How far is to far with automation?
- Thread starter Taiine
- Start date
-
FTB will be shutting down this forum by the end of July. To participate in our community discussions, please join our Discord! https://ftb.team/discord
For automation it really is a personal choice. I can not honestly tell someone that there is such a thing as to much automation. Personally I like automation that is interesting to build and configure. To one person a quarry to smelter is to much for another it is their bootstrap they do on every build.
About the only thing I do not like automating to much is mining I personally dislike the use of quarrys and autominers of most sorts. At least on a large scale. But that is just my personal take on things.
to each their own.
About the only thing I do not like automating to much is mining I personally dislike the use of quarrys and autominers of most sorts. At least on a large scale. But that is just my personal take on things.
to each their own.
Its too far when I everything in the game is automated and you could just stand in a 1x2x1 space forever.
For me, the other parameter that defines too far is when I am running out of memory on my server which has 8 gigs of memory and I am the only one that uses it. Take that in for a second. I had a single player world that required more than 8 gigs of memory to run.
For me, the other parameter that defines too far is when I am running out of memory on my server which has 8 gigs of memory and I am the only one that uses it. Take that in for a second. I had a single player world that required more than 8 gigs of memory to run.
I have to find that video where a guy made turtles that mined resources to make more turtles that programmed them to do the same thing, endless turtles...
This is the earliest one i could find. One of the forumers here was also working on a Swarm Mining program, his/her name slips me now...
Do keep in mind, though, that at the time of this video, turtles did not require fuel to run. To actually manage this, AND the refuelling of every new turtle created, would require a much more complex program... I currently don't know of any CC turtle program that has managed to overcome those hurdles. BUT if that was made possible, you could have a real World Eater (And CPU destroyer) turtle program. One that could possibly fill a DSU in a reasonable amount of time, but would require some MASSIVE hardware to actually handle it =S
Creates enderchests filled with fuel and one for dumping. Supplies each new turtle with one of each.
It's SkyNet gone wrong... I really need to learn Java now. Picture this:
A turtle, that can make other turtles to do his bidding, makes attack turtles to attack anyone near the base of operations, crafting turtles to craft, mining turtles to mine resourses, engineer turtles, WAND TURTLES! (that have wands with warding foci).
Good game humanity, good game.
Beacons aren't that hard to get in the first place.
Too far would be pressing a button and creating ichor from ThamicTinker with no effort. (Although, planning on trying that. XD)
Are you sure its not the nazi in your screaming how far is too far? JK of course
Too far is when you automate something that you used to find fun. I stripped some of the automation from TiC smelteries because I liked watching the liquid metal pour out into the molds. I also built my pack to have more adventure aspects so I wasnt sitting in my fully auto base and fiddling with the machines. At one point in 1.4.7 I felt like my base was playing me.
just curious, what did you add to your pack to encourage adventuring. Right now im using gregtech iwth the 'better dungeon loot' config enabled, so beating dungeons gets you iridium, lots of steel, etc. We also use Dungeon Pack too, so we have a lot of options.
The ones I know of and like are Twilight Forest, Witchery, Ars Magica 2, Better Dungeons, Dungeon Pack, Roguelike Dungeons, Aether 2, Natura, Christmas Festivities, Deadly World, Dimensional Doors, Farlanders, Biomes O Plenty, Highlands, Gravestone Mod, Hardcore Ender Expansion, Ruins, and Unique Artifacts.
I guess it depends on how often you die and lose everything. I guess if you tear through everything with an Unstable Resonant 9-Bladed Lightning Sword while wearing invincible Runic Nano-Quantum Force Armor, and slurp up the entire world into an Applied Energistic Enderpouch of Fortune, then yes, you might get too much loot.
In this program you can see each turtle only mines a column straight down to bedrock then returns to base to drop resources. Without some serious modification of the code, there's absolutely no way to guarantee that each newly created turtle will have the amount of fuel needed in order to reach it's designated digging spot and go back to base. What would end up happening is that some of the turtles will get stuck in mid-air with no fuel to continue operating, and just clogging up the air-space for the turtles that do have enough fuel to continue operating.
This program was made when Turtles did not require fuel, and will probably still work if you enable that option in the CC config, but to actually manage a self-replicating turtle swarm mining program is a DAUNTING programming task filled with challenges at every step of the way. First, since turtles started requiring refueling, the most efficient method of mining with turtles in regards to fuel consumption is the rabbit-hole method, where each turtle goes and mines in a sort-of checkerboard pattern a hole straight down, and mines everything that don't match the "noise" blocks you programmed the turtle to ignore, in all four directions. Each "turn" of the turtle does not consume additional fuel, only the actual digging and moving up/down/sideways consume fuel. That way each turtle mines a + sign pattern. By clever distribution of turtles, you can mine an entire chunk, not leave one block unmined, and consume as few fuel as possible. Now for the creation of turtles, that's easy to do, there's all sorts of automation nowadays to automatically construct the things, and it's not really that hard to auto-install programs in them, and also name them too. But you'd need to program some sort of pathfinding algorithm for the swarm, so each turtle knew it's relative coordinates in relation to eachother (which you can do via rednet), in order to avoid collision, which is a very hard thing to do, and also every newly created, named and programmed turtle must know exactly where it would need to go and mine down, and also request enough fuel to complete the operation to go there, and get back with the resources, which is extremely tricky to do. Relative coordinates via RedNet GPS helps, but don't solve the problem altogether - especially as you get farther and farther away from Home Base and the location of the GPS, so you would need to continuously expand the GPS network along with the swarm, which means auto-production of advanced computers, wireless modems, etc... You'd also need to update every turtle that ever was created with new coordinates every time they got back to home base, and make sure that no other turtle has been given the same relative coordinates as any other turtle, which is also a very hard thing to do. You could do it less efficiently fuel-wise and only make it so the turtles continue to mine exactly like in the video, which would make pathfinding and coordinate-giving easier, but still non-trivial.
It's no coincidence that no-one i know so far has managed to create such self-replicating world eater turtle design with respects to fuel, pathfinding and constantly updating new coordinates for every turtle - especially since the production of new turtles can and will rise indefinitely as more turtles are introduced in the swarm and more resources gathered. If it was easy someone would have already done it inside the new fuel-requiring environment, but so far i haven't seen a design that does work indefinitely and continuously out of it's own resources. If you introduce outside fuel production for the new turtles being created, that might ease the problem. But still, it doesn't solve the problem you have when more than one turtle returns to main base at the same time, which is also something you have to program for and avoid happening.
There are plenty of swarm programs, though. Just not self-replicating ones and self-expanding ones. And they deal with many of the problems i've outlined above, but not all of them, and as a result are very complicated, and many times, not as efficient as more traditional turtle-mining programs, or even the Buildcraft Quarry when supplied with appropriate power. They're still an interesting programming challenge for those who want to go up to the challenge. Me? My LUA skills are barely even enough to understand the code written in these programs. I could never write a program like that on my own, but i am familiar with the hurdles the turtle programmers go through to make possible for turtles to do what they do, at least enough to know that there isn't a simple "fix" you can apply to some old code and make a functional world-eater/skynet.
I'd love to see your implementation of the idea, because i tend to accompany the CC forums and as of yet no one has been able to overcome all those hurdles and get to the point where world eater/skynet is doable. Like i said, there are no shortage of swarm programs, but no world eaters. So, you're saying, either the community isn't capable enough to see how they could solve the problem (by either using other mods as parts of the contraption or by improving on the code), or they aren't interested enough in such a project that it hasn't popped up simply because of lack of interest. I tend to believe it has not surfaced yet because it's hard to implement and code, because of all the problems i outlined... Like i said i'm no real programmer myself so i may well be underqualified to judge the difficulty level of all these problems or see a way around them, but then again, neither has the CC community as far as i know, and there are many many many swarm programs there, but no world eater yet, and that's what i judged my observations on. You are correct, i was very longwinded and could simply have stated "there are many hurdles that CC programmers have not yet surmounted to make such a build possible" and be done with that, but i wished to elaborate a bit more on the complexity of the task in order to better illustrate how difficult it can be to approach the world eater problem. Pathfinding in particular has been a problem tackled by the CC community for ages, and even the best ones still fail now and then, because of the way CC handles some minecraft-related issues, like chunk loading/unloading, resetting computers/turtles and a resume function for when that happens - writing to a text file the current relative positions of all turtles and reading from that at startup tends to work well, but sometimes there's still a fail to register and a turtle gets "lost" eventually... It's a minecraft, not a CC problem when that happens, but still. Hurdles that have yet to be overcome.
Im actually working on a CC project that could begin to start to kinda-sorta make a self replicating world eater horde.
It will dispatch up to 64 turtles (1 stack) to build a pattern. The builders will get resources from a stack of chests, head out to build the pattern. when they need more blocks or fuel, they will return for more. It handles the turtles running into each other when coming and going. its not perfect (yet) -- I still get major traffic jams when they are done building.
Ive done a lot of pondering on how create a setup that will build a base for you -- NOT an 9x9, but a BASE which you can move into and start playing. Something with a crafting table, furnace, chest, a spot for a smeltery, a spot for your MAC, and whatever else i want to add.
@brujon has hit on a bunch of the problems with CC and turtles. but they boil down 2 problems: Fuel Aquisition and Block detection. Fuel can be disabled via config (cheating IMO), but figuring out what kindof block the turtle picked up is a hard problem. Still working on that one.
But back on topic - how much automation is too much? -- When you have automated the game to the point where its no longer fun to play.
It will dispatch up to 64 turtles (1 stack) to build a pattern. The builders will get resources from a stack of chests, head out to build the pattern. when they need more blocks or fuel, they will return for more. It handles the turtles running into each other when coming and going. its not perfect (yet) -- I still get major traffic jams when they are done building.
Ive done a lot of pondering on how create a setup that will build a base for you -- NOT an 9x9, but a BASE which you can move into and start playing. Something with a crafting table, furnace, chest, a spot for a smeltery, a spot for your MAC, and whatever else i want to add.
@brujon has hit on a bunch of the problems with CC and turtles. but they boil down 2 problems: Fuel Aquisition and Block detection. Fuel can be disabled via config (cheating IMO), but figuring out what kindof block the turtle picked up is a hard problem. Still working on that one.
But back on topic - how much automation is too much? -- When you have automated the game to the point where its no longer fun to play.
Skynet would be difficult but turtles that create more turtles to mine more to make more turtles: not so hard. GPS is a virtual non-issue if you don't try to control the coordinates of the turtles you send out to mine. That's what ender chests removes. There are a few ways to solve chunkloading(Turtles that go up to some y level the mining turtles won't go to and place some kind of chunkloader[most preferable would be Chickenchunks set to the line] with it set to a time that it would go out and place and every 16 turtles would get a turtle for that purpose) most optimal would be misc peripherals turtles but meh. Whatever turtle is crafting the mining turtles would keep a count of the number to set for building the new turtles for things like tree farming for charcoal for fuel as one example. It'd be more down to calculating how often new things would need to be made so on and so forth. Ore processing turtles and such could even be made to make it more efficient.
To your question about whether I think it hasn't arisen due to complexity/difficulty or lack of interest; I'd go with the latter. Mostly because a great bulk of the work would simply be the huge amount of coding involved(not difficulty of said code), annoying-ness of troubleshooting problems, and the calculations necessary. It's also a quite pointless endeavor, at least in my opinion.
To your question about whether I think it hasn't arisen due to complexity/difficulty or lack of interest; I'd go with the latter. Mostly because a great bulk of the work would simply be the huge amount of coding involved(not difficulty of said code), annoying-ness of troubleshooting problems, and the calculations necessary. It's also a quite pointless endeavor, at least in my opinion.
You just went full retard. Too much loot? More like not enough loot(and storage space).