perfect aerodynamic model. I think a good middle ground could be struck with a simple aerodynamic model that's also believable.

FAR has such simple and believable model. It's far from perfect, and sometimes the results are really ridiculous (for example, try atmospheric reentry just behind a big asteroid), but it works well most of the time.

[Red Iron Crown]

I've been playing for a long time, and I'd like better aerodynamics and reentry heat. They're the same feature in my mind, a good aero model will include frictional heating.

An iterative implementation like the OP suggests is a bad idea, IMO. As others have mentioned, it will break designs and require adaptation to the new model at each iteration. Better to do it all in one go, like tearing off a band-aid. One large adjustment to gameplay is better than many smaller ones.

Better aero would be a boon for new players; currently rockets that look like real rockets are inefficient and less effective than pancakes, which don't look feasible at all and aren't an obvious design to newbies. It's a boon for older players like me as it would add a fresh challenge and new things to learn. As far as I can see it's a win for everyone. Of course some people are going to complain, but that is true no matter what course development follows. Haters gonna hate.

Edited June 20, 2014 by Red Iron Crown
Typo

[marach]

I currently don't use FAR because I'm not sure what to what extent it is compatible with MechJeb*, but should improved aerodynamics in the stock game render my current crafts invalid, I'll simply learn to adapt.

*My play style is more of a mission planner and infrastructure manager, rather than an ace pilot

MechJeb doesn't get FAR's Atmosphere density data so you lose the super precise landings and you have to play with the flight path on ascent guidance to get a full auto pilot for launch. But other than that it /mostly/ just works. There is a patch made by Sarabian that lets mech know the FAR data out there though.

[Whirligig Girl]

My problem with fully realistic aerodynamics is that it isn't fun. I don't want to have to very carefully keep my ship almost exactly on the prograde without flipping out irrecoverably. I need to do launches as much as ten times to get one out of the atmosphere with FAR, especially RSS/64:10 Kerbin.

It seems just purely un-fun to have ships spin out like that. I'm all for fairings and nosecones working and all that jazz, but flipping out, while it may be realistic, is not fun. KSP is not realistic, it's Authentic. Authentic means fun+some realism.

[Sky_walker]

Again: Noone in an entire topic asked for perfect, fully realistic aerodynamics.

Just something that would make sense (eg. nose cones would be beneficial instead of just crippling your spaceship) and be more believable (eg. no rocket-powered chairs).

[meve12]

2) Will I have the parts to properly deal with it? (I think this pretty much means something like pFairings)

Can we have PWings with that? And a 2.5m fuselage?

[Greenfire32]

I think the best way to handle it would be if Squad advertised that the current areo model is going to change with version x.x

That way, people would know that it's coming and not get visibly upset that it was "forced" on them.

[m4v]

The reason being that the current aerodynamic model is artificially soupy in order to create an additional ~1km/s of "difficulty" getting to orbit. FAR's drag model works quite well at a realistic scale, like with RSS.

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, I can do cartwheels with my rocket and still reach orbit with less fuel consumed than in my best stock launches.

The only thing I can see Squad improving is making drag dependant of the cross-section and angle of attack instead of just the part mass. But the "soup physics" will have to stay as long as kerbin stays at 1/10 scale.

[regex]

But the "soup physics" will have to stay as long as kerbin stays at 1/10 scale.

This can be solved by either doubling the size of the system and retaining the current surface gravities (density still ends up being stupid) which means a 75km orbit around Kerbin will then have a velocity of 3.32km/s instead of stock's 2.28km/s, or by decreasing atmospheric isp ala KIDS. Personally, increasing the size of the solar system would be a much better option; even double the size would make the game "feel" that much better, but it affects every body. OTOH, decreasing isp would be easier since you wouldn't have to rebalance everything else.

Of course you could also address KSP's ridiculous notion of isp affecting fuel flow rather than thrust, which might actually solve some of the ease issues (I'd have to test).

[coolitic]

Use the farram aerospace mod. Honestly though, that should become stock.

[Alshain]

Use the farram aerospace mod. Honestly though, that should become stock.

Been using the Ferram Aerospace Mod for just a little bit recently. I have successfully put planes into orbit and landed them on the runway using FAR. Despite this I feel it should never ever become stock. The truth is, it just adds tedium to the process that is going to anger a lot of people. I enjoy building the planes, but trying to make sure the wings are withing a millimeter of the perfect to get some obscure derivatives of which I don't understand because I'm not really an aeronautics engineer to turn green is just annoyingly frustrating. It can't possibly be that difficult and still be called a "game". I'm sure some people can figure it out, but I want to launch space crafts, not spend all my limited playing time in the SPH. So in short, FAR is good for some people but if Squad were to require it, KSP would be destroyed.

I'm not saying the stock model can't be improved, but it can't be FAR either.

Edited June 20, 2014 by Alshain

[m4v]

This can be solved by either doubling the size of the system and retaining the current surface gravities (density still ends up being stupid) which means a 75km orbit around Kerbin will then have a velocity of 3.32km/s instead of stock's 2.28km/s, or by decreasing atmospheric isp ala KIDS. Personally, increasing the size of the solar system would be a much better option; even double the size would make the game "feel" that much better, but it affects every body. OTOH, decreasing isp would be easier since you wouldn't have to rebalance everything else.

Of course you could also address KSP's ridiculous notion of isp affecting fuel flow rather than thrust, which might actually solve some of the ease issues (I'd have to test).

Using a real scale solar system will make launches take at least 15-20 minutes long, this is a game, fun comes first, realism later.

Nerfing the rockets, while i'm not opposed, in essence is just another way of doing the same, you either end with unrealistic drag or unrealistic rocket performance (not saying that current rockets are realistic though)

edit: So you have to make a compromise either way, I would rather stick to the soup physics than changing the rockets, the result will be the same and you avoid drastic changes in the game.

Edited June 20, 2014 by m4v

[regex]

Using a real scale solar system will make launches take at least 15-20 minutes long, don't forget that this is a game, fun comes first, realism later.

I'm not talking about a real scale solar system I'm talking about double stock's size, which is still @ 1/5th the size of our solar system. And no, actual, real life rocket launches don't take 15~20 minutes; a Soyuz TMA, for example, takes about 9 minute to reach LEO. Doubling Kerbin's size wouldn't radically affect launch times.

Nerfing the rockets, while i'm not opposed, in essence is just another way of doing the same, you either end with unrealistic drag or unrealistic rocket performance (not saying that current rockets are realistic though)

The drag and aerodynamic model needs to change, regardless. Building a spaceplane (hell, any plane) is a completely unintuitive process that results in planes that shouldn't be able to fly. The launch profiles for rockets in KSP is downright stupid and equally unintuitive. With a better model players can draw upon real life examples of how to do things. FAR isn't the answer for stock, obviously, but stock does need to change.

[r4pt0r]

I'm not talking about a real scale solar system I'm talking about double stock's size, which is still @ 1/5th the size of our solar system. And no, actual, real life rocket launches don't take 15~20 minutes; a Soyuz TMA, for example, takes about 9 minute to reach LEO. Doubling Kerbin's size wouldn't radically affect launch times.

The drag and aerodynamic model needs to change, regardless. Building a spaceplane (hell, any plane) is a completely unintuitive process that results in planes that shouldn't be able to fly. The launch profiles for rockets in KSP is downright stupid and equally unintuitive. With a better model players can draw upon real life examples of how to do things. FAR isn't the answer for stock, obviously, but stock does need to change.

im ok with all of what he said.

[Kasuha]

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.

XF-85 Goblin

Stipa-Caproni

Guppy

Bumble Bee II

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.

[xcorps]

I don't care about the aerodynamics model at all.

I think it's not a very good idea to have an option to switch between what we have now and FARish if it gets changed. Either don't change it, or change it and make it stick.

What other "choke points" do you see in the refinement of various subsystems from placeholder to complete feature?

We really need an improvement to the whole science thing. A tech tree is a fine thing, unlocking new parts as your program matures. However, using random experiments like grabbing some dust from Minmus to develop a new engine is just...bleh. It's like..go to Minmus 5 times and click on 10 experiments and you get Mainsails. Science should be important and should certainly be a primary goal in the game, but using it as an arbitrary scoring system for unlockables is doing a disservice to the game.

This isn't Kerbal Rocket Game. It's Kerbal Space Program. Time to add some Space Program management instead of part management.

[TheCanadianVendingMachine]

-snip-

We really need an improvement to the whole science thing. A tech tree is a fine thing, unlocking new parts as your program matures. However, using random experiments like grabbing some dust from Minmus to develop a new engine is just...bleh. It's like..go to Minmus 5 times and click on 10 experiments and you get Mainsails. Science should be important and should certainly be a primary goal in the game, but using it as an arbitrary scoring system for unlockables is doing a disservice to the game.

This isn't Kerbal Rocket Game. It's Kerbal Space Program. Time to add some Space Program management instead of part management.

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.