Infinite world size

[rubenwardy]

A game with an infinite map would be one where the hard drive disk size is the only limiting factor. If you had an infinite hard drive, there are no technical reasons why you cannot walk forever.

Minetest and Minecraft are both limited by the way they store chunks and blocks in the ram, in limiting integers and floats.

[Sokomine]

We could do something like "Congratulations! You've reached the end of the game! Click here to restart" - and teleport the player back to spawn. It would be silly, but so is trying to reach the end (of the map) for no other reason than complaining about there beeing one :-)

[thetoon]

We could do something like "Congratulations! You've reached the end of the game! Click here to restart" - and teleport the player back to spawn. It would be silly, but so is trying to reach the end (of the map) for no other reason than complaining about there beeing one :-)

I won't go all the way again, but there's clearly more to "we want a bigger map" than "we have nothing better to do than walking to the end of a video game world". You might want to map part of real-world into Minetest (and you can't even map a small country like Belgium!) . You might want to host a server with BOTH plenty of players AND a sense of loneliness (and you might even be willing to pay for the hardware it would demand). You might just want to explore the same world for years without losing the feeling "there's a lot more than what I've already saw".

I do understand it's not THAT easy to increases MT world size (heck! I even tried to look into it and got lost, so I'm perfectly aware it's no matter of changing a constant's value). I do also understand that no core developer is willing to take time for this. But I'm starting to get a _little_ annoyed by the usual "you want a bigger? no you don't" approach. Minecraft's sale has brought a bunch of new guys here (and we only see the registered ones), and that's a good thing.

Some of them probably have interesting ambitions for MT, some of them might even know how to code. We shouldn't discourage them from looking into the world size "issue" (pardon the quotes, but just because it's not an issue to all of you doesn't mean it's not an issue at all).

Then some others of those newcomers are just MC players who look for an alternative. I know some of them will compare MT and MC, even though we keep on stating there wasn't any reason to. Some of them will, and probably have already, compare world size. It's one thing to read that horizontal space - the one that matters to the explorers of us - is much smaller in MT. It's a whole other thing to read that it's not even considered as a problem or as an eventual improvement.

Now, for the technical parts. Of course an infinite - and even a finite-but-big-enough - world will bring a whole lot of problems for server admins. The storage needs for the complete world will grow cubic - no pun intended. But there's already means of limiting this (esp. by deleting the un-modified or the oldest parts). And there's still hope of a storage engine with some delta capabilities. And then again, if one guy is willing to put what it takes in his hardware to host the biggest world ever, why _should_ software go in his way? We'd clearly want to warn him, but not prevent him.

One last thing : just because we allow really big worlds in the engine doesn't mean it's a default setting. We could perfectly well ship MT with a world size similar to what it already has, but let those who want more be a setting away from it.

[Enke]

We could do something like "Congratulations! You've reached the end of the game! Click here to restart" - and teleport the player back to spawn. It would be silly, but so is trying to reach the end (of the map) for no other reason than complaining about there beeing one :-)

"Hey, you reached the end! Nice job! Now howzabout you go back to spawn and do it in the other direction?" Would be a rather funny feature to implement

[Sokomine]

I won't go all the way again, but there's clearly more to "we want a bigger map" than "we have nothing better to do than walking to the end of a video game world".

Sure there is. But it's not a "we don't have enough space to build what we want to" issue. It's psychological. There's a border, something...almost..forbidden (you can't walk past) - and that's a challenge to human nature. People have to check if the border is really there, how it looks like, weather they can get past it, if there are any nice bugs to explore, whatever. Without there beeing such a border, the place would be utterly unintresting. I love to explore such "borders" in other types of games sometimes - see how well the simulation works, what happens if the player does something he/she's not supposed to do etc. - which can be more entertaining in some games than the main game course. Or disappointing, if it becomes too soon too obvious that there's no way to affect the world other than that what the story demands.

You might want to map part of real-world into Minetest (and you can't even map a small country like Belgium!) .

That's right. On the other hand, you wouldn't notice it anyway. View distance is usually so short that you might be standing at the foot of Mount Meru and be unable to distinguish it from an almost world-spanning ocean.

You might want to host a server with BOTH plenty of players AND a sense of loneliness (and you might even be willing to pay for the hardware it would demand). You might just want to explore the same world for years without losing the feeling "there's a lot more than what I've already saw".

Perhaps it's the "sense for loneliness" that ought to be stressed here. That's definitely present in some players, and it might well be justified. It's even possible to reach such spots, even on a server that has been up for years. Without advanced travel methods, getting to and from such places takes time, and neither inhabitant nor potential visitors might be too eger to spend that time more than once. And we get people crowding up on spawn and complaining if they can't find a spot 50 m from spawn anymore.

The storage needs for the complete world will grow cubic - no pun intended. But there's already means of limiting this (esp. by deleting the un-modified or the oldest parts). And there's still hope of a storage engine with some delta capabilities.

That's right, a *full* world would be an awful lot of data. But that's not what you'd get if the limits could be dropped. If it's a server, it has to send out data to the players which are exploring the terrain - which puts a limit to the amount of data that can be accumulated. Also, most of the time players will stick to one place. Thus, unless you get bots emulating the explore_map mod, it's no practical limitation. Even after years, it'll most likely still be far less than if the entire surface area of what is currently possible as map size would have been explored. Just diffrently. You *will* run into trouble if you want to give your players an overview image of where their houses are - because you'll have a huge, empty white map with some random paths accross it where people went exploring.

[twoelk]

Minetest is an infinite-world block sandbox game and a game engine, ...

by now I really think using the word "infinite" in the description on the Minetest-Homepage in this way leads to wrong ideas about the possibilities and utter disapointment when theory clashes with the at the moment practically possible reality, be it code or accessable hardware.

I do agree though that there should be no limit to think about ways that infinity could be achieved even if the sollution lies somewhere in the infinity of time.

[digitalmouse]

I like the idea of having huge maps but if someone asks for your coordinates.. ehm. Okay.

Then you just say "walk due east from the spawn point for 2 hours (real time), then turn left at the big tree, and follow the river north for 3 hours until you reach the cliffs...which might take you 20 minutes to climb." :D

Or we start defining political boundaries: ie, "the Nation of Quirm", which are subdivided into Counties or Kingdoms, and *then* provide directions from there. :)

[Enke]

I like the idea of having huge maps but if someone asks for your coordinates.. ehm. Okay.

Then you just say "walk due east from the spawn point for 2 hours (real time), then turn left at the big tree, and follow the river north for 3 hours until you reach the cliffs...which might take you 20 minutes to climb." :D

Or we start defining political boundaries: ie, "the Nation of Quirm", which are subdivided into Counties or Kingdoms, and *then* provide directions from there. :)

And then we build a subgame off of that concept. Factional warfare begins. The forum is split into many divided groups. Many great wars are fought, taking years to complete. Eventually, one kingdom rises above the rest, declares themselves Lord of Minetest, and proceeds to take over the forum. Then he finds that it's just him, and some kid. He crowns the kid as Prince, and rules his kingdom of two with an iron fist. Forever. And then the server dies.

[TBC_x]

The major problem of "Tiny world" lies IMHO in lack of vertical content. I don't know the diversity of mods adding vertical content, but I assume there is little to none. Making a modding competition of some sort could bring new places to explore. I would love to see Dwarf fortress-like caverns in MT.

[philipbenr]

Agreed. I am having to blast my own. It is rather time consuming...

[Kilarin]

don't know the diversity of mods adding vertical content, but I assume there is little to none.

Check out:
SkyTest <my skytest journal>
CaveRealms
and WebWorld
Just to name a few.

[Sokomine]

Then you just say "walk due east from the spawn point for 2 hours (real time), then turn left at the big tree, and follow the river north for 3 hours until you reach the cliffs...which might take you 20 minutes to climb." :D

...and then you arrive and notice that the building advertised as "great castle" is nothing more than a hollow block of cobble, complete, as those things come, with a locked steel door, furnace and unlocked chest.

[twoelk]

...and then you arrive and notice that the building advertised as "great castle" is nothing more than a hollow block of cobble, complete, as those things come, with a locked steel door, furnace and unlocked chest.

hehe - I have an old mc map with a big castle like that.
From the outside it reminds of a large english keep but if you go inside, instead af a nice hall all you find is a huge hole.

A deep hole, or rather a mineshaft down to the bedrock aligned on all sides by a gallery not unlike the mine under the red glass tower on VE's-vanilla server. - Ok there is a little more to it - designed a little for the surprising optical effect. Of course there is a little tree on one side with vines to the bottom of the pit and a nice waterfall a little further down. several caverns carved to the sides on different levels, two minecart stations with connecting lines to other mine systems, not to forget the natural cave systems it opens up to now and then and - uhm maybe I should revisit that old map.

Oh well, still dreaming of a nice and easy one click converter to bring my first voxel game map into minetest. Though with some 4000*4000 nodes it might be quite a chunk to chew for any script be it lua, python or whatever.

[Don]

It would be nice if you could easily set the size you want. Some servers like hunger games or PvP servers could be much smaller while survival servers could have less height and more vertical. Also I wish that when a player reaches the edge the get teleported to the other side. That way if you walk in a straight line it would be like you walked around the world.

[MirceaKitsune]

Infinite is probably hard to do. But Minetest's world size is way way too small in my opinion! I really hope it can be increased by a great amount eventually.

[Krock]

Might be interesting:
Split block position into separate fields in SQLite3 database

[twoelk]

Infinite is probably hard to do. ...

all you really nead is an infinite amount of memory and infinite time to do an infinite amount of calculations to render an infinite amount of nodes. Everything else should be easy once you got the infinite resources needed to solve those infinite problems.

Of course you may have entered infinity yourself long before any infinite probleme has been solved as infinity is a lot larger than most people think.

[drkwv]

all you really nead is an infinite amount of memory and infinite time to do an infinite amount of calculations to render an infinite amount of nodes. Everything else should be easy once you got the infinite resources needed to solve those infinite problems.

This infinite-cap can be workarounded by deleting chunks that contain no significant player-made changes and are not visited by player for, say, a century. Thus you don't need infinite memory to store these chunks as they can be precisely recreated using seed.

[Don]

Infinite is probably hard to do. ...

all you really nead is an infinite amount of memory and infinite time to do an infinite amount of calculations to render an infinite amount of nodes. Everything else should be easy once you got the infinite resources needed to solve those infinite problems.

Of course you may have entered infinity yourself long before any infinite probleme has been solved as infinity is a lot larger than most people think.

Maybe you will get your infinite resources when someone make a really really good potions mod

[digitalmouse]

Larger worlds wouldn't help people who are too lazy to search ...

...or walk/fly/ride/teleport to a new area. The majority of people playing are kids with no developed sense of exploration.

[Kilarin]

But Minetest's world size is way way too small in my opinion!

Ah, but to understand the significance of your statement, we need to understand what actual experience your opinion is based upon.

What percentage of a MineTest world have you ever explored?

[thetoon]

But Minetest's world size is way way too small in my opinion!

Ah, but to understand the significance of your statement, we need to understand what actual experience your opinion is based upon.

Say "Minetest's very own homepage"? No-one would ever consider a 64km-sided cube "Almost infinite world" (emphasis is from that page). It's not a world, not even a country. It's a pretty cool sandbox, there's plenty of reasons not to enl4rg3 it, and there are many players who don't care about exploring that far. And that's fine.

No, it's not being a capricious child to state the world's too small. It would be if one demanded the world to be bigger "right now", and/or without effort on his own part. But we need to get rid of the "are you _really_ sure the world's too small?" defense. MirceaKitsune was only stating his own opinion, which is perfectly valid (and stands easily against MT's own description), not saying MT is crap because of it's world size, even less insulting developers for having made said world too small.

Yes, the world is a bit small for some uses (and no, just because these are not yours doesn't mean they're any less valuable or valid). I've already written about them (Denmark, Mars, motorized vehicles...) and won't go into detail again. I would understand (and Mircea as well, I'm sure) there's no one willing to work on that (at least for now) and as I can't spend as much time as I would on MT I wouldn't mind world-size being lowest priority. But thinking about how we could enlarge the world won't hurt anyone, will it?

Don't try to show people how they're wrong and how the world is perfectly big enough : it is for you (and plenty of others), it isn't for some others. That's about it.

[Kilarin]

Don't try to show people how they're wrong and how the world is perfectly big enough : it is for you (and plenty of others), it isn't for some others. That's about it.

Actually, you will note that I didn't deny his claims. I simply asked what they were based on. We need that information to understand what the actual complaint is. If the complaint is "I've tried doing something that wouldn't fit", then that is very valuable information for the developers. If the complaint is, "It just bugs me knowing that there is an edge, even though I've never gotten anywhere near it" then, while this is STILL a valid point of view, it's a very different problem.

There is discussion of separating the coords in the database. I suspect that might make expanding the world size a little bit easier.

[thetoon]

Actually, you will note that I didn't deny his claims. I simply asked what they were based on. We need that information to understand what the actual complaint is.

Which word don't you understand? World, or size? His complaint sounds like mine, and is pretty easy to understand : the world could be bigger. I know I'm calling for bad-tempered comments by saying that, but the mere fact our main competitor (Minecraft, for the few of us who already forgot and/or like to live with their head in sand) has a much bigger map is actually a reason to consider the claim legitimate in the first place.

For the sake of clarity, here's what I (and, pretty much everyone in computer science) would consider "infinite" : only limited by computing resources (memories and computing power). And this is not even something I do expect for MT, at least not in the next decade.

I do know computing is about trade-offs, and "we won't have a bigger map because we have a lot of vertical space" is a valid one (that I'm not too happy with, but let's stick to the point). "We don't have a bigger map because we haven't seen any use for it" isn't, especially when you demand any "use" to be thoroughly proven/documented.

If the complaint is "I've tried doing something that wouldn't fit", then that is very valuable information for the developers. If the complaint is, "It just bugs me knowing that there is an edge, even though I've never gotten anywhere near it" then, while this is STILL a valid point of view, it's a very different problem.

That's exactly what I'm standing against right now. "Tell me what you need, we'll explain you how to live without", as we use to say in France about computer scientists. And I do take my share of the blame here : I tend to act like that myself more than I should.

As far as I read these forums, I've never seen anyone asking (even slightly) for an infinite world. Even a world without edges is easier that an infinite one ("donut" is the way). Asking for an infinitive world is "a very different problem", and not something we should tackle too early (nor too sober). Asking for a world the size of the Belgium, in the meantime, could very well be.

There is discussion of separating the coords in the database. I suspect that might make expanding the world size a little bit easier.

I've read about that, and it's rather promising.

For my part, I'll stop ranting about world-size for a while : patches and pull requests are much more efficient and I don't have time for them (now). What I'd love to is, in the meantime, not reading the same stance about "being sure we need a bigger world" again and again. Some of us are. It might be their (our) responsibility to make it happen, that's fine : this is what open-source is about, scratching our own itches.

If there's a real consensus shared among the developpers about "not making the world bigger", then state it on the homepage, and try to explain it. Many FOSS games have strict limitations, and stand firm about them : Oolite - and Elite clones in general - aren't multiplayer and won't ever be. People understand that. We (as a community) would need to go that way, IMHO.

[ArguablySane]

If the complaint is "I've tried doing something that wouldn't fit", then that is very valuable information for the developers. If the complaint is, "It just bugs me knowing that there is an edge, even though I've never gotten anywhere near it" then, while this is STILL a valid point of view, it's a very different problem.

I'll throw in my two cents on this. I know some people don't like comparisons to minecraft, but I think it does act as a good guide for the things players and modders might want to do in an "infinite" voxel-based world.

In minecraft there are a few examples of projects which have required a linear distance greater than 64,000 blocks.
The best examples are automated mining machines such as tunnel bores. These continuously dig out a tunnel or trench, feeding the extracted ores back to a base station to be processed. Assuming continuous operation on a server and a mining rate of 1 metre every 2 seconds (typical for tunnel bores), it would take roughly 18 hours to bump into the edge of the map if started from the middle. For more complex machines with a larger extraction area you could expect the movement speed to be several times slower, but I still wouldn't expect one to run for more than a week before crashing into the edge of the map.
Another example is the "Far Lands or Bust" youtube series. The last reported distance he'd walked was 1,479,940 blocks in the z direction. This is an extreme example, but obviously it wouldn't be nearly as impressive if he had run into an invisible wall after walking 32000 blocks.