On eXploration, astronomy, random bodies and other overhauls

[monamipierrot]

1 hour ago, tater said:

OP is in the full random, and unrealistic "unknown" camp. That doesn't make randomized bad, it just that particular mechanic not ideal.

I'm not sure what you means here. If you mean I would like to have unrealistic solar system, well, that's not what I want, but I would love that if a player wants, he could tweak this realism and have Kerbin orbit around Mun.

Instead, I would love a random generator that could invent new worlds with some weird features, and still respecting the game physics. Something that a designer could have imagined. Or not. Why we need this? Very simple. When Columbus looked for China beyond the Ocean, he found a entire separate continent. LAter somebody found there were TWO oceans. Separated by a tiny isthmus (and one lower than the other). And that in this continent there were new food. Even good food indeed (discovery of potatoes are said to be the real reason of European domination of most World in the next 4-5 centuries).

Not even Columbus imagined a new continent. Literaly nobody thought there could be 2 oceans. And I'm very very sure about this: not even the wildest insane guy of year 1491 could have ever dreamed, in some horrible fever due to prolonged hunger, about a simple potato.

So IMHO the real goal of a random generator would be to make you say "wow" and feel you discovered something really new, and not "just another frozen moon". Or at least sometimes. And yes: novelty would bring newer technical challenges:

"how the hell can I reach that biome which is 15,000m under average level and come back through its thick atmosphere?"

"is it possible to surf-brake on the surface of this atmosphereless oceanic planet?"

"This small moon's insane rotation speed puts it hard to land halfway between equator and poles. How can I do it?"

"Now, the atmosphere is breathable for jets and kerbals, science is abundant, and the sightseen is unbeatable: perfect for building a base. Absence of a proper planetary surface will give me some headache, thou."

"wow, those 2 moons are very close each other, and their density and size is similar, so I can use the same exact probe cutting down costs. Too bad that one has frozen nitrogen lakes, and the other is mostly covered with magma. Gotta use one of those *cold'n'hot* air conditioning unit".

Tater is right in worrying about the skills of a generator when we are used to - mostly - human designed worlds such as Mun.

But I have seen some great work done with generators in other games. The most known, Minecraft. If you know Freeciv, try fiddling with the "pseudo-fractal" planet generator, too (and this piece has been written by community, not some actual developer). I have come to the conclusion that properly trimmed generators can't be beaten by the finest unaided human handdrawing. Or better said, humans may be superior only by fiddling with generators, tweak them, maybe manual modify some of their output, or just manually SELECT these output to collect the best and most interesting results. This is how I bet KSP developers designed Mun and Kerbin.

But I believe generators with current technology CAN create both Mun, Kerbin, and the wildest potato-moon out there.

47 minutes ago, GregroxMun said:

Uncharted Lands

Wow!!! I didn't know it. Too bad I can see those bodies at the very beginning of the game!

[tater]

1 hour ago, GregroxMun said:

Well if you want some new solar systems, Kopernicus packs have plenty of that. There's Alternis Kerbol Rekerjiggered, Uncharted Lands, and of course Real Solar System.

Yeah, I know. Most requests get answered, "there's a mod for that." Yeah, there is. But stock replay is pretty "meh." If it weren't for mods, I'd have been border long ago (still a good value, though). What mods don't do is the fog of war part.

[Whirligig Girl]

15 hours ago, monamipierrot said:

Wow!!! I didn't know it. Too bad I can see those bodies at the very beginning of the game!

Just don't look at them, then? Maybe pester @KillAshley for ResearchBodies compatibility with UL?

[tater]

I'm not necessarily down on well-done randomized worlds, but in similar threads in the past I have see people argue (effectively) that worlds could use more hands-on TLC by developers. If that is indeed the case, then having a vetted library of worlds still gets us most of the way there. I'd add that I would prefer plausible worlds, so any random generation needs to have some geological rules built in.

[Andem]

55 minutes ago, tater said:

I'm not necessarily down on well-done randomized worlds, but in similar threads in the past I have see people argue (effectively) that worlds could use more hands-on TLC by developers. If that is indeed the case, then having a vetted library of worlds still gets us most of the way there. I'd add that I would prefer plausible worlds, so any random generation needs to have some geological rules built in.

And a limit on how obscenely bright a body's colors are would probably be beneficial as well...

[tater]

Yeah, I'm not a fan of the stock worlds at all in terms of colors. It turns out that the real universe is more than interesting enough, without worlds having to be bright purple, etc.

[Andem]

12 minutes ago, tater said:

Yeah, I'm not a fan of the stock worlds at all in terms of colors. It turns out that the real universe is more than interesting enough, without worlds having to be bright purple, etc.

The stock colors could be toned down a little, but they are pretty. The issue I have is when someone adds a neon pink and green planet to an otherwise perfect planet pack.

[monamipierrot]

3 hours ago, Andem said:

The issue I have is when someone adds a neon pink and green planet to an otherwise perfect planet pack.

Yep. At least they don't add little aliens with big heads and silly green skin. Ooops.

[KillAshley]

On 20/11/2015 at 9:59 PM, monamipierrot said:

Procedural random generated bodies Overhaul

 

The stock KSP bodies are great, really. However, if we implement the Astronomy Overhaul, and we keep the KSP stock bodies, we lose 50% of its purpose, and 99% of it beginning from the second career we play.

That is the reason we should give the option (well, it should be the actual default option) to start a Non-Kerbol universe. At the start of a new (career) game, the program generates a new procedural random Solar System, either based on fixed default balanced parameters (to make sure we keep the pace and difficoult level of the game as close as possible to stock KSP) or on manually tweaked parameters (to have very different solar systems, maybe harder or easier to explore than stock one). Kerbin also should be somewhat random, both on physical properties (but only very slightly differing from stock ones) and in its map and biome configuration.

Some have pointed out that random generation could last hours. Well, if this is a real issue for some CPU, then a workaround could be building a whole online library of already built "balanced default" - and maybe manually selected - downloadable solar systems. I bet it would be easy ot host literally dozens of them to make sure each player will not to come back to an old one.

This is part that I have issue with. You seem to be unaware of the complicated nature of procedural generation of bodies with the current PQS system. Sure it can technically be done, and the stock game itself has bodies that are generated procedurally, however theres a BIG difference between that and the randomization that you seem to want. Suffice to say that the procedural system that you want implemented will leave you with either extremely buggy/unplayable worlds, or small generic bodies that will leave you bored withing 5 minutes. I'm afraid that the only way that you will find new adventures is with mods.

[AdmiralTigerclaw]

Were this me, before I started attempting to create or implement any form of procedurally generated bodies or systems, I would first create an implement a system to help the player automate his space infrastructure.

Before you launch your rockets, you need a base.  Stock game, you have KSC as the base.  Said base gives you access to management tools relevant to the KSC surface base.  Even with mods that help you build stuff in orbit or on the Mun, your interface is BASED on the KSC surface base.  All your flights and operations are pinned in relation to KSC.  A kerbal is 'on a mission' if he's anywhere other than lounging in KSC.

 

Before I would add any form of expanded exploration and procedurally generated systems or bodies, I would add in methods for players to construct 'static' bases that act as resting points for operations.  No more ad-hoc methods or unstable 'bases' that explode at random when the physics engine derps.  Effectively, (and simplified) you land a bulldozer, it builds a 'static' base.  Said static base becomes a new staging point and you interact with it like you would KSC.  The mechanics of these offworld bases would be slightly different of course.  You'd manage it like a Real Time Strategy base.  Before it can build rockets/recruit kerbals, you have to set up prerequisite industry and support 'sites' in the same SOI or Body.

 

Only after I have this expanded infrastructure automation would I even CONSIDER adding more exploration elements.  We are but a single player.  The larger our game gets, the more we must manage at the same time.  That must be streamlined to keep the workload level as our gameplay SoI expands.

[monamipierrot]

11 hours ago, KillAshley said:

You seem to be unaware of the complicated nature of procedural generation of bodies with the current PQS system.

Yep. I am. If you don't prove me wrong I'll continue to think that the task to tell the same generator they already used for planets and moons, in order to have some random results within a tolerance, may be a joke comparing to - say - tweak the physical models of craft to introduce distortion and stress of parts.

If you do prove me wrong, and reverse the above guess of mine, still we can just grab ourselves to Tater's idea of manually generated bodies, which may be just slightly random changed to ensure you'll never find a Solar system similar to the previous one.

So if the problem is just TECHNICAL, as you put it, then I just smile: Squad didn't implement dozens of vital aspects because it was TECHNICALLY hard for them. Videogame industry has proved that if you put enough money, many tecnical problems fade away. Just wait till Squad will be big enough to handle this.

[monamipierrot]

11 hours ago, AdmiralTigerclaw said:

Were this me, before I started attempting to create or implement any form of procedurally generated bodies or systems, I would first create an implement a system to help the player automate his space infrastructure.

Fine for me. And then, let's implement procedurally generated bodies, astronomy overhaul and the rest ok?

I thought it was clear enough that ideas of mine are definitely not the kind of elements you put on a "what's new" list for next 1.1.x version.

This is definitely a general overhaul, and a vital one.

While your idea is great, but I still see it as a new (major) feature in a upcoming 1.x version.

If you read my OP, you may have noticed I also said that introducing the eXploit side in KSP (career, mining...) has been a good move, althou I think it is largely insufficient. Your idea links to the classic scheme for XXXX games, or for RTS games (which is often very similar). Thumb up. We definitely need this. Please just don't have a "war of the poor" between supporters of different area of improvement for KSP. If we agree that KSP is a uncompleted game and that it still has to find its own personality (and again, I repeat that this is great news for everybody!), we can convince Squad to put more effort to move on and don't stop at this crucial point of the development.

The only kind of "suggestion" of this sub-forum which I don't like and I don't support are those about cosmetical overhaul, plain parts addition or even plain bodies addition itself.... I mean, I see all of these tasks legitimate, but they are that kind of work you do AT THE END of a proper development  process. IMHO, those suggestions right now are not only a waste of time, but even dangerous because they may contribute to reinforce the idea that this game has come to a final stage of process, an idea which I strongly deny. I love this game in a similar way as a father loves his child, let's say. It is natural that I want it to grow, learn more and mature. I'm giving strong advices because I know real improvements are not easy, but not because of that we should abandon them. "We chose to do that because it is hard".

Again, hail to Squad for this great piece of software. Please make it more than just a piece.

[tater]

I think the key is not the bodies. It's the career/science/tech overhaul required. 

It requires quite a lot about their career and science systems to change.

[monamipierrot]

3 hours ago, tater said:

I think the key is not the bodies. It's the career/science/tech overhaul required. 

It requires quite a lot about their career and science systems to change.

It has been said dozens of times. The c/s/t system is barely a scratch. Every time a user suggests something, one can have a glimpse of an entire new universe of fun. Still, the current scratch is enough for hundreds of hours of delights, with just a few repetition and grind.

Again, to this eXploit aspect of the game you must also remember adding the colonization and the mining things, and yes as I said before we strongly need the STMCC (Science tech mining colonization career) overhaul.

 

[AdmiralTigerclaw]

9 hours ago, monamipierrot said:

Fine for me. And then, let's implement procedurally generated bodies, astronomy overhaul and the rest ok?

I thought it was clear enough that ideas of mine are definitely not the kind of elements you put on a "what's new" list for next 1.1.x version.

This is definitely a general overhaul, and a vital one.

While your idea is great, but I still see it as a new (major) feature in a upcoming 1.x version.

If you read my OP, you may have noticed I also said that introducing the eXploit side in KSP (career, mining...) has been a good move, althou I think it is largely insufficient. Your idea links to the classic scheme for XXXX games, or for RTS games (which is often very similar). Thumb up. We definitely need this. Please just don't have a "war of the poor" between supporters of different area of improvement for KSP. If we agree that KSP is a uncompleted game and that it still has to find its own personality (and again, I repeat that this is great news for everybody!), we can convince Squad to put more effort to move on and don't stop at this crucial point of the development.

The only kind of "suggestion" of this sub-forum which I don't like and I don't support are those about cosmetical overhaul, plain parts addition or even plain bodies addition itself.... I mean, I see all of these tasks legitimate, but they are that kind of work you do AT THE END of a proper development  process. IMHO, those suggestions right now are not only a waste of time, but even dangerous because they may contribute to reinforce the idea that this game has come to a final stage of process, an idea which I strongly deny. I love this game in a similar way as a father loves his child, let's say. It is natural that I want it to grow, learn more and mature. I'm giving strong advices because I know real improvements are not easy, but not because of that we should abandon them. "We chose to do that because it is hard".

Again, hail to Squad for this great piece of software. Please make it more than just a piece.

 

I'm thinking more in terms of prerequisite functions and just expressing my thoughts.  I've mentioned this before elsewhere, except I don't remember where.  The topic's buried in my post history somewhere.

If procedural system creation came before player asset automation, then you have a situation in which the player is eventually stuck doing every little maintenance task for their growing space agency until you reach a point where it becomes impossible to expand because all of your work is geared to just keeping your existing operations afloat.  Any and all procedural generation beyond this point is pointless because you cannot access it (short of cheating).  There is only so many times you can fly the same resupply mission before it becomes a chore.

 

Really, I think that certain features and issues do need to be addressed in a logical order.  But one also has to remember that these changes do not have to occur sequentially in series.  You express that you don't want visual improvements to be focused on when you feel the program needs the depth of play to improve.  You believe that focusing on visuals means the main content work is 'over'.  Thing is...  Visual Content and cosmetic work is a different field of expertise from the gameplay code overhauls and bug fixes for new features.  They can literally be in production at the same time with other people on the team.  And really, you WANT them in production at the same time. 

[KillAshley]

10 hours ago, monamipierrot said:

Yep. I am. If you don't prove me wrong I'll continue to think that the task to tell the same generator they already used for planets and moons, in order to have some random results within a tolerance, may be a joke comparing to - say - tweak the physical models of craft to introduce distortion and stress of parts.

If you do prove me wrong, and reverse the above guess of mine, still we can just grab ourselves to Tater's idea of manually generated bodies, which may be just slightly random changed to ensure you'll never find a Solar system similar to the previous one.

So if the problem is just TECHNICAL, as you put it, then I just smile: Squad didn't implement dozens of vital aspects because it was TECHNICALLY hard for them. Videogame industry has proved that if you put enough money, many tecnical problems fade away. Just wait till Squad will be big enough to handle this.

yes you can create a way to subtly change the bodies within a predefined range, however the main issue is the overall stability of the planets after the changes. The problem is one that is already there on stock worlds, and that is buggy terrain issues such as; negative terrain issues (most bodies), underground easter eggs(tylo), errors in the mesh resulting in invisible terrain(pol), incorrect terrain coloration(Laythe). Don't get me wrong, the general idea of what you are saying is something that I've considered myself to implement into a pack, but in their current state i doubt squad know enough about their own terrain system to implement such a drastic overhaul. Since Novasilisko left all updates to planets have been full of errors, and frankly that spells doubt in everyone who knows about PQS manipulation

I might revive my idea about doing this myself though, so that you can at least have a pack that does it. I mean if you look at my UnchartedLands pack, which is generated completely procedurally, you can see there is promise..however it takes a lot of work

[Pthigrivi]

This is all kind of moot though right? This has come up many times in the past and Squad has pretty conclusively ruled out procedural planets. 

 

What at does seem randomised are the resource maps, which makes me think if the planets had more interesting surfaces and if different biomes on the same world had slightly different science and experience multipliers these fine details could be randomised. Duna would still be there, but whether it was more profitable to land on the poles or on a mountain peak could change from save to save. 

Edited April 18, 2016 by Pthigrivi

[tater]

4 hours ago, Pthigrivi said:

This is all kind of moot though right? This has come up many times in the past and Squad has pretty conclusively ruled out procedural planets. 

What at does seem randomised are the resource maps, which makes me think if the planets had more interesting surfaces and if different biomes on the same world had slightly different science and experience multipliers these fine details could be randomised. Duna would still be there, but whether it was more profitable to land on the poles or on a mountain peak could change from save to save. 

Yeah, this is another reason I suggested a library of planets that could be picked semi-randomly. I say semi, because I think that the right kinds of worlds need to be in the right kinds of locations (no ice world in close orbit of the sun, for example).

[Pthigrivi]

Hmm. Maybe. But its still clashing with the reason Squad ruled out procedural planets to begin with, which is that a single system allows players to have a common experience and lets them go online communicate with each other about what they've found and how to tackle a common set of problems. Another thing to consider is how many players even make it to all the planets we already have? You and I have, plus a handful of regulars here who have played kerbal for ages. But most players barely make it past the Mun. For them what does it matter if there's 20 or 100 different planets that may or may not pop up? I mean things get a bit silly when we speculate on dev time, but just in principle lets say they're investing x amount of time into working on contracts, y amount of time modeling new parts, z on UI, they've got to be thinking about what the return on investment is going to be modeling new planets. Even if you have a procedural system you're talking about mapping hundreds of biomes, defining atmospheres that make sense, applying balanced multipliers, huge numbers of new textures and vetting potentially dozens of new planets. Its a lot of work. I feel like just in terms of delivering the most across the board boost to player enjoyment there's way more milage to be gained helping and encouraging players get farther and find more to do when they get there. I'd rather see players actually rewarded for things like brining and using rovers and planes and devising new strategies than fussing with lander permutations. 

Edited April 19, 2016 by Pthigrivi

[tater]

This would be explicitly an option for replay. The default game would be unchanged. So anyone that wants the "common" experience can have it. Note that if you apply a seed to the random choices (regardless of how the planets are actually created), then people can share those as well. People play 10X Kerbol, 6.4X, 3.2, 2X, RSS, etc, etc. I'm not sure it's an issue. I didn't play very long before I abandoned regular scale completely (though I make a point of playing through a new career with each major update to test it).

[Kosmonaut]

For the "Kid's mode" idea, I think that there should be a dulled down mode but not way less complex like you are suggesting. One of the fun things about KSP is it's complexity. I am 14 now, and have been playing KSP for at least two years and the demo for longer, and I never really had problems and still don't.

As for the procedurally generated solar system, I think that there are multiple ways to do that. You could A) have multiple different solar systems with different difficulty levels on each. The player could choose each one. B) Keep the Kerbol System and have a procedurally generated "galaxy" of ten or so other solar systems. C) Do what you are suggesting, with choosing procedurally generated or stock with DLC and mod solar systems as an option, or my favourite, D) in which there are multiple planets, but not procedurally generated, as well as several example systems (with the different difficulty levels as suggested in A. An RSS-like solar system might be a good idea too for the ultra-realism players. You would then have the choice of a pre-made system (including kerbol and the possible RSS system), a randomized system, or you could create a custom sloar system to your needs and wants.

Thanks for reading, I hope someone might find something useful in here.

[tater]

The only issue with RSS (or rescales much above maybe 4X) is that the parts need to scale, or it becomes frustrating (hence RO with RSS).

3.2X is so playable with stock parts I'd actually prefer that to be the "easy" stock game. It just feels better, I don't know how else to put it, and you barely notice the rescale once you do it. I've done 6.4X with stock parts as well, and that actually works, but it's more challenging perhaps for new players.

[Kosmonaut]

What I mean isn't full scale, but our solar system that we currently inhabit

[Atlas2342]

1 hour ago, KosmonautwithacapitalK said:

What I mean isn't full scale, but our solar system that we currently inhabit

Isnt that full scale? Otherwise i need more sleep

[Kosmonaut]

1 minute ago, Atlas2342 said:

Isnt that full scale? Otherwise i need more sleep

No you are right, I just can't express myself properly... What I mean is a real world solar system at kerbal scale... Everything is smaller 

8 hours ago, tater said:

(no ice world in close orbit of the sun, for example).

Wouldn't you in theory be able to have a reflective atmosphere, thus creating a sort of reverse greenhouse effect?