Truly infinite worlds (poll)

Would you like to have truly infinite worlds in Minetest?

Yes
30
48%
No
18
29%
That depends/difficult to say
15
24%
 
Total votes : 63

User avatar
voxelproof
Member
 
Posts: 769
Joined: Sat Aug 05, 2017 08:13
Location: Europe

Re: Truly infinite worlds (poll)

by voxelproof » Thu Sep 27, 2018 15:13

I've realized recently that this thread better qualifies for "Minetest-related projects" than "Feature discussion" for I'm pretty sure that this sort of enlargement of MT worlds will never happen in official development course and, well, I'll be happy enough if I see the 5.0 version released at all. Maybe this is due to the autumn blues, but I have a strange feeling that MT is just a shadow of MC and is set to be like that :)

That said, I return to the Eden world. I've found the second super-tall mountain of remarkable beauty, Mount Neverest and because I have a presentiment of the third huge massif existing within this realm, which if found will be named, of course, Mount Minetest, the search will probably take some time. I'll let you know what I have determined in that matter. So long, guys :)
>>>there is life beyond the Cube>>>
 

User avatar
voxelproof
Member
 
Posts: 769
Joined: Sat Aug 05, 2017 08:13
Location: Europe

Re: Truly infinite worlds (poll)

by voxelproof » Wed Oct 03, 2018 13:53

Hi, guys, I'm back alive. Mount Minetest found! 8D
>>>there is life beyond the Cube>>>
 

cy
Member
 
Posts: 59
Joined: Sun Jun 24, 2012 17:25

Re: Truly infinite worlds (poll)

by cy » Thu Dec 27, 2018 04:51

I'd like to point out that with the current way of storing coordinates, if all you stored was the coordinates and 16 bit node type (no entities, no metadata) and you filled an entire map of -32000 to 32000 in all three axes, that would add up to 2,097 terabytes of data.
Code: Select all
(2^16)^3 * (2+2*3)
It takes a long time to sift through 2 petabytes, no matter what algorithm you use. With 32 bit coordinates (-2.1 billion, to 2.1 billion) the largest size of your map file would be around one million yottabytes.

Most worlds fit inside -2000,2000 each way, and even then they're really spread out on the edges. Try running to the end of the map sometime. Heck, try flying noclip to the bottom of the map sometime. 32,000 blocks is a long way.

Only think I've ever considered is sacrificing 1 byte of depth to add a nibble each to the map's width and breadth, like minecraft does. That way maximum depth would be -256, and maximum width/breadth would be 1 million. And that's not so much for the benefit of being able to travel a million blocks west, but because falling for 32,000 blocks is utterly terrifying!
 

User avatar
voxelproof
Member
 
Posts: 769
Joined: Sat Aug 05, 2017 08:13
Location: Europe

Re: Truly infinite worlds (poll)

by voxelproof » Thu Dec 27, 2018 20:26

cy wrote:I'd like to point out that with the current way of storing coordinates, if all you stored was the coordinates and 16 bit node type (no entities, no metadata) and you filled an entire map of -32000 to 32000 in all three axes, that would add up to 2,097 terabytes of data.
Code: Select all
(2^16)^3 * (2+2*3)
It takes a long time to sift through 2 petabytes, no matter what algorithm you use.


As far as data needed for making a map of the world's surface is concerned, this amount would be probably confined to less than 30 GB (I've checked this covering almost whole Eden world - the map.sqlite is 13 GB, made by flying along y meridians every 2000 nodes with viewing range 600).
>>>there is life beyond the Cube>>>
 

Exilyth
Member
 
Posts: 71
Joined: Sun Jul 28, 2013 18:46
Location: Earth

Re: Truly infinite worlds (poll)

by Exilyth » Sat May 11, 2019 10:34

[x] depends/difficult to say

Allowing users to set the map size for new worlds would be nice to have and would benefit all kinds of people: older hardware could run smaller maps and hightech research clusters could use their petaflops to simulate the whole globe.

The option for using a large datatype for larger coordinates (and thus larger worlds) would be nice too.

An option for toroidal maps (e.g. reaching the world edge just wraps to the opposite edge) would be great to. This would give the illusion of a spherical planet without requiring more memory/data. Mapgen would need to be able to work over map edges though to prevent dscontinuities in the noise when jumping from -mapsize to +mapsize.
 

User avatar
Walker
Member
 
Posts: 145
Joined: Tue Oct 03, 2017 09:22
In-game: Walker

Re: Truly infinite worlds (poll)

by Walker » Sat May 11, 2019 12:20

i have create a sector based map saving algorism ... but it gone because of Forum-Crash
 

User avatar
Walker
Member
 
Posts: 145
Joined: Tue Oct 03, 2017 09:22
In-game: Walker

Re: Truly infinite worlds (poll)

by Walker » Sun May 12, 2019 05:14

THANKHS GOD

sorcerykid have recovered my thread ^^

you can see it here > https://sorcerykid.github.io/minetest-f ... _22571.htm

but i will take it too this forum again ^^
 

User avatar
voxelproof
Member
 
Posts: 769
Joined: Sat Aug 05, 2017 08:13
Location: Europe

Re: Truly infinite worlds (poll)

by voxelproof » Wed May 15, 2019 02:21

Exilyth wrote:[x] depends/difficult to say


Yeah, I wholly agree and the larger size of the worlds are desirable only if the mapgen generates sufficiently complex and diverse landscapes. There's no point of pursuing huge worlds when the overall impression of the terrain features remains largely the same over the whole map. However if such interesting mapgens are implemented, it would give the exploration a wholly new dimension and meaning, making it a basis for a new kind of gameplay, so it's imo worth considering if only such techical possibilities exist.
>>>there is life beyond the Cube>>>
 

User avatar
paramat
Developer
 
Posts: 3333
Joined: Sun Oct 28, 2012 00:05
Location: UK
GitHub: paramat
IRC: paramat
 

User avatar
voxelproof
Member
 
Posts: 769
Joined: Sat Aug 05, 2017 08:13
Location: Europe

Re: Truly infinite worlds (poll)

by voxelproof » Sat May 18, 2019 15:38

paramat wrote:'Truly infinite' is also impossible =)


That depends on what do we mean by 'truly'. It is theoretically possible to create a non-procedural terrain generator in which each world would be unique and not related to a given seed, based on pure randomness. Then the only constraint would be the capacity of the hard drive on which the map is stored.
>>>there is life beyond the Cube>>>
 

ShadMOrdre
Member
 
Posts: 248
Joined: Mon Dec 29, 2014 08:07
Location: USA
GitHub: ShadMOrdre
In-game: shadmordre

Re: Truly infinite worlds (poll)

by ShadMOrdre » Mon May 20, 2019 16:07

In this case, IMHO, "truly infinite" does not mean something that a computer can generate. That, I would agree, is not possible.

However, "truly infinite" can simply mean "without limit". In this sense, a computer does not need to generate infinity data, it simply needs not limit the amount of data that CAN be generated. Rather than rely on a 16 bit value, that essentially limits everything, the code could use a much larger larger value, and either use a portion of the larger type to downsize to the 16bit value, or better yet, use multiple 16bit values that can be stacked in a meaningful way.

Walker, in the other thread, posts code examples of how something could possibly work.

The Opensimulator project, referenced in my sig, did something very similar. In that software, the same limits apply. The work around was very similar to what Walker has suggested, and in the Opensim project, it was realized.
MY MODS: lib_ecology lib_materials lib_clouds lib_node_shapes ---- Inspired By: Open Source Virtual World Simulator Opensimulator.
 

Previous

Return to Feature Discussion



Who is online

Users browsing this forum: No registered users and 2 guests