The amount of effort it would take to implemented compared to how many people have talked about wanting this kind of feature.
1 Like
Bearing in mind that half of my suggestions were for other peoples interests, and I do realize some of them are wanted by a minority, if someone, not necessarily me, was to do a survey, what kind of numbers would you be looking for before an idea was “worth considering” and not “crazy”.
It’s not a hard threshold, just that it needs to be a higher priority than any other new features on the list of potential features to develop. So the feature needs the highest effort / number of people wanting the feature ratio.
2 Likes
Putting aside the AI’s behavior, no was has commented yet on the question: If the AI evolves to a higher life form, would it necessarily have to evolve shorter generations than the player? The simple, if not necessarily scientific, way would be for, as the player advances their microbe a million years at a time, the AI tribal creature lives one tribal generation a million years at a time. Is their already code that would need to be changed for that? And if all there is currently is a cap on how far it can go compared to the player, would it take that long to move the cap? A simple toggle “let AI advance past you at an unscientifically slow rate” would likely fix half, though certainly not all, of the issue. A meet in the middle of sorts.
I will grant that a microscopic creature making it into a macroscopic creature’s intestines would require creative programming, but would stage 3 interactions with stage 4 or 5? Assuming at some point, i realize this is still some time off, you want AI tribal creatures to act a certain way when the player is in the tribal stage, why not, when the game has progressed that far, just let that be its behavior even if the player is still in stage 3. Would tribal AI NEED to act different if the player is something other than a tribal creature?
Edit: Also
Would the AI be prevented from researching things the player does not want in their game?
For performance reasons we could just make it so that the galaxy simulation only kicks in once the player gets out of the early space stage.
3 Likes
And the most important stars in the sky would be generated in aware or awakening when you may start using them for navigation and stuff?
1 Like
Well only star positions would need to be generated once a night sky is needed, if we go that hard into realism and don’t just use a fixed night sky star pattern.
Using stars to navigate sounds like again one of those super hard to implement features that only like a couple of people would ever appreciate or notice. That idea easily overtakes the AI overtaking the player on the craziness scale.
2 Likes
By “navigating using stars” I mean something like how people had used polaris to note where north is a long time ago. Perhaps your civ could also discover how a northern/southern star works in the society stage?
1 Like
That was exactly what I was expecting. It’d take a lot of programming effort to ensure stars would appear that could be used for specific navigation like that and then implementing the surrounding features related to travelling in a way that makes sense. Also such a focus on stars could make players realize more readily that we just have a fixed number of pregenerated star patterns. So a feature like this would also put more pressure on making dynamic night skies (instead of just a couple of easy to make premade textures). So in summary, again this would be a real nightmare in terms of how much effort it takes to implement.
2 Likes
(Post deleted by Author)
Never mind, I miss read your post.
1 Like
Also for galaxy simulation, do we want to actually somehow simulate an accurate-sized (most likely spiral) galaxy? I can guess there are some tricks which should allow us to decrease the amount of computing resources that will need to be spent on simulating such a large object.
1 Like
It’s highly unrealistic to try to even store a single full-sized galaxy (so not even talking about trying to simulate anything in it), because if we assumed that each solar system could be stored with just 32 bytes of information (which is basically not enough for anything more than just a few star properties and a couple of planets with no life whatsoever) then it will take terabytes of storage per galaxy. So Thrive saves would be terabytes in size in the space stage. That’s just completely impossible.
I’ve previously talked about this here:
2 Likes
On the topic of handling galaxies, I have an actively horrid, overkill idea.
Without good ftl, you’ll never need to simulate more than a few dozen light years in real time. The problem with a chunking system like minecraft or something is that once you place a block in the chunk, you have permanently increased the size of your save file. Simply running in a straight line and breaking one block every chunk will eventually fill up your hard drive. If thrive worked like that, and actually tried to simulate a galaxy, you could simple hurl a single rkv (relativistic kill vehicle) at one planet in every star system the second you git Kardashev 2, regardless of anything about it, and wait, as your save file ballooned in size at a rate proportional to the speed of your missiles. That’s unworkable. Your whole computer would die. But, logging that you fired one missile at a ton of chunk isn’t that hard, and generating what would happen if a given chunk got hit with a missile is doable. So, if you set the save file to forget that you did anything a million years after it goes into effect, you’d need to shoot more than a million missiles a year to balloon your save file that much (assuming you store them efficiently, say a list of 64 bit chunk ids for each activity worth logging or maybe even a 64 bit chunk idea and hashed target descriptor,of some kind, and that the player doesn’t mind a save of a few gigabytes). So, how is it immersive to not remember the planet you shot specifically? To trick the evolving changes? Well, either the planet is within a few dozen light years, or there’s no way you’re getting a response for a century. Say, when the timer on the thing you shot into the chunk runs out and it hits it’s target, the chunk is procedurally generated. Any time you procedurally generate the chunk, it’ll be the same for a given seed, so we can throw out our chunk. Then we generate an effect. The effect has to be deterministic so we can throw out everything there too. basically make it such that when a ship of yours enters the chunk all events can be responsibly reconstructed without loosing anything, so the only data that needs to be save is huge lists of packed binary or player activity logs or something. Well, how is this realistic either? You can generate uch more, non deterministic detail when the player is doing something with a chunk, which is fine, because think about it, either the chunk is within a few dozen light years, or you aren’t seeing it again for a century or more! If you come back 200 years later and everything has changed you won’t feel like you’re scammed. basically, detail isn’t permanent, all broad strokes can be reconstructed from solely the world seed and the players actions in the last million years. The player can do a LOT but, as what they do scales up, you can blur more. If the player is conquering all words they can get, they won’t mind nearly as much as the first world they specifically target, you could make the amount of save/ram space taken up by a player scale slower than linearly by loosing detail. The imperium of man won’t much mind if one of their colony ships colonized the wrong world, and they have FTL. Once you’re on that scale, who cares if instead of storing every ship you launched the game just stores a colonialism seed and procedurally generates your target worlds.
Yes, this is comically ambitious, and would require a TON of dev work, but I think allowing for comical scale works fine in theory, it’s not actually impossible, it’s just a lot of work. Fully deterministic procedural generation wouldn’t necessarily work for ascension, but just say the player is offered to save their changed worlds to a file before they fall out of scope and get purged.
1 Like
Is that idea from the perspective that the player controls a single ship that can be only in one place at a time? Thrive space stage is explicitly not planned to be like that. Instead it is a continuation of the industrial stage, meaning any player structures or probes or whatever must show up as units at all times, and any colonies are controllable at any moment. This is pretty much totally incompatible with the fact that even loading all the positions of stars into memory would not be doable (as we need to assume that someone with just 8 GB of RAM will try to play Thrive).
1 Like
I have no mouth but i must scream.
1 Like
Not necessarily one ship, but without ftl it’s not like there’s any unity between a galactic government. you could imagine a colonial power in the 1700s seeing a summarized, limited version of their colony’s government menus, and a several month lag before any policy goes into effect. I think the omnipotent RTS gameplay design makes sense with ftl, but without it, I’d say only the space within a light year of your head of state could get the kind of control. You could dictate policy outside of that, but might not know the local head of state until after they die. I imagined moving around in my chunk description because there is no actual reason to limit the player. An itinerant court is a thing, etc. If the game has ftl, again, my ideas are double crazy, but if not, I don’t see how controlling everything equally makes any sense anyways. And uh, if the game is planning ftl, I wasn’t conclusively informed.
3 Likes
This is probably why space stage civs are, well, knows as “civs” and not “states”, since there is a limit to how far they can stretch their control against the upper barrier of speed and large distances.
3 Likes
Yes, because my scifi dreams need to be fulfilled (and to correct the single biggest mistake in the Big S game that was controlling a single ship in space stage). Also otherwise Thrive would just end at our current human technology level, because well all technologies we don’t currently have would be science fiction.
2 Likes
Alright, very sensible, even if i fully disagree with the logic that restricting ourselves to known technology restricts us to tech we already have. I am not discussing “Thrive with a Sci-Fi ban”, I’m discussing “Thrive based on things we know.” Alien microbes are sci-fi, too.
Examples
Known =/= acquired. We cannot build a space elevator, why? Not because we don’t know how it would be done. It’s because we don’t know how to get enough of a material strong enough into existence in a single line, how to get such a line into space, or have anywhere near enough money to do it. How to get to other star systems or build megastructures or whatever? Dyson swarm. Requires no new tech, just a lot more of it than we’ve ever built. We can do a lot of stuff, we know most of the steps to making most of an entire tech tree, we just don’t have the resources and/or the means to do something we understand in theory. We know how to make antimatter, if need be, we just haven’t ever made more than a few atoms (I think maybe just separate protons and electrons but don’t quote me) of it. With only an invention of a magnetic bottle, we have an antimatter bomb. Add something that somewhat reflects gamma rays, we have antimatter rockets. The inventions we need don’t violate physics, they’re just more than we have right now.
I have not sufficiently used hide details in the past, it’s great!
Anywho, nice to know ftl is definitely planned.
1 Like
Without building it we won’t know for sure that such a material can even exist. So any technology not built by humans (either right now or in the past) needs to be considered science fiction because we can’t be sure that the problem is actually solvable. And I don’t think being reasonably confident is enough, because fusion energy has been talked about and worked on for a really long time and talked about being feasible relatively soon, but so far no success. So even if right now there’s scientifically plausible ideas about a space elevator, actually building one might run into roadblocks indefinitely.
Okay, these are fair counterpoints, basically situations where the science is proven but the problem is manufacturing or scaling up the manufacturing. Though I’ll still note that for like antimatter manufacturing a totally different approach is needed to make it anywhere near economically realistic for it to be used for some technology. So again we are speculating that if there is a need scientists or manufacturing process engineers can figure out how to make the process a few orders of magnitude more efficient (without needing to break the laws of physics).
3 Likes