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

[SkyRender]

As Kerbal Space Program approaches scope completion, we are coming upon a time when existing systems which were left in a prototypical state are going to be revisited and revised. This has the potential to be one of the most awkward and upsetting parts of development, for both the devteam and the players, because of one simple truth: we are used to the systems that don't work right.

Let's take the biggest one, the elephant in the room, aerodynamics. The current aerodynamics model in KSP is decidedly off in how it works. There are mountains of problems with it, and completely un-physical side-effects are possible as a result. Fixing this placeholder engine has always been on the list of priorities, but here's the thing: fixing it has the very real potential to make most players of KSP very, very angry.

Realistic drag modeling will cause many of the most popular rocket structures to be completely infeasible. The lack of stock fairings will severely limit how craft can be designed, and the addition of them will serve to cause even more problems as players have to learn not only how to deal with a new aerodynamics model, but also how to design their craft in an entirely new context. Just look at the results of pretty much anyone getting into the Ferram Aerospace Research mod without knowledge of how aerodynamics actually work to get an idea of just what we have to look forward to. Now picture that on a scale of every single user being switched over to using FAR, like it or not. It's not pretty, is it?

How on Kerbin are we going to deal with a change of this scale, then? The only real solution is a gradual implementation, making the aerodynamics model slightly better with each new release for several versions until old and new players alike are comfortable with having to deal with aerodynamic design principles. New indicators will have to be added. We'll probably have to have a "wind tunnel simulator" option at some point. A lot has to change in small steps for this feature to reach a point where it is both acceptably realistic and accepted by the player base as a whole. It can be done, but going straight from the current model to FAR-level aerodynamics modeling in one version is the worst possible way to go about it.

There are other system upgrades and changes that are going to cause upsets if not done gradually as well, of course. What other "choke points" do you see in the refinement of various subsystems from placeholder to complete feature?

[Eric S]

Having a preference for playing with FAR, I think you overstate it a bit (at least for rockets), though I do agree that more realistic aerodynamics is probably going to be the biggest relearning-causing improvement.

My normal rockets mostly just required nosecones on radially attached stacks with FAR even with my more aggressive asparagus-staged launchers. Yes, when you get into asparagus pancakes that are wider than they are tall, there will be issues, but I don't think those are really as common as some people seem to think.

When I'm balancing a large, non-aerodynamic rover on top of a rocket, yes, I have to pay a lot more attention.

[RidingTheFlow]

I think increased difficulty due to aerodynamics could be actually good thing.

Because with gradually improving joints/physics over time KSP rockets became a lot less affected by wobble/breakups/spontaneous deconstruction - which was originally the major difficulty factor ("need more struts"). Making rockets not to tumble/break up due to aerodynamics forces could take it place, and it arguably should be more interesting/educational.

[Random Tank]

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...

[regex]

here's the thing: fixing it has the very real potential to make most players of KSP very, very angry.

So what? You gotta break a few eggs to make an omelet.

Realistic drag modeling will cause many of the most popular rocket structures to be completely infeasible.

Yes, but people will then be able to build rockets that look like rockets. They will work like we expect them to in real life and new players can draw on examples that actually exist.

The lack of stock fairings will severely limit how craft can be designed, and the addition of them will serve to cause even more problems as players have to learn not only how to deal with a new aerodynamics model, but also how to design their craft in an entirely new context.

This is only an issue for old-school players who expect a completely unrealistic and terrible drag model. See above.

Just look at the results of pretty much anyone getting into the Ferram Aerospace Research mod without knowledge of how aerodynamics actually work to get an idea of just what we have to look forward to. Now picture that on a scale of every single user being switched over to using FAR, like it or not. It's not pretty, is it?

It's not pretty because players are introduced to a complete pile of crap before finding FAR. They expect craft to behave in ways they shouldn't because KSP teaches them otherwise. If new players can draw upon real life examples of what should fly and reasonably expect that it will, they won't have to figure out the pile of crap and then deal with the "culture shock" of a better model.

How on Kerbin are we going to deal with a change of this scale, then?

Just do it. Seriously, the only people it's going to hurt are the grognards who expect something completely unrealistic. Newbies can draw upon real world examples and thrive, old hands at the game get new challenges, and the grognards get something new to complain about.

At the end of the day it's not like we're just going to get FAR anyway; the stock implementation will likely be far more forgiving.

[r4pt0r]

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...

then lets quickly discuss. i think OP raises alot of good points, and a good solution in upgrading aero in steps. but stock fairings do need to come at some point.

edit: actually, this is a discussion not a suggestion so it should be just fine.

Edited June 18, 2014 by r4pt0r

[Kasuha]

One thing is, KSP aerodynamic model is IMO not all that bad. I can imagine it staying with the final game. Yes it is unphysical in many obvious aspects but KSP is a space discovery game, not flight simulator. For many purposes it is realistic and physical enough, including spaceplanes.

Another thing is, if Squad replaces the aerodynamics model, it will hurt, but it will only hurt once. Yes, many designs will be rendered useless but if it's done right and fast, people will stop complaining soon and will start enjoying new, more realistic physics. Of course there will be players who will keep posting about how they're staying with old KSP because they prefer the old model but such thing comes with each change.

I also believe many players understand that current aerodynamics model is a placeholder and is going to be replaced.

It can be done, but going straight from the current model to FAR-level aerodynamics modeling in one version is the worst possible way to go about it.

No, it is the best possible way. Because if you do it iteratively, you will break designs with each release over multiple releases . If you go straight from one to another, you'll do it just once.

Edited June 18, 2014 by Kasuha

[Jouni]

Just look at the results of pretty much anyone getting into the Ferram Aerospace Research mod without knowledge of how aerodynamics actually work to get an idea of just what we have to look forward to. Now picture that on a scale of every single user being switched over to using FAR, like it or not. It's not pretty, is it?

In my experience, FAR is the easy mode for rockets. I switched to it a few days ago, and noticed two main differences to stock aerodynamics:

1. Rockets need more babysitting during the ascent, because you have to start the turn almost immediately.
   
2. I used to build rockets with two boosters, a lower stage, and an upper stage. Now I replace the upper stage with more payload, because the payload fractions are so ridiculous.
   

[regex]

In my experience, FAR is the easy mode for rockets.

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.

[Superfluous J]

I feel like I'm the easiest guy to please sometimes.

I like the current aerodynamics model. It lets me launch rockets and fly planes in a fashion that sufficiently balances between ease, difficulty, and realism. Is it perfect? No. Is it horrible? Not even close.

I like the concept of Ferram's aerodynamics model. I've not used it mostly because most of my interaction with it is in videos where people bring up GUIs with so many readouts and buttons that my old college graphing calculator shudders in the corner. That said, if it was forced on me I'd buckle down and figure it out.

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

[Random Tank]

I feel like I'm the easiest guy to please sometimes.

I like the current aerodynamics model. It lets me launch rockets and fly planes in a fashion that sufficiently balances between ease, difficulty, and realism. Is it perfect? No. Is it horrible? Not even close.

I like the concept of Ferram's aerodynamics model. I've not used it mostly because most of my interaction with it is in videos where people bring up GUIs with so many readouts and buttons that my old college graphing calculator shudders in the corner. That said, if it was forced on me I'd buckle down and figure it out.

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

I feel EXACTLY the same way; I even have saves with and without FAR because... well.... I don't mind the stock aero model, but I like FAR's... Even for building planes, I don't mind either version. But in the mind of a 'mainly realistic rocket simulator' that this game is trying to be, it does need changing; as for the slightest bit of realism you want rockets to look like... well rockets really. I don't mind when it happens though!

[the_bT]

I'm just going to say this: I have no background in aerodynamics, I'm mildly interested in the subject but that's about it.

I switched to FAR.

I discovered that none of my existing crafts could take off.

I then build new crafts making them look similar to real aircraft and they flew very well. I never looked back.

It might be related to installing B9 at the same time and thus having many new parts to learn that masked the transition...

So in conclusion: I don't think that a new aerodynamics model that is as detailed as FAR is will alienate anyone. At least not if all parts get rebalanced at the same time as well. But that is just common sense right?

Also the GUI things of FAR are nice and all but I mostly ignore them and do well enough

[MR4Y]

Well, this is assuming the aerodynamics will be redone, which is highly unlikely.

[Dodgey]

Why is it unlikely that the aerodynamics would be redone??

[LethalDose]

Why is it unlikely that the aerodynamics would be redone??

IMHO, I don't think the aerodynamics will be redone for the exact reasons that the OP raises. Primarily, players will get angry.

I could make pointed remarks about the way Squad deals with issues like this, but I don't see the point anymore.

[Dodgey]

I don't the issues raised as issues at all. Rocket designs come and go as it is already. Changing the aerodynamics would just cause people to design new rockets, I don't see any reason why new aerodynamics would have to be as unforgiving as FAR, the OP said that a solution would be to implement it over several releases, if this could be done than why wouldn't a more forgiving system couldn't be designed.

Also the lack of fairings in the stock game isn't an argument for anything. It's like saying re-entry heat shouldn't be implemented (another discussion for another time) because there aren't any heat shields in the game.

[sumghai]

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

[Vaporo]

You know, we could have both aerodynamic models. Just put a slider/check box in the settings to enable realistic (FAR-like) or simplified (like what we have now) aerodynamics. That way, we can have realistic aerodynamics AND aerodynamics that allow for crazy unrealistic Whackjob-esque designs, depending on what each player wants.

That's how I would go about it.

Of course, then you have problems with people designing rockets in simplified mode, then sharing them with others who play in realistic mode and discover that the designs don't work for them. It would create a minor divide in the community.

Edited June 19, 2014 by Vaporo

[sumghai]

You know, we could have both aerodynamic models. Just put a slider/check box in the settings to enable realistic (FAR-like) or simplified (like what we have now) aerodynamics. That way, we can have realistic aerodynamics AND aerodynamics that allow for crazy unrealistic Whackjob-esque designs, depending on what each player wants.

That's how I would go about it.

Of course, then you have problems with people designing rockets in simplified mode, then sharing them with others who play in realistic mode and discover that the designs don't work for them. It would create a minor divide in the community.

A better option would be to have the improved aerodynamic models be the default, such that all active players adapt to the new system (just as they adapted to the new SAS). The option for reverting to the old aero model should only be available in the Alt-F12 debug menu - it is, after all, a cheat option.

[Vaporo]

A better option would be to have the improved aerodynamic models be the default, such that all active players adapt to the new system (just as they adapted to the new SAS). The option for reverting to the old aero model should only be available in the Alt-F12 debug menu - it is, after all, a cheat option.

That may be a better option, but it could be difficult to implement as the aerodynamics are a core game function, and simply "swapping out" realistic aero for simplified aero could be hard on the game's code. (Which I don't know anything about, by the way. Feel free to contradict me.)

[Sky_walker]

and the addition of them will serve to cause even more problems as players have to learn not only how to deal with a new aerodynamics model, but also how to design their craft in an entirely new context

Oh horrors. They had to learn how to get into the orbit - they'll learn how to build rockets that don't look like a monster strapped to cartoon-alike set of fireworks.

Current aerodynamics model is pretty much this:

Seriously, if I'd be about to point out just one thing that bothers me with Kerbal - it's areodynamics. Everything else is either great or good enough. But areodynamics are horrible.

And that obviously is a problem not only with going up the atmosphere and space planes, but also descending down through the admosphere and how pretty much every shape and kind of ship can get into the atmosphere without even damaging it's structural integrity, not to mention breaking up completely.

Luckily - it seems like both: Better aerodynamics and re-entry heat&damage are on a list of planned features. It only saddens me that they picked to implement it after the multiplayer - mode which equally well could be kept as a mod, just like it started, while the core mechanics of the game remain waiting :/ :/ :/.

IMHO, I don't think the aerodynamics will be redone for the exact reasons that the OP raises. Primarily, players will get angry.

For those that will get angry - give them an option to use the old model, preferably like sumghai suggests.

Edited June 19, 2014 by Sky_walker

[Pulstar]

They'll get confused and maybe angry for a day or week at most, after that everybody will get used to it. Save wiping is more likely to cause outrage than having to redesign lifters, and anybody that actually likes building lots of planes probably uses FAR already and wants the aerodynamics changes to happen anyway. Doesn't seem like a very large group to me, especially considering that parft of the appeal of the game is realism.

I wouldn't worry too much about this. You spend most of the game outside of the atmosphere, people don't get attached to lifters much since they're one use only and everything you leave in space will remain unchanged.

[Capt Snuggler]

I don't currently use FAR. i installed it once. my current aircraft did not work and got really nasty lag, so I uninstalled it.

*BUT*

if a better aerodynamic model (FAR or similar) was introduced and made stock, I would just deal with it.

its an opportunity not a threat. trouble shooting, redesigning, improving, redesigning, testing, enhancing, etc = fun.

that's why I play KSP in the first place so bring it on.

[LethalDose]

For those that will get angry - give them an option to use the old model, preferably like sumghai suggests.

Oh, man, don't get me wrong, I agree with you and the OP 100%. Both improving the aero model and allowing players to toggle between the original and improved are great ideas.

I just don't think Squad will ever implement it.

[MrAnonymous]

Learning to play this game has a HUGE learning curve. Still people loves the game even though everytime they launch their rockets, it blows up at the launchpad. I think adding proper aerodynamics would be like that. Isnt that what KSP is about? Building things to test it and see if it works. Adding aerodynamics would teach people about how rockets and planes work (for the most part). It is called a space sim for a reason.