Procedural sub-biomes to motivate more surface exploration

[cantab]

I must have over 2000 hours in KSP by now. Many other players have hundreds or even thousands of hours. I'll admit mods are a factor, but seriously, replay value is not a problem for KSP.

[tater]

I must have over 2000 hours in KSP by now. Many other players have hundreds or even thousands of hours. I'll admit mods are a factor, but seriously, replay value is not a problem for KSP.

I have a large number of hours (not Steam, so can't guess), and I start a new career with each patch, and with substantial mod changes, plus RSS configs, etc. And I'm doing OK in terms of replay. I tend to play with life support (I prefer low-resource count versions, like Snacks!/USILS, with kerbal death turned on), and even in stock I only send manned craft long distances that look like a Mars reference mission in terms of habitat space just for roleplaying. As a result, I have visited all the SoIs with probes, but have only sent manned flights to a subset, and have not landed on all bodies yet as any failure that I cannot possibly rescue before LS runs out is a major issue (I rarely have fatalities, and most have been kraken attacks with no quicksaves).

Without mods I'd have played a lot less, I think.

I also don't look at the wiki pages, or read up before hand on what to do. I made a probe "bus" that spread a few craft in a line to try different periapsis points to test aerocapture/reentry on Eve, for example, so I could get a feel experimentally instead of just reading a forum post. This is my way of keeping "fog of war." I never used tab to even look at planets I had not been to, and never even played with sandbox aside from orbital testing for similar reasons. I want to be surprised as much as possible.

In short, I don't disagree with you, I did't think it was something worth squad actually doing (total, randomized kerbol system), I said it was "the best possible game from a replay standpoint'" not that it was what they should bother with seriously. It's more like "a perfect KSP for decades of replay of my dreams" not a serious suggestion.

They need to fix what they have

edit: to be clear, I'm against procedural sub-biomes, but I think there is some merit in random, procedural, scientific, "areas of interest" to investigate within rover range of a given landing site on a small scale-size that would only be discernible on th ground.

Edited August 3, 2015 by tater

[Kegereneku]

Change scale, though, and ask a simple question: Could an astronaut or probe ever find something of interest at a landing site that is not apparent from orbital photography? Yes, or no?

You misunderstood what I meant. When I said "visible from orbit", I meant that the User-Interface would display the good places. displaying them after a Scan or Report is mostly a game mechanic doubling as an anti-frustration features.

That aside, we sent Curiosity at place which Interested scientist MORE in the first place, we don't have the budget to land rovers on every single square-meter of Mars.

Again : "Random generation" is a tools, not a goal.

It all tie up to the same reason we have a problem with the more or less random generation of contract, always asking for a more careful procedural generation.

The purpose of KSP is really to explore a solar system with a space program.

What you are describing is more a painful trial&error than a fun exploration. Failure is an tools as well, not a goal.

That aside, I have nothing more to add, I invite you to reread what I said about randomly generating 10 borings planets, and put it against understanding what make a planet/science/design interesting and making sure those aspects are (1) present, (2) balanced with the rest, in the best solar system of the Kerbalverse.

Randomization is involved, and it is unambiguously better for replay, with none of the negatives you mention (I granted all your positive improvements, so you'll need to explain how not having a wiki page for every world makes exploring less cool---the game will generate all that data as a function of what science you do (atm height, composition, etc, etc).

Your statement is unambiguously wrong for many reasons, some stated above plus simply the wide diversity of players expectation.

And this is not a question of semantic. Randomizing aspects do not necessarily increase diversity and diversity don't necessarily improve gameplay/fun, just like play time(replay) do not define how 'objectively great' a game is.

(philosophers here can ponder the significance of glorifying the ability of a game to keep us doing the same things over and over again as longly as possible)

...at which point we can debate FOREVER over the longevity of any video game for anyone.

My point is that a relatively more caring, focused and discrete generation of place of greater interest (made visible from orbit) would be more fun than a brainless spam of inferior-biomes

[tater]

Sigh.

You misunderstood what I meant. When I said "visible from orbit", I meant that the User-Interface would display the good places. displaying them after a Scan or Report is mostly a game mechanic doubling as an anti-frustration features.

That aside, we sent Curiosity at place which Interested scientist MORE in the first place, we don't have the budget to land rovers on every single square-meter of Mars.

What's your point? Of course you would do that. In KSP this would be a "contract" telling you where to land, frankly. You design/fly rockets, a planetary geologist would tell you where to land, you'd not pick place randomly when you are in orbit based upon what goofy orbit you managed to come up with. Everything would be planned. The site would be picked because you flew photographic orbiters first, and mapped, then you'd send something like Ranger to possible landing sites, then Surveyor to actually land there.

None the less, the choices made are because they think there will be a better chance of finding particularly interesting stuff, Curiosity was sent to a location with a high-probability of interesting stuff, but how many particularly interesting things it finds are effectively random. I can go for a walk not far from my house in a site where there is a higher than average chance of finding dinosaur fossils. What I find on a given walk is random. Erosion determines what is on the surface any given day. On top of that, I could walk right over something, missing it, that a paleontologist would have spotted 10s of meters away. In KSP terms, the "biome" is a place where they'd expect to find certain stuff. So lowlands might have large flows like the Mare, and other, related geology. You'd not find mountainous geology there. So within that, you can either add hand-drawn areas of interest to maybe 2 places on the map, that even at tiny KSP scale are like needles in a haystack, or you make some that match the "biome" but can be placed randomly. Without intentionally looking for one, how many "anomalies" have you accidentally found in KSP? That's about what you can expect from hand-made "areas of interest," and the Mun has way more than most worlds, right? I made a point of seeking out the monolith, and even knowing exactly where it was, I had difficulty finding it (it was buried). I've seen no others, If I happen upon them, great.

Again : "Random generation" is a tools, not a goal.

It all tie up to the same reason we have a problem with the more or less random generation of contract, always asking for a more careful procedural generation.

I don't think we disagree here even a little. Of course random is a tool. Random need not mean "utterly random," though. I can imagine a game that randomly generates worlds using a reasonable (simple) model of planetary geology---but I have no where said I want this for KSP, so it's off topic.

What you are describing is more a painful trial&error than a fun exploration. Failure is an tools as well, not a goal.

I disagree entirely here. It's not painful at all. You add the tools to provide the data one would find on a wiki page. If certain data requires mapping, you send a flyby, or an orbiter. The actually useful science drives your choices instead of "Acme corp wants a satellite in polar orbit over Duna" or temp scans, etc. Meh. So we'd have camera parts, or other stuff to actually do that would give is the same data. KSP is all painful trial and error, the dev won't even give people dv data without mods.

That aside, I have nothing more to add, I invite you to reread what I said about randomly generating 10 borings planets, and put it against understanding what make a planet/science/design interesting and making sure those aspects are (1) present, (2) balanced with the rest, in the best solar system of the Kerbalverse.

I never said planets themselves would be randomly generated. Reread what I suggested. I said Squad would make (the same way they do now) more planets and moons, and that these handmade planets would be used as the "pool" to make a random kerbol system drawing a subset of the planets they had made for any given kerbol system. The number of total planets might be 8 +-3, and the number of moons variable per planet (weighted based upon planet size, larger planets tend to capture more moons). Their orbital positions, and perhaps scaling would vary. So in some games the next planet out is Duna, in others it would be an entirely different world. Heck, it might be possible that Kerbin is a large satellite of Jool. Anyway, individual planets are just as good as they are now.

Your statement is unambiguously wrong for many reasons, some stated above plus simply the wide diversity of players expectation.

And this is not a question of semantic. Randomizing aspects do not necessarily increase diversity and diversity don't necessarily improve gameplay/fun, just like play time(replay) do not define how 'objectively great' a game is.

It is wrong for no reasons at all.

State one, please. For long-term replay. Again, if you can reread, the worlds are created in exactly the same way as they are now, or indeed, even as you might wish them improved. Squad creates some extras at the this same level of quality, and gives the return player the option of a randomized system---it builds a new kerbol system using the parts it has (pre-designed worlds). The only change might be scaling, done in such a way to not break things (a 2-3X RSS config is entirely playable with nothing but stock parts, so I'd limit resizing of planets you'd land on to that sort of scale, and might allow the gas giants to get a little larger since they are just gravity-well eye candy. My suggestion was that the thrill of exploring would be greater, and there is more replay fun if the solar system can change if you wish. Say 24 planets made by squad, only 8 of which are ever in any given game, and Kerbol and Kerbin remain the same. Perhaps Kerbin always has at least 1 moon for gameplay, unless it is itself a moon, then there is always an additional moon. What's the downside?

a brainless spam of inferior-biomes

I nowhere suggested inferior biomes, indeed I explicitly stated I am against sub-biomes, and only think that scatter level detail/sites should be random---we have scatter, why not make a subset of it actually interesting?

- - - Updated - - -

Here's what I mean by random, I'm going to clarify, again.

Imagine a card game.

Kegereneku seems to be saying that I want people assigned 8 randomly generated cards. 8 rectangles of paper with random stuff printed on them.

What I am actually suggesting is that the game draw 8 random cards---from a standard deck of cards.

The first has infinite possible cards, most of which will be gibberish nonsense (his point). My idea had 8 choices out of 52, carefully crafted cards that all make sense (in a card game, anyway). In any given game, your hand would be the planets you get from the deck of planets (placed sensibly in zones around the sun, BTW). A different experience each game, but not random in the bad way Kegereneku suggests.

The more planets in the "deck," the more chances to play without a 100% repeat experience. If some small RSS-like scaling is thrown in, then it results in even greater replay value.

Edited August 4, 2015 by tater

[WinkAllKerb'']

I must have over 2000 hours in KSP by now. Many other players have hundreds or even thousands of hours. I'll admit mods are a factor, but seriously, replay value is not a problem for KSP.

yup but when you know where things are, medium/long term, you loss the search and seeks excitement , some procedural science content could be interesting regarding that.

(kinda like the contracts list, we can have a special location list and make them pop randomly in the system @ some coordinate, celestial bodies, original ground structure, wreckages, fonctionnal stations, bases, ship, rover,sat etc. )

-----

edit:

-----

(& also why not some part of this extra content within the spawn pick up list could either be premade and/or procedurally generated itself from some assembling algorythmics basis)

Edited August 4, 2015 by WinkAllKerb''

[Mr. Scruffy]

[...] Randomizing aspects do not necessarily increase diversity [...]

*cough* evolution? *cough*

[Kegereneku]

Sigh.

What's your point? Of course you would do that...

I really don't think I'm overreading your answers.

First point was that you are trying to justify random generation of sub-biome with an improper analogy to real-life, IRL some place are simply not worth exploring because we can get the same meaningful data at better place.

Second point is that we are talking of game-design here, meaning that you shape game-mechanism to represent the fun part of reality, not the boring part.

My stance in short : mindless quantity do not replace skillful quality

1000 sub-biomes which all share 99% of the same simplified characteristic and underuse game-mechanic (like EVA).

...are less fun (euphemism)...

Than 10 "place of interest" made by an human mind with roughly the same amount of work. (more on this later)

(many players don't like being forced to play 100 hours to achieve a simple result)

Next on another misunderstanding : I meant that we don't need to stick on the "Contract given marker" if we want an User-Interface to display place of interest. We can simply display them on the Map-view without contract (after building upgrade or orbiter-scan or else).

The ability to delimit biome from space or at least see them in Real-Time without needing experiment is an often requested feature.

Planning a mission (and having fun) mean having the ability to predict where and what amount/quality of data you will have BEFORE committing a lot/all effort in it.

Lastly , Trial&Error is hardly the best game-mechanic we ever made (euphemism again).

I don't think we disagree here even a little. Of course random is a tool. Random need not mean "utterly random,"

Clarification time : I'm not on a position purely antagonist to yours.

So far I'm saying the randomness you described so far, do not meet (my) criteria to take sub-biome from 'mindless copy' into 'place actually interesting'.

Your suggested use of "Random procedural generation of content" to "increase the game length" isn't new... but it is more often than not what I would call "polishing turd", hiding an inability/unwillingness to generate REAL content with copypasted crap. (Opposite to more skillful use of procedural generation)

I disagree entirely here. It's not painful at all. You add the tools to provide the data one would find on a wiki page. If certain data requires mapping, you send a flyby, or an orbiter.

It depend entirely of how much data and tools you intend to require a player to micro-manage. Being forced to send cheap probes at every single place to know if it worth considering a rover or a manned mission is precisely what is painfully frustrating in Trial&Error gameplay.

Beware of fake difficulty.

I never said planets themselves would be randomly generated.

Not what I understood.

But then I fail to see a point to customize your system at all, nor in doing that "each game" as you suggest. There's strictly nothing good in separating all players experience (plus DeltaV map). So at best it would be an optional Joke-feature.

I see no interest either in having around thirty planets/moons alternating.

Rather than making 8 awesome planets with all interesting feature, you are suggesting to make each game have only 50% of all features in a non-optimized way... and for what ? Fake longevity ? Because that's what that would achieve, forcing players to rebuild all infrastructures with no added-value.

I nowhere suggested inferior biomes, indeed I explicitly stated I am against sub-biomes, and only think that scatter level detail/sites should be random---we have scatter, why not make a subset of it actually interesting?

The inferiority ensue from the balance requirement to overcome the handicap caused by the randomness.

Randomness mean that you cannot customize/finely tune "scatter level detail" (Area of Interest) to fit optimally :

- Interesting reliefs (crater/hills/crest/access to other site and ISRU).

- distance with others AoI.

- distance with ISRU place.

- planetary balance in term of science-point quantity.

- predictability and fairness of career playthrough. (if enough "scatter level site" are too easily accessible, why explore the rest ?)

Worse, it lead to a FAILSAFE generation of bland but FAILSAFE area that have FAILSAFE parameters so it stay FAILSAFE anywhere.

No bets you know the importance of Game-design, but the impression that random generation is a good idea every time is hard to shake off. That's why I insist so much on this.

Again, Random generation isn't a goal, if it bring nothing good to the game -not even replay-, don't do it.

And let's be clear : We are NOT talking of crafting Area-of-Interest with an unrealistic attention to details here. This is simply improving what SQUAD did often like the Mun's Crater-ravine-Crater area, except visible from space, with more sub-biome and with a "slightly painted" road to make it more enjoyable to plan and send a rover.

*cough* evolution? *cough*

Natural selection don't determine who is right, only who is left.

If you wanted all the coolest animals and reliefs of Earth at one place, it can't happen randomly.

[Mr. Scruffy]

Skillfully handcrafted designed interesting spots, will not only be old once you explored them, and boring once you know where they are, they will also have the problem of contrasting to the areas around them, that are not so skillfully hand-crafted, especially considering, that alien-civilizations are a no-go (unless that changed).

I wonder if any of you two even read my rather longish post a page ago. I have the impression that both of you want to creat little bubbles of interest, Tater wherever the player needs them, you, as them being predifinded and designed fixed locations. And both of you are right in the points you bring up against each other. In essence, both would feel like ´Mass Effect´ - as they are both ´bubbles´; one convenient, the other distinctive.

The bubbles we have to consider here are the existing geomes though. Within each of those we want special places. We need not generate them. They already exist. Each geome already features a bunch of different places. They are defined by the terrain properties, altitude and slopyness. Also just taking samples/doing experiments at different (otherwise not varying) lacations within the same geome should have scientific value, upto to a threshold, just by confirming existing data, so to say.

To design interesting places would be like ´creationism´. But to generate them for each of the player´s landing site would be ´anthropocentric´. Those interesting sites are not all that unique in real life, either. The conditions have to be ´right´ (e.g. water can be found in deserts by digging at the foot of the dunes, rather than their tops). Crop said conditions from what the game provides, and you have your interesting spots, based on (a simplyfied) logic and ´natural laws´, as should they be.

[tater]

I really don't think I'm overreading your answers.

First point was that you are trying to justify random generation of sub-biome with an improper analogy to real-life, IRL some place are simply not worth exploring because we can get the same meaningful data at better place.

Second point is that we are talking of game-design here, meaning that you shape game-mechanism to represent the fun part of reality, not the boring part.

Having to discover the solar system is fun. That issue is entirely separate from the OP here, and regardless, simply will not happen.

My stance in short : mindless quantity do not replace skillful quality

1000 sub-biomes which all share 99% of the same simplified characteristic and underuse game-mechanic (like EVA).

...are less fun (euphemism)...

Than 10 "place of interest" made by an human mind with roughly the same amount of work. (more on this later)

(many players don't like being forced to play 100 hours to achieve a simple result)

I explicitly stated I am against the sub-biomes, who are you yelling to? I said there might be some merit in adding something, anything interesting to do on EVA on a scale-size that would not be visible from orbit. By definition, those are areas of interest NOT driven by scientists back home on kerbin, because they cannot know they are there, as they are unresolved in orbital imagery. At best there are areas that might have more of such interesting stuff---in KSP this would require landing where a contract tells you to, or alternately, we can simply assume anywhere a player lands is because his scientist think that area is of interest. Doesn't matter, I'm looking for something interesting to do on EVA. I use KIS/KAS, instead, and tweak my facilities with engineers.

Next on another misunderstanding : I meant that we don't need to stick on the "Contract given marker" if we want an User-Interface to display place of interest. We can simply display them on the Map-view without contract (after building upgrade or orbiter-scan or else).

The ability to delimit biome from space or at least see them in Real-Time without needing experiment is an often requested feature.

I would personally prefer to have to map a planet.

Add camera parts to all probe cores---the most basic instrument on probes to other worlds is a camera, after all. Do something like scansat with some modifier for distance, and camera (telescope) quality. A simple interface for the player would be map view, as you suggest. One, high-res images of a world can simply allow zooming the map view to the appropriate fidelity. If you have 10cm pixel maps of some parts of the Mun, in those areas you can zoom in that far---the map view can show you where, too. At certain resolutions, the geologists can tell you something about what might be of interest---then the map can be marked as you suggest. Note that this is not grossly different than what players have now---"science from orbit above" or "visual survey of are X." The missions would actually be less tedious/clicky, you'd not have to click the thermometer at the right instant, you'd simply put the probe in orbit, and let it take pictures. Done.

Planning a mission (and having fun) mean having the ability to predict where and what amount/quality of data you will have BEFORE committing a lot/all effort in it.

Lastly , Trial&Error is hardly the best game-mechanic we ever made (euphemism again).

Sadly, trial and error is THE game mechanic in stock KSP, like it or not, else we'd have dv readouts, etc.

I am not suggesting trial and error, however, I'm not sure where that is coming from. Data would be provided as earned by the player. If the first orbiter of Jupiter had tried aerobraking, you can bet that would have been "trial and error" at some level. There is a reason that the model for planetary exploration has been: flyby --> orbiters -->landers. Space exploration in RL tends to be iterative, in KSP most people send Jeb to the far extent of the solar system for years in a room the size of a hall closet.

Clarification time : I'm not on a position purely antagonist to yours.

So far I'm saying the randomness you described so far, do not meet (my) criteria to take sub-biome from 'mindless copy' into 'place actually interesting'.

Like I said, my idea for a randomized KSP can have all the possible worlds decked out in great detail as YOU would prefer (I would, too, actually, I'm for better worlds, period, what tool they use doesn't matter to me at all ).

Your suggested use of "Random procedural generation of content" to "increase the game length" isn't new... but it is more often than not what I would call "polishing turd", hiding an inability/unwillingness to generate REAL content with copypasted crap. (Opposite to more skillful use of procedural generation)

It's not the "opposite" of anything at all. Right now there is scatter, entirely for (meh) aesthetics. If some of the scatter was actually interesting, and moved the game forward, then it gives the player a reason to drive a rover around for some reason other than seeing how fast it can go before it flips.

It depend entirely of how much data and tools you intend to require a player to micro-manage. Being forced to send cheap probes at every single place to know if it worth considering a rover or a manned mission is precisely what is painfully frustrating in Trial&Error gameplay.

Beware of fake difficulty.

That's not "trial and error" gameplay. Trial and error would be sending the rovers with parachutes not knowing if the atmosphere would allow that. Planning would be finding out if there is an atmosphere first. Realistically, you'd know the broad details from ground-based astronomy, anyway, so I'm unsure what your complaint is. Jupiter was broadly understood before Pioneer. The cool trial and error might be tiny moonless, or rings (if dangerous) that would not be visible, and you'd not be prompted even for SoI interactions until the probe/crew could observe them. So your first mission to Jool as a nice periapsis set, and when you actually get inside Jool SoI, you start discovering moons that were not visible from Kerbin. Your craft might have to make a few burns if the planned trajectory intersects one. To me this would be nothing short of awesome. Apparently you might not want to discover new moons.

Not what I understood.

But then I fail to see a point to customize your system at all, nor in doing that "each game" as you suggest. There's strictly nothing good in separating all players experience (plus DeltaV map). So at best it would be an optional Joke-feature.

Joke?

I don't care even a little about other players' experience relative to mine. I don't care about a delta v map, either*---for a stock KSP argument, you can't care either, as KSP does't provide the player with dv information, so a map is semi-worthless. (*because it's not relevant to stock KSP, I actually DO care, KSP should have a dv readout). That said, there are apps to calculate transfer dv, so you include one in stock along with readouts for stages. Dv is now not a problem. Besides which, as I said explicitly, any randomized Kerbol system would be an OPTION for players who are burned out on the same ole system. It's a replay feature.

I see no interest either in having around thirty planets/moons alternating.

Rather than making 8 awesome planets with all interesting feature, you are suggesting to make each game have only 50% of all features in a non-optimized way... and for what ? Fake longevity ? Because that's what that would achieve, forcing players to rebuild all infrastructures with no added-value.

Forcing a rebuild? People actually keep saves? I start a new career from scratch any time I change anything, from a KSP patch, to a new mod. I'd rather explore a new system than care about some "infrastructure" I built X versions ago.

And again, you presume that 30 planets means each is worse. Right now, almost all the planets need a major overhaul. Adding 20 more would change quality... not at all. Kopernicus has resulted in at least enough cool new planets to choose from. All as good as current stock. The reality is that if a squad artist has to hand-draw every cool feat on every planet at a scale-size under 1 meter, we'll get a cool new Gilly in maybe 10 years. Any such detail needs to be automated at some level, or will be a proverbial needle in a haystack for a player to stumble upon.

Randomness mean that you cannot customize/finely tune "scatter level detail" (Area of Interest) to fit optimally :

- Interesting reliefs (crater/hills/crest/access to other site and ISRU).

Untrue. You could add random scatter predicated upon any number of factors, including the existing "biome" (perhaps all craters could be given a "crater" biome), better applied "biomes," the math related to procedural cratering (look for the edges, that's a crater rim, etc).

- distance with others AoI.

Because computers are bad at calculating the distance between 2 points?

- distance with ISRU place.

Because computers are bad at calculating the distance between 2 points?

- planetary balance in term of science-point quantity.

Balance is easy, they just need to have a plan (which they don't). Roughly figure out what the ideal science per planet or typical mission should be, and tweak. Trivial.

- predictability and fairness of career playthrough. (if enough "scatter level site" are too easily accessible, why explore the rest ?)

Given the dumb, "biome" system, it's trivial as well. Only allow 0-X interesting areas per "biome." You land in biome-1 6 times, and manage to find one interesting scatter, and the game has decided there are to be 2 in that biome. Most people won;t try for them, they'll realize that they only get one every few EVAs in the same "biome," and if they get it, bonus.

Worse, it lead to a FAILSAFE generation of bland but FAILSAFE area that have FAILSAFE parameters so it stay FAILSAFE anywhere.

I suggest no failsafes, I actually wish it was possible to lose career, short of playing "grind" mode (I fond the "difficulty levels" just tedious ways to make you grind in early career).

No bets you know the importance of Game-design, but the impression that random generation is a good idea every time is hard to shake off. That's why I insist so much on this.

Embrace the random. You can do rather a lot with paper, pencil, and good ole dice. There is a tendency in the modern world to assume everything should be "modeled" causally, when realistic outcomes can be done quite well with the odd random number generator.

Again, Random generation isn't a goal, if it bring nothing good to the game -not even replay-, don't do it.

It is a goal. An exploration game needs "fog of war," period. There is no exploration in an "explore in the age of sail" game to sail east from Spain if you know the Americas are there instead of the Indies. What would be the point of watching the New Horizons stuff come in if we already knew what Pluto looked like? Last month was literally "what KSP should be about," IMO.

And let's be clear : We are NOT talking of crafting Area-of-Interest with an unrealistic attention to details here. This is simply improving what SQUAD did often like the Mun's Crater-ravine-Crater area, except visible from space, with more sub-biome and with a "slightly painted" road to make it more enjoyable to plan and send a rover.

I agree with this. Who wouldn't? That's one mission, though. One career. Then what? Always land in the same spot. I suppose squad could concentrate just on one cool spot per planet, and leave the rest as cruddy as Minmus, then you know to only land in the cool spot.

Edited August 5, 2015 by tater

[tater]

Skillfully handcrafted designed interesting spots, will not only be old once you explored them, and boring once you know where they are, they will also have the problem of contrasting to the areas around them, that are not so skillfully hand-crafted, especially considering, that alien-civilizations are a no-go (unless that changed).

This is spot on.

I wonder if any of you two even read my rather longish post a page ago. I have the impression that both of you want to creat little bubbles of interest, Tater wherever the player needs them, you, as them being predifinded and designed fixed locations. And both of you are right in the points you bring up against each other. In essence, both would feel like ´Mass Effect´ - as they are both ´bubbles´; one convenient, the other distinctive.

I'm for both, frankly. Alternately, you can populate all the words with interesting detail at the scale it happens in real life, then astronauts will find it about the right amount of time.

That's the "goal" really, to have interesting outcomes about the right % of the time. Most EVA on places like the moon will be routine, areas like Hadley Rille might be expected to have more interesting probabilities, but if you happened onto an open lava tube you could walk right into---win!

The bubbles we have to consider here are the existing geomes though. Within each of those we want special places. We need not generate them. They already exist. Each geome already features a bunch of different places. They are defined by the terrain properties, altitude and slopyness. Also just taking samples/doing experiments at different (otherwise not varying) lacations within the same geome should have scientific value, upto to a threshold, just by confirming existing data, so to say.

Good point.

To design interesting places would be like ´creationism´. But to generate them for each of the player´s landing site would be ´anthropocentric´. Those interesting sites are not all that unique in real life, either. The conditions have to be ´right´ (e.g. water can be found in deserts by digging at the foot of the dunes, rather than their tops). Crop said conditions from what the game provides, and you have your interesting spots, based on (a simplyfied) logic and ´natural laws´, as should they be.

Yeah.

What drives me are nice vignettes of a view, frankly. Certainly places you land just feel authentic, and that's where exploring---entirely for its own sake---is fun.

[youkofoxy]

I think the term for someone who does it (what I did) is "HALO drop"

But "Bomb Drop" fits best.

procedural good? well, the XCOM fans were screaming with signs reading: "Procedural Maps now"

and they were happy when it was announced that each map would be procedural.

On the other hand, fans of The Elders Scrolls, know that Arena was criticized for "more of the same everywhere"

Furthermore, Portal does not use this technique in their levels, but Invisible Inc uses.

and Okami does not use procedural in its characters, but it's not hard to find a game that has its characters generated by a program such as Human make.

like any piece of code, there are those who are incredible and those who are a disgrace.

and the matter interesting? Mars is the sheer lack of variation, but the Earth is incredible on variation...

in fact, although forest be ... well, forest. every corner of the amazon forest is different.

Now, walk in a pine forest ... is incredible despite the view is almost the same as of a desert, in terms of change ... at least that's what I think, I never walked into one.

the only solution is to mix even the best key to making a unique environment.

handcrafted with procedural more seasons.

this is my solution to your dialema.

when more techniques better, is what the Terminator 2 taught us about tricking the brain.

when you think you understand it, everything has changed and you are lost again in a familiar environment, but unknown for some reason.

[Mr. Scruffy]

[...]What drives me are nice vignettes of a view, frankly. Certainly places you land just feel authentic, and that's where exploring---entirely for its own sake---is fun.

I am not sure if i get the meaning of this sentence (vocab issue). But if it wants to say, that special places should also look special, at least from close by, that shouldnt be impossible either. Upon landing (and at intervals while being landed) the engine could check the surrounding terrain for ´special conditions´ (those would have to be predefined for each geome) and texture them accordingly. Say on one planet/geome a sloopyness of >75° over a length of more than 10m would constitute a special site, cause it would show some sort of geological layering, telling us about its history (let´s not get into wether this makes sense for an atmosphere-less planet, please). The engine would just have to check for such terrian within the ´physics-distance´ and when it finds the right conditions, supercedes the normal texture with the ´layered´ one.

This, so far, was about giving rovers a point in career.

----

For procedural generation of the kerbol system, i´d even expand "tater´s ideal", the system in which you´d have to explore its present state on your own, mostly. Ideally, the system would be generated by roughly the same forces that govern real system formation and thus also have a history, when we start to play in it - and science should give us clues to that, also. In a perfect game, we could form our own hypothesis and, if we want to spoil ourselves, check them, by actually watching the formation of our kerbol system as it happenend. But that´s dreaming, i know...

- - - Updated - - -

[...] There's strictly nothing good in separating all players experience [...]

Just think about the stupidity of this argument: If it were true, every civilization game ever would have taken place on earth. There would be one and only one Tropico island, for everyone to play on... I trust i need not bring up more examples. The quoted statement is rediculous. It´s only merit is from a dev´s point of view, as it makes balancing easier. So, when i hear it from a dev, that´s one thing. If i hear it from a player it sounds like ´but terrorism and WMDs!!´.

You know, when i play KSP today, i have to constantly actively ignore the fact, that whatever i do - it has been done a thousand times before ; and whatever i see, it has been seen by a thousand people before. "To boldly go where thousands have gone before!"

"The dV tables are invalid for other systems!" - yeah exactly, that´s sort of the point. It appears to me, that you want some sort of Fallout/Elder Scrolls in space. A sequence of ´quests´. The same for each game. Nifty handcrafted content, but dull as hell gameplay-wise (IMHO). I dont want a KSP, for which you can write a ´walkthrough´ with detailed steps on how to ´win´ (or complete) the game - as would be possible for a civilization game with no random map-generator (and in which i may only play one civ, too - no choice there).

If player experiences are not to be seperated - why have difficulty settings as detailed as KSP already features? Why have moddability? This attitude against procedural system creation almost seems like ´but the earth must be the centre of the universe´ attitude 500 years ago - just because there would be possibly millions of other systems within the realm of possibility, it wouldnt mean the end to Kerbol as we know it. It would still be the stock system and the one most people talk about. Jool will stay Jool, Duna will still be Duna. But imagine the explosion of diversity that would open up and all those interesting AARs, reporting about other systems...

The seperation of the experiences of players is why i spent way more time playing games, than i do watching movies. I might even go as far as to say that ´seperating the experiences of the users´ is sort of the ´raison-d´être´ for games in general, especially sandbox-games.

----

Edit: Back to rovers, again.

Intertesting sites based on terrain logic (rather than random or handcrafted) would come with an inherent challenge in designing and driving a rover that is actually able to reach and draw samples from said sites: It may not be all that easy to get to a point of 75° slopyness and stay stable there for the time it takes to do the experiment. Reaching the lowest and highest point of elevation within a geome comes with challanges all on their own, too (Moho-holes f.e.). It may happen that you´d discover such a site, and despite the fact that you came with a rover and the site is rather close-by to the landing site, you are still unable to reach it, due to the rovers design. So, such a concept would not only give a reason for rovers to drive around some, but also reason for putting extra thought on their design based on where you want to go with them and what exactly you plan to achieve with each individual one.

Edited August 5, 2015 by Mr. Scruffy

[Mr. Scruffy]

[...]- Small rover become useful within AoI without having to drive for hours to already covered biomes

[...]

You wanted an example : Imagine Duna

- One area of interest is at the equator, deep in the basin (large regions of dark soil), it's a "magnetic anomaly" [...], if you look closely, you realize that the anomaly is at the center of an absolutely flat place over a great distance, like... a runway.

- One 'cave' is 50km away North, at the edge of the depression/ravine. It is ISRU friendly.

- Next one is closer at the top of the depression/ravine it's just a big noticeable rock... if you look closely you'll notice a "road" to the 'cave'

- Still toward North is a mountainous crest. One huge point of XP at a place hard to land or drive to.

- Another one is 30km away, it is basically Olympus Mon. If you look closely you'll notice something like a roads.

- Next one is 30km away at the limit of the polar cap, it's a large crater with something at its center [...], it is also ISRU friendly.

- Last one is at the pole.

...if my calculation are right, you now have a line for an awesome Rover-trek from the equator to the pole with refueling place for rocket/shuttle, base, and so much science you can train your scientist here (meaning regular travel)

[...] [emphasis changed by me]

How long do you suppose it takes to drive a rover from duna´s eqautor to one of its poles, while exploring stuff in-between? Not to mention that "runways" and other mysterious stuff pointing to alien influence is pretty much announced to not be in (and would be pointless, unsolvable mystery in the ´Lost´-sense, as long as there is no story to them). Just recently, i tried to drive a rover, that had landed quite close to the edge of the southern polecap to ice-free terrain, maybe 5°, but no more than 10° north of its previous position. After driving around for like an hour, i gave up on it. Based on that, I´d expect covering 90° to take days.

EDIT: Well there is an event announced (

), as i just learned, that will feature just that: A roverdrive from duna´s eqautor to one of its poles. The distance is given as apporox. 1,000km and it will be streamed over the duration of 5 days. Edited August 5, 2015 by Mr. Scruffy

[micha]

As someone else said, once you've done it once, any hand-generated special terrain will not really have any replay value. It's also a LOT of work for the devs, something which may come eventually, but for now I daresay they have their hands full with other work. And while procedurally generated content may not have the sparkle of hand-crafted content, it is better than nothing and essentially available for free once the underlying routines are written (which is no small taks either). After that it's just setting (or generating) seeds and having the engine create content for you. Still a lot of work, but not as much as modelling precise scenes. And it's also a LOT kinder on your Computer specs (no need to keep procedurally generated content in memory when not looking at it).

Replay value in KSP is different things for different people. However, in the context of (surface) exploration, it is having more stuff to do. So I fully support having some sort of "interesting feature" generator which can populate the area around you with some special stuff. Although I'd argue against random generation and have deterministic generation instead - that way if you find something neat, you can tell other people to land at X and they'll also see it.

Even though it's outside the scope of this thread, I also like the idea of a procedurally-generated solar system with a "fog of war" forcing you to send out probes first to discover the details of planets and moons before sending out landers or kerballed missions. Again, this should be deterministic in the sense that you should be able to share the seed value with other players so they can enjoy the same system.

Finally if the deterministic generator only places "interesting" stuff in a few locations around a planet, rather than everywhere, that could tie in nicely with the contract system which could then generate contracts to land people near the special content. So yes, you can still discover it on your own by luck (I found an anomaly on the Mün by luck), it would mostly be by following contracts, and those contracts should only be generated after some sort of planetary scan.

[Mr. Scruffy]

Well, Micha, if interesting sites would be determined by terrain properties, and the terrain would not be known without scanning it first, it neednt even tie into the contract system: You could chose to land at some easy, flat but featureless location for some easy science, or you could land at some harder, rugged spot with lots of different terrain features, each of which offering new opportunities for experiments, harvesting which will not only place higher requirements on your lander (and choice of site, and your skill in spot-on landings) but also on your rover to actually be able to get to them. In other words: It would be entirely up to you, the player, to chose how far you want to take it - you get to chose you own mission difficulty and, in the case of success, get rewarded accordingly.

---

Going off on yet another tangent to this thread: I dont like the contract system. I think IT is the heart of grind. It´s not the fact that we need money and stuff to proceed, but that we have to subdue our plans to whatever funky missions the game conjures up for us. If i would just be rewarded for succeeding in what i wanted to do in the first place and wouldnt have to launch another 5 satelites into some random orbit, destroying immersion rather than promoting it (and those are among the more immersive contract-types - dont get me started on the others...), before i may steer the program back into some reasonable path of progression...

Yet another EDIT (rambling): I think Squad got the career mode ´wrong´ (=not according to my personal taste - yours may of course vary) in almost all aspects. If i was in charge, i´d probably do away with ´science´ as we know it alltogether. Rather, it would be ´data´ on the bodies. We´d want it, for its own sake. To collect all data there is to collect on all the bodies would be the goal of the game, the precentage of your having accomplished that your score, if you will. You´d have a budget based on time and reputation, tied to time-intervals. And you´d have limited time (amount of which setable at gamestart, with ´infinite´ being an option), or REP would decay and you´d have to prevent it to hit 0, or else -> game over (no one on kerbin cares about the space program anymore and it thus gets canceled entirely). You´d have goals, active at all times, and fullfilling contracts would be an exception at best - you´d only do that if you hit real dire straits (they´d probably come as pop-ups, not in a facility of their own). Parts would have an unlock date tied to each, which could be accelerated by funding additional research towards it. Parts would have XP like pilots (to be hired, not bought) do, which could be used in conjucture with some time and some money to improve certain aspects of them. Reaching ´goals´ (unsurprisingly they are about collecting as much data about the kerbol system as you can, mostly) would give you nothing but REP, but a higher REP-score will increase your monthly (or whateverly) budget... Also there would be no data or REP to be gained on Kerbin (but above it, starting in the upper atmosphere, i´d say), and most certainly not on the KSC itself.

Example: You start your game with 5 REP and 50,000 roots, say. Your basic buildings, as they are all most basic, do not cost maintainance, and the big4 (Jeb, Bill, Bob and Val) are the enthusiastic silverspoon initiators of the program, who volunteer to work for it for free. Each Kerbin month (day is 6 hours, 30 of them is a month) REP decays by a fixed 1 point, after having paid you 10,000 roots for each point (suddenly difficulty settings boil down to just these two values: REP-decay and roots_per_REP). On your list of goals for Kerbin you find (among others): ´reach upper atmosphere´ and ´suborbital trajectory´. As thermometers should be avaiable from the get-go (how else would heat-tolerance of parts be known?), you also find ´temperature in upper atmosphere´ in your list of currently attainable goals (those for locked science parts do not show up yet). Each of these 3 goals will give you 1 poínt of REP upon completion, but the latest only for a complete series of measurements (say 3-5, for simplicities sake) and otherwise only the according fraction of 1. Ideally you manage to slap together a rocket that achieves all three in one go. Say this rocket costs you 15,000 roots. You are now down to 35k, but have between 7 and 8 REP (depending on how complete your series of temp-scans is), paying you between 70,000 and 80,000 roots next month (after which REP will drop to between 6 and 7). With that, you may be able to do your first buildings upgrade. Be careful though: It will start to cost maintainace from then on.... You might also choose to promote research of a specific part, so that it unlocks before its scheduled date. Or you could, though this would be more common mid-to-late game cause you´d need the part-XP, have work on an improved version of an already unlocked part started. The later two could require their own building (the one we currently spend SCI in, i suppose), with the latest needing at least one upgrade. The funding might not be directly, but through employées working in them, number of which determining progress (think x-com) - firing people costing a bit of REP mayhaps.

Alternatively you might choose to wait for better parts even before you do your first launch. Some of them might be only 100 days away. By then you will have 50k+50k+40k+30k = 170k roots, but only 2 measily points of REP left. You better show some results before the month is over, then, and they better be impressive or at least continious, or else your REP-bar will stay dangerously close to zero and "game over" and your coming budgets will be insufficient in the long run, despite your saved up ´fortune´. You really dont want to spend your last money on a duna (or any other longterm) mission, when you only have 2 REP left - if you let it come this far, the public wont care about your promises of great discovery next month, when the mission arrives - when REP hits 0 the program gets canceled, no matter what.

Edited August 5, 2015 by Mr. Scruffy

[micha]

Mr.Scruffy; agree with your points, but I think there's still scope for "interesting" areas even away from "interesting geometry". Also the poles would just be chock-full due to the way the geometry system works.

But yes, the contract system does need a bit of an overhaul. OTOH, you're not forced to take contracts - I only take contracts which further my own in-game goals. (Eg, I'm building a base around the Mün, so I'll take high-paying contracts to, eg, send a satellite there. Along with it, I'll send another station module. Then I leave the satellite until I get an Impact contract (mod), or use the satellite as a ScanSat (mod) satellite, irrespective of what the original contract wanted me to do with it). So still somewhat constrained by contracts, but I never spam 5 launch a satellite contract for no reason other than funds.

Better would be a larger overhaul - but anything I can think of requires time-based mechanisms, which Squad hasn't, and, by the sounds of things is unlikely to, implement. For example, periodic budget payments, based on rep which decays over time, and missions attempted/completed, so you can't just timewarp to get infinite funds. But you'd also need other time-based expenses and events (eg, life support, Kerbal Construction Time, etc) to balance it. There'd still be a place for contracts, which would be private companies wanting you to do stuff for them.

That way your primary mission (get all the science) is mainly funded by the budget, but you can also earn pocket-money and extra rep with contracts.

Not a trivial thing to design and balance so it's fun.

Fully agree on using Science as the "high score" though - this is something I've posted before too.

[Mr. Scruffy]

[...]Better would be a larger overhaul - but anything I can think of requires time-based mechanisms, which Squad hasn't, and, by the sounds of things is unlikely to, implement. For example, periodic budget payments, based on rep which decays over time, and missions attempted/completed, so you can't just timewarp to get infinite funds. But you'd also need other time-based expenses and events (eg, life support, Kerbal Construction Time, etc) to balance it. There'd still be a place for contracts, which would be private companies wanting you to do stuff for them.

That way your primary mission (get all the science) is mainly funded by the budget, but you can also earn pocket-money and extra rep with contracts.

Not a trivial thing to design and balance so it's fun.

[...]

lol - read my edits...

And here´s another one: I find it sooo wierd to choose to go without a meaningful time component in a game that has the words ´space´ and ´program´ in its title, when both these terms are so closely linked to time. As is the word ´career´ (which just strikes me now as yet another misnomer by squad. We are not playing a career in KSP - this is not ´Pirates!´ after all. We are not playing as an in-game individual, like say a submarine captain. Hence we are not playing out a career. Call it ´full campaign´ or actually just ´program´-mode instead).

A second edit: If i had to judge KSP, i´d say that its basic idea (exploration of a fantasy star-system at feasable scale with seamless flight from launch pad to smoking crater + building your own rockets) is excellent - something i had to wait for way too long. Full score there. Basic implementation? Still really good, imho. Other people may complain about lack of the various models (aero, or whatnot) and the "outdated" (as i read somewhere) graphics, but to me, both are adequate. 3/4 score there, i´d say. Campaign-mode, though? Ugh. Barely sufficient, i´d say (D- in school grades), ´mistakes´ in almost every single sentence, if it was a written text. But it is more like a multiple choice test - with ~3/4 of the crossed answers being the wrong ones, imho.

---

"Should a game like KSP, being about building up a space-program, have a flat* time-dimension?"

a) Sure, it will need (and already has) timewarp anyways, being a space-game. It will span a long game-time almost by definition. This will serve well to fill ingame-time during the long transfer waiting periods.

No, it could be exploited if we fail to balance it properly. Rather have otherwise hired staff be ´enslaved´ with a strange exponential cost-model and an obscure, unrealistic contract system, instead of a budgeting-dynamic.

*) "flat" as opposed to "looped" as the orbits are with self-repeating states if you just wait long enough (talking about "exploits"...).

---

They crossed b. I think they should have opted for a. And if i was to put every single decision about ´career´-mode (right down to terminology!) into questions like this (granted: hopefully with less biasedely phrased answers), i think, i would end up disagreeing to 75+% of the answers given by the devs.

I will admit though, that the ´career´-mode as i thought about it, say a year ago, would have been substantially different in parts, to how i´d like to have it, now. The process is iterative, after all. But my answers to such basic questions as above would not have been much different.

Edited August 5, 2015 by Mr. Scruffy

[Mr. Scruffy]

BTW: I am not sure you need life-support or contruction time in order to balance ´flat time´. Wages for your crew and scientists as well as maintainance costs for upgraded buildings might suffice. For crew in flight, the wages could just double with each passing month (or so) until they return to kerbin, giving an incentive to not drag out manned missions for too long and to exchange stations´ crews (which life support does not, i assume?). And of course decaying REP.

I think the game would actually be easier to balance. It is a simplification over the current model, really. Two currencies, instead of three, and one of them is merely the growth rate of the other, which is in essence the only thing left to balance against (well and time). And i think for simplicities sake, to get a foot on the ground, so to say, with ´flat time´, life support and construction can be left out, ´for now´. Not to say they couldnt be introduced, if the implemented simple stuff proofs to profit from a little more complexity. I´d like to build rockets x-com style (hire engineers, number of which limited by facility upgrade and determining constuction speed), but i dont think its really neccessary for proof of concept of the basics and could be introduced later.

[Kegereneku]

NOTE : Trying to answer everything despite the repetition have been pretty long and straining. If you see 'sign of exasperation', please take them on a purely rhetorical value. I don't have anything against anyone personally but some arguments are best addressed... directly.

Have a nice reading.

I wonder if any of you two even read my rather longish post a page ago. I have the impression that both of you want to creat little bubbles of interest,

I did, we are just not forced to agree with your personal opinion or consider it objective.

To avoid repetition, let's just say that trying to make a game "like reality" do not necessarily make a better game, no matter how realistic you believe that game is in the first place.

Next, assuming you read our posts, you fail to address the point that simply adding 1000 sub-biomes will only increase the already grindy nature of Science-point. In result it is pointless IMHO to be alarmed of "Place of Interest" becoming boring when your proposal so far suggest worse : more clic-fest, no reason to go to interesting relief, no reason to do a long rover trek...etc

Consider the following :

- Randomly generated sub-biomes : You land a Kerbal/rover, gather immediate science around like a automate (roll, stop, science, roll, stop, science).

If the Reward was balanced down, you only do that 100 more times to break even.

- 10 'Places of Interest' per planet :

You have the choice of launching 10 classic missions to classic-biome and gather (say) 50 sciences each,

Or you land Kerbal/costly rover at 'Place of Interest' and you can get up to 200 sciences if you manage to reach all more complicated micro-biomes.

The goal is more or less to recreate the : "Cheap Probes" versus "Huge Mission" paradigm choice many dream of.

Just think about the stupidity of this argument: If it were true, every civilization game ever would have taken place on earth. There would be one and only one Tropico island, for everyone to play on..

For starter, this is quite a logical fallacy to equate my sentence into "you want to railroad players".

Especially since taking your logic people should be FORBIDDEN of sharing similar experiences and playing by the same rules just because you pretend that it is inferior to what you personally want.

You do realize that the whole "Challenge" forum depend entirely on physical and gameplay parameter being the same to all, right ?

Simply said : Different game = Different game-design. You do not necessarily need to randomize a map and doing so wouldn't necessarily improve thing. There is too many game example of that.

So, since you talked of stupidity you feel like a child who think he can put everything randomly in his cake and it will somehow become the best cake ever even if he don't know the first thing about cooking.

Sorry to not give a more throughout answer to the entirety of yours posts. But most of it have been covered already.

How long do you suppose it takes to drive a rover from duna´s eqautor to one of its poles, while exploring stuff in-between?

I don't know but it should easily be more entertaining than exploring copypasted sub-biome pattern for the same duration.

The way you didn't realize that "magnetic anomaly" can be interchanged with ANY science-you-deem-acceptable (like Mystery-Goo) hint that you are just being contradictory here.

The reasons I talked of adding "subtle road" was precisely to alleviate the travel time by making it easier to go very fast.

Right now, I wonder why I keep trying to make you (& tater) understand that Realism don't necessarily lead to interesting Gameplay.

It's incredible how some players can denounce "unrealistic feature" that are more often than not the very reason the game is playable in the first place.

** Sniped for brievety **

I'm sure you'll understand if I avoid a Fractal-Quote-Answer apocalypse.

I'm not yelling (at last wasn't at the time), I was emphasizing points that I think you are missing.

What you are describing is -to me- roughly the same as spamming sub-biome (in the form of thing to do at scatter-level), and randomly generating "Place of Interest" into "whatever the player consider as such" miss the whole point of the things.

Anyway, I can tell you just want a gameplay fundamentally different from a "improvement of previous game mechanic".

That idea of yours to simply gather data in a Completionist way until all is known, is a pretty distinct form of entertaining with different expectation.

From my point of view 99% of a planet reliefs is decoration and only 1% is fun to play in, so we should encourage design-variety through them (ex : a Rover for cave could not use solar energy). Exploring where those 1% are should NOT be a dumb use of Trial&Error, it should be a voluntary choice. "Do I go to that place or do I not for equal reasons ?"

Equally, what is the point of taking 10hours to scan entirely a planet when this time is not interesting in itself ? It is simply Masochist and potentially PUNISHING !!! "did you scan for 10 hours the planets ? No ? Then you don't get to know the fun place..."

Again, you are pushing a feature for realism, despite it being having no point in term of gameplay. That's one of those features many would support once they discover the full implications. Like Aerodynamic-realism, Deadlyreentry, AntennaRange, LifeSupport, RSS...etc.

Next : The minor Trial&Error aspect of KSP only concern easily reversed or bypassed aspect of the game. (A crash can quickload, running out of fuel can change the mission objective).

But the longest aspect of KSP pretty much avoid trial&error : No luck based mechanic (random failure), everything is predictable (+ Maneuver node), all data required to do the mission is available (because this is the thing to do) and you are more and more encouraged to plan contingency plans for long mission rather that abandon/redo them entirely.

The lack of dVreader is a game flaw that ought to be corrected once SQUAD have the Engineer do it, maybe stop underestimating players ability to learn, or as pointed out in my signature : realizes that they built features pointless without it.

At that point I leaning to believe they KNEW it was needed but wouldn't put it until they could insert it within a game-progression.

The utter inability of current Career-mode to make you feel you are slowly-building a space program is certainly at the root of your apparent disdain for "keeping a Savegame" and wishing to increase game-length through forced micromanagement. Beside your clearly different expectations.

If you can't understand how everybody playing on the same system and sharing the same experiences is a GOOD THINGS for KSP, we might as well stop the discussion here. Not caring about other wishs & expectations is not constructive.

Untrue. You could add random scatter predicated upon any number of factors, including the existing "biome" (perhaps all craters could be given a "crater" biome), better applied "biomes," the math related to procedural cratering (look for the edges, that's a crater rim, etc)

You still don't understand... A "mountain biome" don't actually care about what is hilly, what is a steep slopes, what is a ravine or what is a peak. A crater-biome don't even care about the center of the craters. DETAILS need to be added manually or dumbed-down so much for a computer that they'll lose any interests. NOT FORGETTING HOW as said numerous time, if you are only subdividing a biome/task to do you are only increasing the grind-fest, not the actual fun.

Hence my point that QUANTITY can not always replace QUALITY. If you don't gain anything with random procedural generation, don't ***** use random procedural generation.

Embrace the random. You can do rather a lot with paper, pencil, and good ole dice. There is a tendency in the modern world to assume everything should be "modeled" causally, when realistic outcomes can be done quite well with the odd random number generator.

I retract what I said, you don't actually know much about game-design.

Paper pencil and good old dice game required and still REQUIRE A LOT OF HUMAN-INTELLIGENCE !!!!! There is STRICTLY NOTHING RANDOM in the way they where created. And even if today computer-TOOLS are used, it take extremely subtle decision and arrangement from Sentient entity such as human to turn random data into a GAME.

Now you can say I'm yelling at you.

It is a goal. An exploration game needs "fog of war," period.

You are being stupid here... "Random Generation =/= Fog of War" no matter how you look at it. It's confusing the actual gameplay of exploration, with the mean you produce the "contents" that you will be exploring.

As I'm trying damn hard to make you understand : You can get bored of exploring 100 randomly generated/copypasted pattern on 50 planets, faster than exploring 10 handcrafted place (of choice !) on 8 planets.

The more I try to grasp your wish and expectations, the more I think you simply want another game because the feature that make KSP, KSP, simply can't entertain someone eternally.

I agree with this. Who wouldn't? That's one mission, though. One career. Then what? Always land in the same spot. I suppose squad could concentrate just on one cool spot per planet, and leave the rest as cruddy as Minmus, then you know to only land in the cool spot.

Just for the record, you missed the point of the dichotomy between :

- Place that are more science-efficient for cheap easy mission.

- Place that are more science-efficient if the complex mission is capable of getting all sub-zone.

1) You get to chose whether or not you go with one or the other.

2) You get to use more of the gameplay features (EVA or rover)

3) The improvement is of higher quality than using badly "random generation" to generate less refined contents.

As someone else said, once you've done it once, any hand-generated special terrain will not really have any replay value. It's also a LOT of work for the devs,

I disagree with both statement.

- hand-generated special terrain can easily have greater replay value than auto-generated one.

- It is also not necessarily a lot of work. handcrafted don't mean "crafted-with-masochist-mean". Procedurally generating PHYSICAL-ROCK like the Asteroid and simply using your human brain to place them and designate already-generated relief into place that now have a reason to be explored, is not particularly hard... for an human.

FINAL NOTE : To me the discussion is becoming quite long and circular, don't be surprised if I just opt out.

[Mr. Scruffy]

[...]

I did, we are just not forced to agree with your personal opinion or consider it objective.

To avoid repetition, let's just say that trying to make a game "like reality" do not necessarily make a better game, no matter how realistic you believe that game is in the first place.

Next, assuming you read our posts, you fail to address the point that simply adding 1000 sub-biomes will only increase the already grindy nature of Science-point. In result it is pointless IMHO to be alarmed of "Place of Interest" becoming boring when your proposal so far suggest worse : more clic-fest, no reason to go to interesting relief, no reason to do a long rover trek...etc

Consider the following :

- Randomly generated sub-biomes : You land a Kerbal/rover, gather immediate science around like a automate (roll, stop, science, roll, stop, science).

If the Reward was balanced down, you only do that 100 more times to break even.

- 10 'Places of Interest' per planet :

You have the choice of launching 10 classic missions to classic-biome and gather (say) 50 sciences each,

Or you land Kerbal/costly rover at 'Place of Interest' and you can get up to 200 sciences if you manage to reach all more complicated micro-biomes.

[...]

You need not agree with what i wrote, and you may not consider what i said as being objective, but to partake in a discussion you should consider what the other participants said at all, and not what you just assume they said. The ´sub-biome´ proposal, which openend the thread, had been long discarded, mostly, in this discussion, pages ago, and i, for one, never argued for it (and tater explicitely said so as well). ´No reason to go to interesting relief´ is the worst part here, as what i suggested goes diametrically in the other direction, suggesting exactly to make interesting relief worth investigating. No wonder, this discussion seems circular to you, then, i guess.

--

About system-generation: When will it finally get around to you, that this would not kill or even harm the stock system in any way? One reason why it wouldnt being said ´challenges´. An extended common ground as one single map offers is desirable. But so is diversity. In an exploration game, a single setup can only take you so far.

And no, said systems need not be inherently inferior to the stock system, if it is done well. Allow me to expand a bit on this: A couple of years ago, i wanted to write an exploration game based on Dan Bunton (RIP)´s "Heart of Africa" (1985, iirc). But i wanted it to take place on generated continents. It was surprisingly easy to write said generator, even for a beginner like me, complete with biomes (real ones, vegetation-types) and rivers. All i did was:

- start with a flat surface

- go over each tile and roll the die wether a vulcano erupts or a meteor impacts (raise/lower elevation of the tile and surroundings)

- erode (flatten)

- 15 iterations of the above going over the whole surface

at this point the continent´s shape is done

- have ´wind´ carry humidity from the eastern sea

- have it rain off a bit on each tile and a lot more on the westside of mountain ranges

- Have rivers start at extremely humid places and run downhill

- Set temperature according to distance from equator

- Set biomes according to temperature and humidity.

Done. It was quite crude, but the results were quite acceptable (so, i do know a bit about ´cooking´). Overall, i think wether generated systems would be interesting depends on how well the process of making them is done. Maybe we should focus on discussing what it would have to be able to come up with in order to yield interesting results. For the purpose of the discussion so far, though, i was always assuming a quality (as in diversity, plausibilty and detail) that is at least equal to what we have now.

Generated systems, btw, need not even be a contradiction to handmade content - that could still be placed within generated worlds. Considering the effort it takes to make those, i´d think they (the handcrafted ´interesting sites´) would be good DLC material - a nice cash cow for squad: Have an artist department of 1-2 people churn it out as fast as they creativity allows and charge one or two bucks for each. Yeah, to be fair: Of course a system-generator could also be DLC.

Reading your final sentence ("It is also not necessarily a lot of work. handcrafted don't mean "crafted-with-masochist-mean". Procedurally generating PHYSICAL-ROCK like the Asteroid and simply using your human brain to place them and designate already-generated relief into place that now have a reason to be explored, is not particularly hard... for an human."), i think we might actually not be so far off each other, actually. To reiterate: I suggested to have the game crop interesting places from variable within the game and then put a different texture on it. You suggest objects being handcrafted and placed manually. The main difference i see here is the one betwen logic and whim. One can be followed and to some extent predicted, the other can not. I, for once, would like to be able to look on a map of a body and be able to tell ´there should be some interesting spots around here´ rather than ´i hope the devs have sprinkled some goodies for me here´ - but still, one does not exclude the other and i could well live with both.

About the nature of the rover-drives: You´ll have to admit, that your suggested duna-drive is a bit excessive, now dont you? I mean, 5 days of driving around is a bit too much, right? If a landing site gives you enough in range of your rover to intensively explore for 1 or 2 hours, that should suffice, now wouldnt it? And i think ´interesting spots´ based on topology would do exactly that, even on places like Minmus. But then again: No, i wouldnt mind one or two ´rocks´ put on ground of it by the devs as special points of interest - be they randomly put (i´d prefer that) or at fixed locations.

EDIT: Geome setup: Each geome could (and need) be better defined and diversified against the other geomes to make ´topoligically interestings spots´ work (better). Different sets of parameters should make different kinds of spots interesting in the various geomes and the first experiments´ texts should hint you to what those are. Some TISs could also be seen from quite far away, while for others the difference in texture is so subtle that you can hardly notice even from just a couple of meters away in brought daylight, so you´ll have to search by the parameters (like underground liquid water, say). If polar caps pose difficulties because of how planets are made up by the game engine, just define them as (mostly) TIS-less, with maybe only a slight color-variation indicating an unusual element being frozen at some places, but neglecting terrain-properties (this would come closest to the ´sub-biome´ suggestion, but is just a needed exception).

Edited August 8, 2015 by Mr. Scruffy

[Kegereneku]

Mr.Scruffy, with all the respect I give to longly constructed argument that are not quote-war. I do not see answer to my/ours points in your updated-posts and you do not seem to be getting answers out of mine, so we are still in disagreement.

Even if you think we are not that far of each others, generating automatically sub-biome based on randomly-generated topology is still not different from random generation since the result is random (and worse if homogeneous), dixit what I said about increasing the click-fest and grinding.

(btw : so far we also have yet to address the topic of how the player will distinguish in-situ/non-map the different sub-biome)

At least I think we can agree that we do not see "RANDOM procedural generation" the same way. The personal anecdote you described is everything but random (you choose the equator to be hotter for ex) and the finesse of the details you required to actually improve the game was clearly inferior than needed in KSP. Different Game = Different design = Different method = different finesse. Hence to me you are suggesting to improve an AAA game using Candy-crush design.

To rephrase it : What is the point of your sub...topological-biome/geomes if it is only dividing the click-fest from probes unto rovers ?

I'll wholeheartedly agree that Rover really need incentive but making them the next "only efficients mean" of getting science (due to the necessarily world-wide homogeneity of automated topo-biome/geomes generation) would simply change the problem.

A note on my "huge rover trek" : You seem to believe that I want to force people to accomplish the entirety of the trek for it to have any meaning, this is not the case.

By focusing so much over the sheer leng....Epicness of accomplish suck mission, you are missing the implication that -even on a smaller scale- by manually choosing/refining/improving/balancing where Greater quantity of Science-point are localized (with a voluntary dichotomy between high-science area and normal-science area) you are adding a different gameplay that properly make use of the relief in a way that no automated generation can distinguish (unless you've all hidden sentient AI from me).

so, consider the following along what I've said :

- Each Area-of-Interest can go from intricately complex (ex : all science at various place of a giant rock that EVA have to climb) to methodical (ex : you know you have to stop every 100m-1km starting from the center of a crater/magnetic thingy to get all the science)

- Each AoI have been placed to make the best use from an human perspective of the already generated reliefs (this cannot be automatized)

- Each AoI position have been considered & balanced for an optimal user experience (see High-science / Low-science dichotomy)

- Subtle "road" (aka : rover-friendly surface) could be generated in between on the height map (best use of procedural generation) to ease high-speed travel.

- Overall AoI diversity is made unique and fulfilling by not auto-spamming them.

And don't be fooled that it look like more work, this is nowhere complicated or long for human (by that I mean it's been done in other games), it would actually take longer to automatize right (or why so many modern game still use tiles-based map, it's easier to generate and not f***-up).

The way I see it, you can procedurally generate Place-of-Interest, just not their placement and balance. Note that making them random would bring strictly nothing for reasons already explained before, Players like having the sames maps/challenges.

Example of Area of Interest :

-type 1 : concentric circles of science, to be manually placed and sized to match BIG crater size. (plus the mentioned mean of delimiting the sub-biome)

-type 2 : non-concentric circles of science, to be manually placed over a gigantic mountain, a distinct one as you go higher.

-type 3 : area of science, to be manually painted to match canyon/crest

-type 4 : rocky-terrain with big pebble, various size. A sub-zone by themselves, generated or not around other AoI feature.

-type 5 : dozen of small-asteroid-like rock with 1 sub-zone each, randomly generated over an non-randomly positioned/wide area.

-type 6 : dozen of small-asteroid-like rock with 1 sub-zone each, those generated in a random-looking line toward the next AoI.

-type 7 : single BIG-asteroid-like rock with 3/4 biomes, the biggest generated alone, like a rocky mountain.

-type 8 : single BIG-asteroid-like rock with 3/4 biomes, small surrounded by pebble and/or smaller asteroid-like rock.

...etc

I think what describe the best my preferred approach is :

- classic biomes that can be 100% explored with classic means.

- super-Area that can only be 100% explored with more complex means but only break even if ~25% of one is explored.

The goal being to make unique zone that you are encouraged to explore (which wouldn't be the case if they were too numerous to balance, leading to grindiness) and which require Rover and/or Crew.

The only downside I see is that, knowing which Instruments to use is still a trial&error click-fest. It would have to adapt.

[Mr. Scruffy]

Procedural means that you go random within certain parameters following a certain method. That is exactly what the continent generator did. The results were different each time, yet the process of generation was always the same. That it was set to place the highest temperature along the equator is just one of the process´ defining properties. The same, i´d like to see in defining areas of topological interest. Without generated systems, those areas would always sit at the same locations (so for the stock system your assumption, falsified by my mere opinion and that of many others here on this very board, about what "players like", would be served). Those properties need not be global for each body (or even the entire game), but, as i tried to outline before, should be defined individually for each geome. The purpose of these areas would be to give your rover utility on your average landing - not just only for some uniquely special handcrafted locations (which may well be in the game as well, for all i care - the two things are, again, not exclusive to each other).

If you dont bring a rover, you land, do your experiments and that´s it. Chances are though, that there might some interesting sites around (or maybe not, if you just landed on a featureless plane), which, bringing a rover, could be investigated. If you dont bother doing it, maybe because you find it grindy, that´s your choice of course, but if you dont enjoy driving your rover, it begs the question why you brought it in the first place.

If you only have 1 or 2 special sites per planet where using a rover actually makes sense, you will only build and bring rovers to visit those locations:

- "Should i bring a rover with me to duna?"

- "Well do you want to visit site A or B?"

- "No, not really, i am just shooting for pretty much anywhere - it´s my first visit there"

- "Well, then, dont bother."

Somehow, i doubt that any landing site of the various mars-rovers was handcrafted and picked as landing site based on that, while anywhere else on Mars rovers would just be a pointless grind.

But again: I am not arguing against hand-crafted sites - i just think they would not help much making rovers worthwhile outside of them, as they should be, imho.

EDIT: It´s kind of amusing (to put it nicely) that you are telling me that the program i wrote, including the lines calling the random number generator, is ´anything but random´, just because i wouldnt place everything on it based on random numbers, which is kinda the core of ´procedural´. Guess, what: The biomes werent placed randomly once the generation had reached the point to place them, either, but instead by well defined parameters. The random part only takes place in the beginning of the process, in which ´tectonics´ is (very crudely) emulated. Everything from there on is based on logic, aka the process in ´procedural´. And yeah, you do need to sprinkle in boundaries and parameters to have it all make sense in the end - the sea level, for example, was set (iirc) after the tectonics-part, so that i would always have a certain desirable percentage-range between land and sea tiles.

Anyways, let´s just drop system-generation from the topic and stick with the stock system for the sake of the discussion. My mistake having brought it up - sorry for that. If we do, the random part of the placement of the interest sites drops out of the equation and the game would be left with only the procedure of placing them - always in the same spots every game, cause they are defined by logic, the terrian at hand, and the defintions consituting them within each geome. Altering their spawning points would be a matter of modifying said definitions (and NOT a matter of fiddling with coordinates, as would be, if they were placed ´by hand´). The devs decide to change the terrain of any of the planets? No problem - no need to fiddle with the interesting sites: they are procedually generated after game start and not read from a file. It´s virtually impossible to have one that shouldnt sink below the ground due to terrian changes.

Actually, thinking a bit further, you could generate your own interesting sites within the game, theoretically (engine limitation may not allow for this: ) by digging a hole, for example. The game would just have to call the placing method of the geome for its interesting sites while a hole is being dug, in order to update them and spawn new ones (or delete existing ones), if conditions have changed. I think if that could be pulled off, doing experiments would come a lot closer to feeling like, you know, doing experiments. Or, put in other words, we´d start to play on EVA.

There could be geomes with defintions giving sites at conditions that do not exist within said geome at game start, but which could be created by the player. "Hmmm. Maybe if i dug a hole over there at this low spot, i´ll find something interesting."

Edited August 9, 2015 by Mr. Scruffy

[tater]

I'll try to keep the quote apocalypse to a minimum.

So, since you talked of stupidity you feel like a child who think he can put everything randomly in his cake and it will somehow become the best cake ever even if he don't know the first thing about cooking.

This is a Straw Man. We saying nothing of the kind.

That idea of yours to simply gather data in a Completionist way until all is known, is a pretty distinct form of entertaining with different expectation.

From my point of view 99% of a planet reliefs is decoration and only 1% is fun to play in, so we should encourage design-variety through them (ex : a Rover for cave could not use solar energy). Exploring where those 1% are should NOT be a dumb use of Trial&Error, it should be a voluntary choice. "Do I go to that place or do I not for equal reasons ?"

I'm not even a little completionist. Science is organized trial and error. That's what the scientific method is.

Equally, what is the point of taking 10hours to scan entirely a planet when this time is not interesting in itself ? It is simply Masochist and potentially PUNISHING !!! "did you scan for 10 hours the planets ? No ? Then you don't get to know the fun place..."

Yes, because the player must sit there in real time for 10 hours because there is no time warp... Oh, wait, there IS time warp. Place orbiter in polar orbit. Do something else in your program, next thing you know the planet is mapped. If you can only manage a flyby encounter, then you might image part of the world as you zing past. I think this is a cool idea.

Also, we have auto-generated contract content, integration with this is trivial. All the contracts asking for surface science in location X could be predicated on the map data. If you forgot to bring enough fuel, and your probe goes zinging by Duna without orbiting, then you only get surface contracts for the side it took pictures of. If there were cool "areas of interest" as you suggest, you might send a Ranger type probe to image that area, making all the various surface contracts happen in that area (the player driving the contracts, instead of them being 100% random). Note that the contracts could vary based on the resolution, as well (have low and high as options). High-res areas being weighted by the contract system. Space-Based contracts could try to fill in gaps. "Visual survey" is really "mapping," anyway, so you'd be asked to try and fill in the maps via contracts if you care to.

Again, you are pushing a feature for realism, despite it being having no point in term of gameplay. That's one of those features many would support once they discover the full implications. Like Aerodynamic-realism, Deadlyreentry, AntennaRange, LifeSupport, RSS...etc.

Better aero, better reentry, life support, all these have people arguing against them as well. I don't get the point, your 2 statements seem contradictory. Realism can add to gameplay, it's not about realism, it's merely the scale at which you bother to model. The point of the game is exploration, so why not make exploration as interesting as possible? IMO, from a gameplay standpoint, making science interesting means making it actually useful for gameplay, vs it being a grindy point collection scheme.

That means that you send an orbiter not to "collect science (points) from orbit," but because it allows you to see the surface clearly in map mode, and perhaps pick landing sites (that would require a new mechanism)---or spot interesting places to land ("areas of interest" as you suggest). Sending a pressure sensor to a world might unlock something like what exists in that mod that shows a spacecraft path more accurately within an atmosphere---so you get "science," and in addition to points, you can now do things in game you could not as easily do before (land your next duna base part near the first one more precisely). ISRU already does this scanning wise. I think it's good gameplay vs land for points. YMMV.

Next : The minor Trial&Error aspect of KSP only concern easily reversed or bypassed aspect of the game. (A crash can quickload, running out of fuel can change the mission objective).

But the longest aspect of KSP pretty much avoid trial&error : No luck based mechanic (random failure), everything is predictable (+ Maneuver node), all data required to do the mission is available (because this is the thing to do) and you are more and more encouraged to plan contingency plans for long mission rather that abandon/redo them entirely.

I think that from a gameplay perspective both concepts have merit. Gameplay is never one choice is good, and the opposite choice is bad, often both are good, but in different ways. Take part failure. Apollo 13 would be fun in ksp. I had something sort of like that happen once due to a design fail. I did not mirror a fuel line as I thought, resulting an a very unbalanced craft under thrust once the tanks drained. Getting a burn to return was hard as the craft would tumble. Fun. With KIS/KAS, I have had a ton of EVA fun with rovers since I actually crash those, unlike spacecraft. I have then rebuilt them in order to drive back: making a functional tricycle out of a 4-wheeler after a flip trashed a wheel, for example.

The lack of dVreader is a game flaw that ought to be corrected once SQUAD have the Engineer do it, maybe stop underestimating players ability to learn, or as pointed out in my signature : realizes that they built features pointless without it.

At that point I leaning to believe they KNEW it was needed but wouldn't put it until they could insert it within a game-progression.

Yeah, this decision is baffling to me. No progression is even really needed, you can note your total DV, and see that if THAT craft made it to the Mun, then any craft with similar dv should be able to do so.

The utter inability of current Career-mode to make you feel you are slowly-building a space program is certainly at the root of your apparent disdain for "keeping a Savegame" and wishing to increase game-length through forced micromanagement. Beside your clearly different expectations.

I'm all for a time-based mechanic. MY career games take a long game-time, as I don't warp missions to completion. Life support alone forces this if you have multiple missions in flight.

If you can't understand how everybody playing on the same system and sharing the same experiences is a GOOD THINGS for KSP, we might as well stop the discussion here. Not caring about other wishs & expectations is not constructive.

It's completely constructive. By that rationale, only people playing Minecraft using the same seed can discuss minecraft, or share experiences. That's aside from mods. I know you use an alternate tech tree, does that disqualify you from this discussion? I would say, "no, it doesn't disqualify you."

My ideal KSP having a new system most every new career is not going to happen, anyway, so focusing on that is pointless. It was primarily an argument for the benefit in gameplay terms of semi-random content. Semi-random because the quality of the worlds in question can be as arbitrarily good as you require for my argument, Squad could spend many man-hours on each world making them as hand-detailed as you like, I'd only vary their position, and scaling within some range (past 3.2X scaling in my experience they get "muddy"). Anyway, this is off topic, as it will never happen.

You still don't understand... A "mountain biome" don't actually care about what is hilly, what is a steep slopes, what is a ravine or what is a peak. A crater-biome don't even care about the center of the craters. DETAILS need to be added manually or dumbed-down so much for a computer that they'll lose any interests. NOT FORGETTING HOW as said numerous time, if you are only subdividing a biome/task to do you are only increasing the grind-fest, not the actual fun.

The biomes are painted over the map. That's all they are. I have never suggested biomes be randomly painted. Yeesh. I'd go in and apply crater biome to all craters. BY HAND. (except those that might be "special" crater biomes, like "Young crater" or whatever is of special interest. The current "biomes" are sloppy, I'd fix that, ideally. I'm against the sub-biomes, also, as I have said. Regarding crates as an example, I should be able to know what most "biomes" using data you have gained (photographic mapping, for example). This would give gameplay value to science, and would let the player more easily drive the missions.

Hence my point that QUANTITY can not always replace QUALITY. If you don't gain anything with random procedural generation, don't ***** use random procedural generation.

No disagreement. Conversely, if you CAN gain from random, procedural generation, don;t discount it automatically because the word "random" appears. There is a time and place for both kinds of content, period. I am arguing gray, you are arguing black and white (that any/all random is bad).

I retract what I said, you don't actually know much about game-design.

Paper pencil and good old dice game required and still REQUIRE A LOT OF HUMAN-INTELLIGENCE !!!!! There is STRICTLY NOTHING RANDOM in the way they where created. And even if today computer-TOOLS are used, it take extremely subtle decision and arrangement from Sentient entity such as human to turn random data into a GAME.

Now you can say I'm yelling at you.

Well, since I've actually worked on some games of that type, this is funny. Games with dice, yet you say nothing is random about them. LOL.

Games of a simulation variety (at any level of fidelity) are about realistic outcomes (within the "universe" simulated, so Sci-fi or fantasy have "unrealistic" premises, but can be internally "realistic" within those changes to actual reality). Such outcomes can and are achieved using random content. In a single player game, random encounters, for example. The table rolled on to produce them is an act of intelligent creation, and it is tweaked until the outcomes feel right (realistic outcomes). You need not predetermine the location of every facility on every world ahead of time for a sci-fi RPG, for example. Some can be random. Plot drivers you'll drive the players to anyway. Same goes for board games, though those are typically 2+ players, so everything is placed by players (though perhaps hidden, and only discovered randomly). I won't digress farther on board game designs, I was having such working discussions about them on the USENET in the 1980s.

You are being stupid here... "Random Generation =/= Fog of War" no matter how you look at it. It's confusing the actual gameplay of exploration, with the mean you produce the "contents" that you will be exploring.

As I'm trying damn hard to make you understand : You can get bored of exploring 100 randomly generated/copypasted pattern on 50 planets, faster than exploring 10 handcrafted place (of choice !) on 8 planets.

I've managed to argue without calling you stupid even once. I'll leave our relative intelligence for the reader to determine. I have presumed that most issues here were a function of language (presuming english is not your native language), and that you might not be clear on subtle distinctions of language. Regardless, it is possible to disagree without being rude.

I will say for the nth time: even in my random solar system, I granted you all the worlds as hand-crafted as you require. So in arguing that point, you must remove any random quality of the worlds from the argument, and say why it would be less interesting for repeat players (I also stated it would be an option for repeat players) to explore a new system. People seem to enjoy outer planet mods, and other kopernicus alternatives, so apparently some agree with me about new content to explore).

Fog of war (not knowing what you should not know about the Kerbol system) REQUIRES something to be randomized. It is certainly equal. You pop on the forums to explain a problem you are having in a game with a certain world, and people will tell you exactly what to do---say aerocapture. They'd tell you where to put the periapsis for a kerbin-planet transfer. In a randomized "advanced" play mode if you did this, no one could help you that specifically, they'd instead have to tell you how to determine the proper periapasis yourself. Similar result, but one is very mechanical, the other is "do this experiment/science." (note that perhaps the game could give you this data with the right science earned (again, USEFUL science, instead of points to grind).

Just for the record, you missed the point of the dichotomy between :

- Place that are more science-efficient for cheap easy mission.

- Place that are more science-efficient if the complex mission is capable of getting all sub-zone.

1) You get to chose whether or not you go with one or the other.

2) You get to use more of the gameplay features (EVA or rover)

3) The improvement is of higher quality than using badly "random generation" to generate less refined contents.

Sigh. This is a non-argument, particularly 3. How about I argue that you replace 3 with "The improvement is of lower quality than using much better than hand-generated random content."

You would rightfully say that I cannot make that claim, just as you cannot define random as always bad. You conflate random with bad and randomly generated at every turn. Random can just as well be "semi-randomly placed, hand-generated content." A lookup table. If the player is in a crater of type A, then there is an X% chance he discovers an interesting feature appropriate to that type of crater, hand-designed by squad.

You would place such a feature in ONE crater, and have the player have to land in that spot to see it. If such a feature would be "realistic" to find in 20% of craters, then I would simple have it applied with a 20% chance to any particular, appropriate crater a player lands in (one such determination per landing within some distance, so if you landed in the same crater 100 times, and it was;t there the first time, it would continue to not be there, ever, check another crater).

Random, hand-generated, and the game is not forcing the player to land exactly in spot A, every, single game. Just land int he right kind of crater and you might find one.

Edited August 9, 2015 by tater

[tater]

Options on topic (assume all useful improvements on "biomes" are done for all, BTW, so that they make sense):

1. OP's desire for procedural sub-"biomes."

Pros: more science points with fewer missions and more clicking (is this a pro?)

Cons: more science points with fewer missions and more clicking

2. Additional, hand-designed* (unless procedural can do as good a job) content possible nearby a given landing site (limited to X total per "biome"), placed in a "semi-random*" way. (% chance per landing site within XX range by biome or other factors the game is capable of tracking)

Pros: New content. Possibly more science points with fewer missions and more clicking. If the content is interesting to see, then it is also interesting to look at. Some is the same interesting content as #3, but the exact location will vary game to game. Some might be linked to the already random ISRU deposit levels as a visible sign that this area is bad/good for ISRU?

Cons: Even if hand-designed by Squad to be awesome looking, the word "random" is used in reference to their possible placement, which makes them "bad." For reasons. Also, additional science points, which is not really a "pro."

3. Additional hand-designed content certain at a tiny number of places per world.

Pros: New content.

Cons: always in exactly the same spot, making 99.999999% of the places on each body pointless to visit from a gameplay standpoint.

[Balto-the-Wolf-Dog]

[WILD RAMBLING]

I don't support randomly procedurally generated sub-biome (two separate thing here)... because we need REAL contents here !!

Procedural generation is all well&nice to fill up unimportant/decorative stuff like the reliefs, texture... but it is a sign of weak-design when it concern a deeply critical gameplay features that require complex consideration to please players.

After all...it is really a "point of interest" if a sub-biomes is as boring and grindy as every other biomes ? Nah! Procedural generation is how we got those contracts-description that no one probably read anymore and lack of attention to details is how we get brain-damaging contract (warning : drama at 120% here).

And so, I opine that if we really want to encourage surface exploration, like "Build useful heavy-rover and base", you have to create handmade masterpiece Area-of-Interest... around key reliefs, and visible from orbit. So you can actually plan on the long term rover mission and think your mission another way. Prepare ISRU and so...

Ideally it would be a different paradigm in gameplay...

- If exploring a dozen normal biome with cheap probes got you (ex) 20points

- Exploring one of those "point of interest" sub,biome would only give you 10points with a drone too cheap, but 50p with a rovers, 100p if you have a Scientist to reset heavy-science and 200p if you planned a rover to get to the equivalent Area-of-Interest that was put voluntarily 50km away with a hand-made "relief" for Rovers.

A zone you really want to explore yourself. And who know, maybe you could craft actual cave there ?(cave are for now incompatible with the tools used to make planet relief)

[/WILD RAMBLING]

I'd definitely take designed over procedural. KSP is small enough geographically for it to be practical. Procedural really comes into its own in extraordinarily large or infinite environments. Terrain could use as similar treatment, possibly part of the same project. Planets in KSP are certainly a hell of a lot less barren and bland than, say, orbiter (completely featureless) but for the most part everything is smooth, rolling hills and dunes. I'd really like to see some more exaggerated topography, at least in some cases such as Laythe. What we have for Duna makes a fair amount of sense geologically, but Laythe would probably be tectonically and volcanically active. Eve may be as well.