Procedural Parts (The best way to improve the game experience?)

[Guest]

I think that the technology should effect two things: Structural strength and price. The former would set an absolute upper limit on the tanks (could be increased by material research), the latter would set a practical limit that could be overcome with enough money. The price should be a function of the tech level and tank size, with various techs affecting the function's curvature. So, initially, the costs for making tanks beyond certain size would skyrocket quickly, you'd have to develop a new tech before they become feasible. You could probably rush the development and focus on tanks only, but you'd end up with N1-like monstrosities running 30+ 1.25m engines and a zany, grossly overengineered scheme to make it all stick together (much like the actual N1 was, which I realized when working on BobCat's one) in order to get anywhere. This could be a very interesting system, mirroring how real rocket development works. Also, something similar could be useful for wings. Nothing stops you from making a 747-sized wing from wood and canvas, but the resulting aircraft would probably end up looking like the Spruce Goose, and fly about as well.

[boomerdog2000]

... if procedural parts are added to the game this mean that you dont have the normal parts any more! ...

Are you joking? I already answer you this 2 times! Can you please read? Or tell me an example of how this will make worthless the squad´s work.. Please!! tell me..

No I'm not joking. Squad has already implemented the techtree based on locking the current parts until you can unlock them. That is the entire techtree. That is what Squad has been working on for this entire update. Therefore removing a majority of the stock parts and replacing them with 3 procedural ones would render most of this current update worthless.

[Awaras]

No I'm not joking. Squad has already implemented the techtree based on locking the current parts until you can unlock them. That is the entire techtree. That is what Squad has been working on for this entire update. Therefore removing a majority of the stock parts and replacing them with 3 procedural ones would render most of this current update worthless.

It might render all the work that went into the part models worthless, but the idea still remains viable. Instead of a new tech unlocking a larger tank that can hold 2000 units of fuel it unlocks something like 'the maximum capacity of procedural tanks has been increased from 1000 to 2000 units of fuel.'

[Deathsoul097]

Really, I think this is what we need to be procedural;

- Structural panels (Macey deans ships, and asthetics really. Could possibly work as heatshield, or maybe just a P-Heatshield, for aerobraking and LSO missions)

-SRBs (like stretchy tanks GUI, but showing mass burn time and thrust, and each of these changed by changing the length or radius of the Booster) (would be very good for making stuff like a Vega rocket)

- Girders and structural pylons (in the modular way they are now, so each mouseover would increase length or diameter in certain increments)

- Solar panels (ISS alike arrays - Pretty much it until actually start needing monumental amounts of power in the science updates)

- PFarings (should be stock anyway, freakin great at what it does)

- Wheels (adjustable suspension, wheel type, wheel size and elevation. Larger wheels means increased power-draw)

I cant really think of anything else but for now that's my 2 pence.

[oggylt]

I think that modular tanks would have problem of estimating how much fuel you will need for your missions, since most of us are not NASA and we do not calculate how much of which fuel we would need, I just slam one orange tank and half, then put some boosters on it and I have my first stage.. I am not very anxious about this suggestion, but if that would get implemented, I would suggest leaving old tank templates for estimation purposes. An why would you need SQUAD to exactly recreate mod which already does what you want. This game is made so everyone can have copy of game that they enjoy playing, with beloved mods and without parts they don't need, so why force others to use copy that only you love.

PS: It would totally be too hard for me to enter SRB's mass, burn time,thrust. I would end up with something unbalanced or would be forced to study how all that thrust balancing works

[Technical Ben]

For me, the primary purpose of procedural parts would be to reduce part counts by having monolithic components, but I definitely agree that in real life component dimensions are standardized for ease of manufacturing. So for the most part, I agree that procedural parts in KSP should "snap" to these aforementioned standard dimensions.

It would also reduce part count by having parts that can include batteries or RTGs internally. It would remove 3-4 parts even on a small craft, possibly more on a larger one. Then you can still get "damage" from collisions, or make those with internal parts more delicate.

[Outlander4]

I think going all procedural is a bad idea; it changes KSP completely, up to the point of taking away what I personally like about it. The design challenge in real space programs is to use as many standard components as possible to reduce cost after all, so a greatly simplified version of it we have in game is just the right thing.

Also, going all procedural may have some really nasty effects on Challenges that many people enjoy doing. What's the point of fiddling with your designs, trying all sorts of crazy contraptions and questionably sane set-ups if you can just hand-craft the rocket to suit your needs? Creativity should have some limits else it degenerates into something trivial or something completely incomprehensible, and limited number of parts is a good limitation to have.

Procedural parts that 'snap' to certain forms and sizes sounds better I suppose; it's all about performance in this case. However, I'd still prefer a current model with 'manufacturers' offering different parts; and contracts and relations with those manufacturers may be a cornerstone of the budgetary system in the career mode if done in a simple and intuitive way. That's just liberal thinking, though.

[Nachtwind]

though i like certain procedural parts, like fairings for example i also dont like including too much of that kind into the vanilla game as it would make createing stuff too easy. There is no challange anymore when you just have one large tank instead of having like 20 smaller ones that you need to balance, fit into a design etc etc...

Yet for fairings and stuff this is imho the way to go since they are rather cosmetic parts.. i'd also like to see procedural launch clamps and ladders.. but beyond that? Na.. no real need for that stuff.

[ZedNova]

Maybe in the future they'll add a couple of limited Procedural parts to the very end of the tech tree. Or they could even just exclude them from Career mode, while still allowing them in sandbox mode.

[AngelLestat]

I did some tables to help clarify this idea, but first I wanna answer some things.

I think what AngelLestat is trying to say but not getting across: the Development tree would create technologies that could feed into the procedural engine, plus the creation of new procedural parts would require research and development. So, to create a new 1.25m tank of triple size would require R&D to be able to make any 1.25 parts, plus extra R&D to create each length of 1.25m tank.

Using procedural technologies might allow Squad to create a vastly more detailed tech tree with very little extra model-making work, so long as the constraints on part sizes and costs were well-balanced. After all, there's no point in allowing us to create 500m-long 1.25m fuel tanks!

Yeah, that can be the idea. But well, some people love so much the game that Squad becomes like a god and the game in a bible. So if someone suggest something different that is not in the bible. without clarify some things It becomes an heresy

For that reason this steps needs come from Squad, but it will take time convert many parts to procedural, so they need help. If they ask for help giving us the parameters how they want to do it, I am totally sure we all be willing to help, is not the same spend time in something without knowing if someone will use it, that in something that will become a part of the game to improve game experience, almost all modders will help.

I think that the technology should effect two things: Structural strength and price. The former would set an absolute upper limit on the tanks (could be increased by material research), the latter would set a practical limit that could be overcome with enough money. The price should be a function of the tech level and tank size, with various techs affecting the function's curvature. So, initially, the costs for making tanks beyond certain size would skyrocket quickly, you'd have to develop a new tech before they become feasible. You could probably rush the development and focus on tanks only, but you'd end up with N1-like monstrosities running 30+ 1.25m engines and a zany, grossly overengineered scheme to make it all stick together (much like the actual N1 was, which I realized when working on BobCat's one) in order to get anywhere. This could be a very interesting system, mirroring how real rocket development works. Also, something similar could be useful for wings. Nothing stops you from making a 747-sized wing from wood and canvas, but the resulting aircraft would probably end up looking like the Spruce Goose, and fly about as well.

That is kinda the idea.

No I'm not joking. Squad has already implemented the techtree based on locking the current parts until you can unlock them. That is the entire techtree. That is what Squad has been working on for this entire update. Therefore removing a majority of the stock parts and replacing them with 3 procedural ones would render most of this current update worthless.

Is like softweir said. But even if is not. The hard work in making a tech tree is not about setting the name of the pieces that you unlock. That is the easy part. The hard part is make the UI of the tech tree, the quest logic that will allow advance in the tech tree. But if you wanna change some unlock for other one, that is super easy.

I think that modular tanks would have problem of estimating how much fuel you will need for your missions

Estimating some fuel? If you know where you go with a orange tank then use the procedural tank that represent the old orange tank size. But if you are worried about mechjeb estimations or kerbal enginner mod; just relax. This is not a issue, I already use many mods with different fuels types and tanks sizes and mechjeb or KEng dont have any problem. This is becouse they use drymass, amount of fuel, thrust and all the same parameters that the game use to make their calculates. And all those parameters also are in the procedural parts. The game will not work if a part do not include those parameters, the difference is these are include in a dinamic way.

I think going all procedural is a bad idea; it changes KSP completely, up to the point of taking away what I personally like about it. The design challenge in real space programs is to use as many standard components as possible to reduce cost after all, so a greatly simplified version of it we have in game is just the right thing..

What? you are saying that a real space agency utilize standard components to make space vehicles? the only standard component that I saw in a rocket or space station is the notebook that some astronauts carry.

-----------------------------------------------------------------------------

Pople like it or not, I guess there is not much choice.

I am 100% sure that every time that squad is thinking in add a new part, they also think in:

-How more mess will become the parts menu with this addiction?

-How much time load and memory usage will take this?

-There is another part that is not very usefull to discard?

-This will be compatible with the new physsics model updates?

And how it does not depend on them to improve the unity engine for Multi-core and multi-threading performance and allow 64b to extra memory, then they need to search a way to get out of this death end.

Now if squad fix the aerodynamic system making good calculations of the drag, oranges tanks will not be enoght to send heavy payloads to space. So the only way to do it is including tanks of 3.75 or 5m in diameter. If they do that, then need to add reactions wheels, decouplers, adapters and extra structures to support that new size. So they need to do all those questions that I listed a lot of times.

With procedural parts you do not only fix all these problems, you also improve a lot the whole game performance. Right now all the rockets that we build looks "weird" becouse we have limits part to build things.

Also is a lot easier to update procedural parts that hundreds of different parts.

Graphs comming...

Edited October 5, 2013 by AngelLestat

[Outlander4]

What? you are saying that a real space agency utilize standard components to make space vehicles? the only standard component that I saw in a rocket or space station is the notebook that some astronauts carry.

Rockets and space ships are made from a huge number of parts; some spaceship systems are fairly conventional (electronics, much of the hydraulics, some parts of the turbomachinery and/or gas generators). Many parts are created and manufactured with the sole purpose of being used for the space flight; but many are either fairly standard details used in the rocketry-unrelated heavy machinery or based upon standard parts. Building everything from scratch is very expensive.

If you want an extreme example of that, check Copenhagen Suborbitals, especially their YouTube videos. NASA is on the other end of the scale; that's why their projects are so expensive.

[AngelLestat]

This is a rought idea of how intruduce procedural parts

I could not understand sometimes the choice of measures made it by squad, however these can be fully respected or in a convenient case they can be modify, would not generate a compatibility issue with old crafts.

The idea is like this:

You choose the procedural parts of cilindrical tanks. You start with the blue cell choice size. You had 2 arrows, one up and at the side to modify large and diameter. You will have only the unlock parts sizes to choose.

Then you choose the type of material, this can be a lighter one or to allow carry cryogenic type fuels (all depending your tech tree unlock).

You still do not have fuel inside. So you can use the tank just like structural or you can put Liquid fuel, or LF+oxidizer, or RCS, or maybe a different future fuel type as Liquid Hidrogen, Oxigen, etc. (this will help a lot when live support comes in).

What does the game? In the game you have a "case code style" that would set the cfg parameters needed depending the size chosen, I would like a texture that depends on the material that you use and a bump texture that will depend on the fuel that it carrys. With that combo you always will know what has inside and its dry mass.

For adapters is the same thing, you can use them like structurals or you can put some fuel into. The same for spherical tanks (that will use the RCS spherical tanks model)

The hald spherical parts had the same use and also works like nose cone to reduce the drag.

There are some differences, in the FL-T200 that i mark with a *, the real size in the game is 1.107954 and Oscar-B 0.348474. Is weird, becouse if you see the table and the volume my measure has sense but not the real. Also is a measure that it does not follow (in my poor understanding) the standard sizes parameters.

The same with other parts like decouplers, in this case you also choose if you separate 1 side or both.

How you can see, after you unlock all, you have a lot of new parts to improve your rockets designs, additionally we are reducing by a lot the count parts and the files, textures, etc that the game needs to load.

More sizes equal less count parts in your craft, this improve the frame rates by a lot.

This mean that you can use better textures for the planets or have extra additions to the game.

Once you have a procedural part like the reaction wheels or a leg, modders can use that part like base to edit and create a new one. For example an air brake for airplanes with 3 different size and drag parameters.

Right now is annoying had a mod part that only works for one rocket size or the mod has differents part to work with more sizes but that is a bigger problem for the extra parts.

The table examples are just few of many of the possibilities, almost any part can have a benefic from procedural parts.

Edited October 6, 2013 by AngelLestat

[boomerdog2000]

What? you are saying that a real space agency utilize standard components to make space vehicles? the only standard component that I saw in a rocket or space station is the notebook that some astronauts carry.

Generally the launch stage is standard. The payload ( satellite, space station module) is made from scratch. Without reliable reusable launch systems production has to be standardized. Otherwise companies couldn't afford to make the launch stage. NASA is trying to reuse old space shuttle engine for its SLS. SpaceX's Falcon Heavy uses 28 of the exact same engine, 27 in the launch stage and 1 in the transfer. Falcon 9 uses 10 of those same engines. So yes it seems fair to say that real space agencies use standard components.

[Lazro]

My main problem with this is it makes it very hard to make launch vehicles and their respective payloads match up. If it's all free and slide-y, then there will always be some difference in size between one tank and another, let alone between engines, capsules, intakes, etc. and the fuel tanks.

[AngelLestat]

Outlander and Boomedog.. The 95% of the components of any thing made for space are not standard. Stop trying to discuss something that is impossible to win.

Maybe some computer circuit of a satellite is standard (the most advanced are not), but well, here you go, you can use mechjeb made by Mario´s Industries.. But what about those high efficient solar cells with antiradiation layer and cooling method that you want it for christmas?

NASA is on the other end of the scale; that's why their projects are so expensive.

So now we need to pick 1 agency that nobody knows between all others just to see if at least the 20% of the components are standard???

SLS. SpaceX's Falcon Heavy uses 28 of the exact same engine, 27 in the launch stage and 1 in the transfer. Falcon 9 uses 10 of those same engines. So yes it seems fair to say that real space agencies use standard components.

Haha wrong.. that is not standard becouse the falcon heavy was develope by the same space-X. Is not like, "hey I develope an engine, I sell it for 100 $ each and 299$ if you buy 3."

And the merlin-d engine top stage is modify, so there is no definition that allow you take that like standard.

I know very well that in industry you need try to use standard components to reduce cost. But space industry is the worst case that you can pick to search standard components. This is due by the high cost related to send something to space and becouse that medium is totally different, so all the things that have a purpose here in earth, they do not work up there. And more taking into account that you have a very very small margin of payload and efficiency to reach orbit.

Space X is very efficienct and it uses 95% fuel to carry 5% of payload. 1% means that you are in or out of the market.

And even if in ksp you want standard components.. You dint see the table? You had standard measures to chose the components. And in the case you unlock the new materials, you can paste in the tank texture "intel inside", i dont know.. but something to show that is the bad ass version of the tank made it by Crazy´s industries.

Now I want to hear how you will deal with all the problems (that procedural fix) if we keep using the same system?

y main problem with this is it makes it very hard to make launch vehicles and their respective payloads match up. If it's all free and slide-y, then there will always be some difference in size between one tank and another, let alone between engines, capsules, intakes, etc. and the fuel tanks.

all free and slide Y?? what?? you read it?

Edited October 6, 2013 by AngelLestat

[Lazro]

As in, if you allow too much freedom, it will become very difficult to standardize and tanks will all be slightly different if you have to adjust them. It's not ideal. Plus, procedural tanks take away some design possibilities: Inside surfaces between tanks can be used to mount things onto, not to mention how different each tank looks from each other now. It would make shipbuilder's lives much harder.

Also, a simple solution to the too many parts problem:

Menu tabs.

[Scoundrel]

So, this is my humble opinion on the subject, and it should be taken as that: an opinion.

Procedural parts are easy to standardize: you simply control the granularity of the inputs. A slider that locks to 1.25m, 2.5m, 3.75m, etc. isn't difficult to program: it eliminates variances in terms of diameter without negating the advantage of proceduralization. You can even tie the maximum diameter to the tech level so your procedural part won't exceed 1.25m until you unlock the 2.5m tech, and so on.

Likewise you can do the same with length, limiting the maximum length of the tank to say 3m until you unlock the tech that lets you have 6m tanks, and so on; and by keeping the granularity of the sliders - say .25m in length every extension of the slider - you can build consistent tanks every time. Trial and error would reveal the best granularity with which to control tank size to ensure consistency.

That said, near as I can tell, the only real argument against procedural parts is that it changes the feel of the game somewhat because you aren't stacking premade parts together in a (deliberately) childish method of constructing a rocket - it's this limitation that we have to work around that adds some depth to rocket building, and there's apprehension that procedural parts risk taking away some of the charm of the game, which is a reasonable concern, IMHO.

And I don't think that procedural parts will fix the wobbly rocket problem or that it will solve physics issues, because reducing part count by making bigger, more massive parts reveals a completely new set of issues with the physics engine that need to be overcome. Nyrath's Orion Rocket thread is a great example of the problems of stacking massive parts on top of each other. KSP just doesn't like it for some reason.

So, in summary, I submit that procedural parts will fix some issues - especially UI clutter - and will allow for a lot more flexibility in ship building by allowing the player the ability to tailor parts (especially ladders - that one definitely needs to be procedural!) to let them build the ship they desire - this, to me, is a great thing IMHO, and I personally am for procedural parts. I would just add the caveat that I don't think it is the panacea for what ails KSP.

[Outlander4]

Outlander and Boomedog.. The 95% of the components of any thing made for space are not standard. Stop trying to discuss something that is impossible to win.

Here we go again. Unlike you, I actually know what I'm talking about; that's the whole reason why I opened my mouth in the first place. Making a rocket involves striking a fine balance between using parts you can just buy from manufacturers and use them with little or no modification and creating parts in-house, which involves a lot of work and costs a lot of money. And you'll have to test it quite extensively as well.

The current low-cost trend is to use as many standard, off-the-shelf components as possible; even JAXA which would rather craft the rocket from the finest carbon-fibre-reinforced sakura wood (that was a joke) starts adopting this approach in full, as evidenced by their Epsilon rocket (http://www.jaxa.jp/projects/rockets/epsilon/index_e.html). SpaceX buys most of the small parts from other manufacturers; they do fuel tanks themselves I believe, and many parts of the engine are also developed and produced in-house. Even Soviet/Russian space program relies on this approach.

The only Space Agency that seems to develop everything in-house is North Korean Space Agency (and look where it got them!); and I sincerely believe that even they prefer to disassemble old Russian and Chinese equipment in search of fitting parts rather than manufacturing everything themselves.

Now, back on topic - I said that snapping to the fixed sizes is ok, and your table even makes sense (by the way, it's written 'spherical', not 'espherical', no matter how you pronounce it, English is not Castellano).

The problem is that we already have parts that we like, and making e.g. fuel tanks procedural will give little benefit when using complicated parallel-staging designs. It would decrease part count here and there but it won't reduce any of my rockets to 5-10 parts as promised. So I see little sense in that. There are plenty of other optimisations that could be done to the game and game engine; and when Unity resolves its 64-bit problems and goes more multi-threaded we'll surely see an increased performance. Also, stock parts can be optimised as well in the manner B9 parts are optimised to be more memory-friendly.

Edited October 6, 2013 by Outlander4

[DaRocketCat]

What about damage simulation?

If you hit an object with a procedural wing the whole wing will disappear, meanwhile with a normal one you'd only lost the tip.

[Lazro]

That's all well and good until your CPU melts.

[AngelLestat]

Harverster, in a tv interview said that he wanted to made a game where the player could unleash its creativity, for that, the game needs to be the least restrictive possible in this regard. (his words!)

Hundreds of parts will make the game less limitating, but for a new player it will be harded to start. Plus you get all the performance issues that I already listed.

So heres comes procedural, that fix many problems in one shot.

To day I saw that some modders update their mods with low quality texture. And there is also a low quality (stock parts) texture pack too.

The game already uses low quality graphics but we need to lower more? Why is that?

Everybody start to notices that you need to do some choices in ksp.

You need to play with all stock and try to not build things with really hard parts count.

If you choose add some new parts, then you need to reduce the size of your crafts.

If you wanna new textures for the planets, then you can not have many parts or big rockets.

We need to sacrifice always something.

Some arguments against procedural are that (less options and oglys rockets means more fun..) Someone can explain me that?

Here one example:

I acknowledge that I am exaggerating a bit with this image. But in some way I guess everybody understand where I am going.

We need extra sizes. Is the only way to implement a good aerodynamic system. And is time to see the utility in the nosecones.

That said, near as I can tell, the only real argument against procedural parts is that it changes the feel of the game somewhat because you aren't stacking premade parts together in a (deliberately) childish method of constructing a rocket - it's this limitation that we have to work around that adds some depth to rocket building, and there's apprehension that procedural parts risk taking away some of the charm of the game, which is a reasonable concern, IMHO.

And I don't think that procedural parts will fix the wobbly rocket problem or that it will solve physics issues, because reducing part count by making bigger, more massive parts reveals a completely new set of issues with the physics engine that need to be overcome. Nyrath's Orion Rocket thread is a great example of the problems of stacking massive parts on top of each other. KSP just doesn't like it for some reason.

Yeah, those are one of the things that I wanna said but my english lv not let me

I know that there are some problems with massive objects, but I use them all the time, 3,75m or 5m parts and I dont see problems, in fact I need less struts to rise heavy payload with huge tanks than with orange tanks. And less engines and parts means higher performance in frames.

Here we go again. Unlike you, I actually know what I'm talking about; that's the whole reason why I opened my mouth in the first place. Making a rocket involves striking a fine balance between using parts you can just buy from manufacturers and use them with little or no modification and creating parts in-house, which involves a lot of work and costs a lot of money. And you'll have to test it quite extensively as well.

So when I said that almost nothing is standard in space industry "I dont know what I am talking about?" ok...

In your source url I can not find anything according to your words.

Even if you find in a weird agency that a few products are standard, you can not denied that the most important parts (the same that we have in KSP) are NOT standard.

Tanks are not standard. Engines are not standard. Fuels are not standard in mostly all cases, fairings, structural, etc.

The fact that rocket technology also means continental missiles, is one of the "why" there are so few private companies working in this.

The only Space Agency that seems to develop everything in-house is North Korean Space Agency (and look where it got them!); and I sincerely believe that even they prefer to disassemble old Russian and Chinese equipment in search of fitting parts rather than manufacturing everything themselves.

Haha, you think that your example is fair?? USA and URSS work also alone in the 60th and they do fine.

I guess we have different concepts about what is a"standard component". For me a standard component is something that is not made for the case. Is something that is at sale to normal public or companies, or is something that at least buy more than 1 space agency.

Now, back on topic - I said that snapping to the fixed sizes is ok, and your table even makes sense (by the way, it's written 'spherical', not 'espherical', no matter how you pronounce it, English is not Castellano).

The problem is that we already have parts that we like, and making e.g. fuel tanks procedural will give little benefit when using complicated parallel-staging designs. It would decrease part count here and there but it won't reduce any of my rockets to 5-10 parts as promised. So I see little sense in that. There are plenty of other optimisations that could be done to the game and game engine; and when Unity resolves its 64-bit problems and goes more multi-threaded we'll surely see an increased performance. Also, stock parts can be optimised as well in the manner B9 parts are optimised to be more memory-friendly.

So your solution is wait until unity fix their multi threaded and memory problems? I want to go on vacation to Israel, but I will wait until the war with Palestine is over

But you still have some problems, if you wanna add 3,75 and 5 sizes. You need to add like 30 extra parts (at least).

-So you still have the loading time problem,

-how you find a part between so many?

-developer time to update each part increase

-you still has a lot less parts than with procedural

Plus if you have procedural and unity fix all that, you have always an extra performance that you can spend it in anything you want, like better graphics.

About the count parts, I dont know are the rockets that you do, but space stations or rockets will use an estimative of 30% less parts in average. And +1000% in some cases.

What about damage simulation?

If you hit an object with a procedural wing the whole wing will disappear, meanwhile with a normal one you'd only lost the tip.

I still i am not sure the best way to implement procedural wings. Procedural wings mods is a good addition, but i guess it will be better limit the size parts that you can get.

Also there are some assets from unity that allow you to cut meches and make 2 or many pieces of 1 single mesh. But that is not the exact case scenerary that you can have in a collision.

I will like to know what squad think about this. I'm pretty sure that they thought this over and over, but the time that they need to spend to make almost each part again with a procedural approach is enoght to discourage.

For that reason I'm looking the ways that Squad some how made the design parameters and ask modders for help.

[boomerdog2000]

Even if you find in a weird agency that a few products are standard, you can not denied that the most important parts (the same that we have in KSP) are NOT standard.

Tanks are not standard. Engines are not standard. Fuels are not standard in mostly all cases, fairings, structural, etc.

The fact that rocket technology also means continental missiles, is one of the "why" there are so few private companies working in this.

Maybe this is where we are encountering a problem. What I, and I think Outlander, are referring to when we say standard is, not custom-made.

For instance the Falcon 9 launch stage. SpaceX can build 30 of them and then store them to use with custom built payloads later. While the Falcon 9 may have been custom-made at first, now that they plan on using it more often they presumably have assembly-line techniques to speed up the process. This kind of assembly-lining makes the launch stage standard. While since the payloads are designed for a variety of different tasks they are not standard.

[Outlander4]

So when I said that almost nothing is standard in space industry "I dont know what I am talking about?" ok...

In your source url I can not find anything according to your words.

Even if you find in a weird agency that a few products are standard, you can not denied that the most important parts (the same that we have in KSP) are NOT standard. Tanks are not standard. Engines are not standard. Fuels are not standard in mostly all cases, fairings, structural, etc.

The fact that rocket technology also means continental missiles, is one of the "why" there are so few private companies working in this.

Please, hear what other people say. And read things carefully, not just skim them. Real-life rockets are not like that in KSP, and instead of several dozens parts they consist of many thousands of parts. Engines alone may go well over thousand. Granted, reaction chamber (where fuel burns), injector(s) (thingie that pushes fuel into the chamber), and nozzle are manufactured especially for use in space. A lot of other things (tubes where fuel and oxidiser run, vents, controls, many parts of turbomachinery/gas generators, hydraulics that drives the gimbal etc.) are either standard or something serious manufacturers produce for spaceflight but based on the existing technology (e.g. parts for steam turbines, gas generators for torpedoes and God knows what). You don't hand-made all the stuff in-house, you buy most of it. The same goes for electronics - it's pretty standard; most of it (especially for interplanetary missions) is radiation-hardened but manufacturers did it first for military and only then - for our anaemic space programs, using the same components and technologies. Fuel tanks and rocket bodies are in-house developments, but made from metal/composite sheets available for almost anyone with enough money. Fuels are pretty non-standard, but then Russians use quite standard mixtures of RP-1 and LOX, which is actually the same for Atlas V since they are using Russian RD-180 engines. Did I get my words across this time?

Haha, you think that your example is fair?? USA and URSS work also alone in the 60th and they do fine.

I guess we have different concepts about what is a"standard component". For me a standard component is something that is not made for the case. Is something that is at sale to normal public or companies, or is something that at least buy more than 1 space agency.

Don't quote URSS (USSR in English) to me, I was born there and know perfectly well how things had worked. Company that did structural parts for Energia rocket also produced metal beds, most if not all engine companies did also work on aviation engines; NPO Lavochkin that did most of the Soviet interplanetary probes produced planes during the WWII and had a lot of experience integrating lots of different components together. Energia corporation itself produces a lot of things for consumers because some technologies and process are equally well applied for making both cutlery and rockets (no joking here). Basically, the company with most expertise in something was given a task to produce something they are good at for space program, and it did. If it did not, bad things happened, and no, I'm not talking about reducing of chocolate rations.

Just go through Wikipedia you seem to love so much, find all companies involved in the rocket building and you'll find out from where all those parts are coming.

So your solution is wait until unity fix their multi threaded and memory problems? I want to go on vacation to Israel, but I will wait until the war with Palestine is over

It's fixed on Linux already. 64-bit works there fine. And I don't know about you, but Windows 8 showed me that Microsoft lost any resemblance of care for their customer, so my next big PC will feature either some Linux or PC-BSD. I prefer the latter (force of habit), but I have no idea how would KSP behave with their Linux Compatibility Layer.

But you still have some problems, if you wanna add 3,75 and 5 sizes. You need to add like 30 extra parts (at least).

-So you still have the loading time problem,

-how you find a part between so many?

-developer time to update each part increase

-you still has a lot less parts than with procedural

- There are more optimisations to be done.

- The answer is 'with ease'. If you want to elaborate, more tabs and search box would be a welcome addition.

- They made part module system for a reason; making a part is quite easy now since they don't code them from scratch.

- We don't need an awful lot of parts, that's exactly the point you're missing.

Plus if you have procedural and unity fix all that, you have always an extra performance that you can spend it in anything you want, like better graphics.

Graphics in KSP matters little, physics occupy most CPU time here.

About the count parts, I dont know are the rockets that you do, but space stations or rockets will use an estimative of 30% less parts in average. And +1000% in some cases.

I tend to use one, maximum two fuel tanks for a stage, and pack stages in parallel. Oversized launch vehicles are for wusses anyway

My landers tend to be small and use many small fuel tanks in parallel and several engines; before my PC video card broke I quite enjoyed using side-mounted Rockomax engines. It all uses a copious amount of decouplers; most stages have sepratrones and probe cores to deorbit them, and interplanetary stages have independent RCS supply and probe core with solar panels to further use them after spending fuel for the main engines. My stations usually resemble Mir space station with an addition of an awful lot of trusses forming a ring-type docking structure (usually reinforced) so incoming ships won't go anywhere near sensitive things like solar panels and habitation modules.

A couple of old images, if you're interested:

http://i.imgur.com/sKiF4.png

http://i.imgur.com/w1h77.png

So going procedural makes little sense for me, I'm perfectly fine with the current parts.

EDIT: Also, I recommended Copenhagen Suborbitals because, while representing an extreme approach, they do everything more or less exactly like the big agencies did in 60s, so you may find out a lot of interesting things about designing and building rockets in real life. In general guys who build rockets using donations and volunteers and publishing everything they did under open-source licenses deserve respect, not bashing. They are living the dream we all cherish.

Edited October 7, 2013 by Outlander4

[Van Disaster]

As in, if you allow too much freedom, it will become very difficult to standardize and tanks will all be slightly different if you have to adjust them. It's not ideal. Plus, procedural tanks take away some design possibilities: Inside surfaces between tanks can be used to mount things onto, not to mention how different each tank looks from each other now. It would make shipbuilder's lives much harder.

So use two procedural tanks: just like existing things like pWings, you don't have to make your entire tankage out of one component, I use 3-4 pWings to make one wing as it is. I've already suggested using multiple selectable textures so tanks don't all look the same ( it would be rather nice if you could switch model like that too, have the structure as some sort of meta-tank ). Any sizing issues can be solved with presets, and people treating it like a puzzle game can just use presets & not notice any differences.

Also, a simple solution to the too many parts problem:

Menu tabs.

Doesn't solve anything. The part count issue is because the game loads & builds all parts at launch, and there is a finite limit ( around 640 parts it seems ). And really, pages and pages and pages of similar parts? how is that at all sane.

[John FX]

I like procedural tanks, wings and fairings. I`d like to be able to add my own graphics to them. I would like them to `snap` to fixed size parts. I would not like procedural engines, pods, science or utility although a few extra sizes of NTG would be nice.