Infinite world size

User avatar
Linuxdirk
Member
Posts: 3217
Joined: Wed Sep 17, 2014 11:21
In-game: Linuxdirk
Location: Germany
Contact:

Re: Infinite world size

by Linuxdirk » Post

MirceaKitsune wrote:
Wed Apr 13, 2022 23:50
I can say I've been thinking of making my own Minecraft / Minetest like engine in Godot, […]
Don’t reinvent the wheel. There are multiple stable Voxel engines for Godot.

Voxel Tools for example: https://www.youtube.com/watch?v=WxZK_Yg5kU0

User avatar
Blockhead
Member
Posts: 1622
Joined: Wed Jul 17, 2019 10:14
GitHub: Montandalar
IRC: Blockhead256
In-game: Blockhead Blockhead256
Location: Land Down Under
Contact:

Re: Infinite world size

by Blockhead » Post

Linuxdirk wrote:
Thu May 05, 2022 17:54
Don’t reinvent the wheel. There are multiple stable Voxel engines for Godot.
As much as the maxim is useful sometimes, I like to add an exception: 'Don't reinvent the wheel, unless you're competent and dedicated to building a better wheel than we already have'. We still need new better wheels all the time - cars used to have spoked wheels similiar to a bicycle or motorcycle!
/˳˳_˳˳]_[˳˳_˳˳]_[˳˳_˳˳\ Advtrains enthusiast | My map: Noah's Railyard | My Content on ContentDB ✝️♂

Astrobe
Member
Posts: 570
Joined: Sun Apr 01, 2018 10:46

Re: Infinite world size

by Astrobe » Post

Blockhead wrote:
Fri May 06, 2022 06:06
Linuxdirk wrote:
Thu May 05, 2022 17:54
Don’t reinvent the wheel. There are multiple stable Voxel engines for Godot.
As much as the maxim is useful sometimes, I like to add an exception: 'Don't reinvent the wheel, unless you're competent and dedicated to building a better wheel than we already have'. We still need new better wheels all the time - cars used to have spoked wheels similiar to a bicycle or motorcycle!
Same, except I hate this maxim because it is too often used to justify (or rather disguise) risk aversion (obviously not the case here).

The wheel has been invented circa 3500 years ago from the potter's wheel, and has evolved during all this time. Its most recent "re-inventions" are the mecanum wheel (1972) and the Sbarro wheel (1989).

0siribix
Member
Posts: 123
Joined: Tue Nov 17, 2020 20:54
GitHub: 0siribix
In-game: 0siribix

Re: Infinite world size

by 0siribix » Post

I'm beginning a project with Ronoaldo [Mercurio] to build a server/client in Godot. We have already begun to look at Voxel Tools but haven't decided whether we're going to build from it or use it as a reference. We do not currently plan to try to support communication with MT server or client. We do intend to support using current MT scripts and assets. We will a license that is compatible with Apple so that we can have an official client on iOS

User avatar
Blockhead
Member
Posts: 1622
Joined: Wed Jul 17, 2019 10:14
GitHub: Montandalar
IRC: Blockhead256
In-game: Blockhead Blockhead256
Location: Land Down Under
Contact:

Re: Infinite world size

by Blockhead » Post

0siribix wrote:
Mon May 09, 2022 02:06
I'm beginning a project with Ronoaldo [Mercurio] to build a server/client in Godot. We have already begun to look at Voxel Tools but haven't decided whether we're going to build from it or use it as a reference. We do not currently plan to try to support communication with MT server or client. We do intend to support using current MT scripts and assets. We will a license that is compatible with Apple so that we can have an official client on iOS
You will probably want to start your own thread about this - but before you go.. Can I hazard a guess this project will have infinite worlds? What licence will it be under, and what do you think you will do if anybody violates it?
/˳˳_˳˳]_[˳˳_˳˳]_[˳˳_˳˳\ Advtrains enthusiast | My map: Noah's Railyard | My Content on ContentDB ✝️♂

0siribix
Member
Posts: 123
Joined: Tue Nov 17, 2020 20:54
GitHub: 0siribix
In-game: 0siribix

Re: Infinite world size

by 0siribix » Post

Blockhead wrote:
Mon May 09, 2022 05:42
You will probably want to start your own thread about this - but before you go.. Can I hazard a guess this project will have infinite worlds? What licence will it be under, and what do you think you will do if anybody violates it?
We are still in planning phase but design goals include 32-bit coords (meaning the same thing proller has been doing with MT) as well as 4th dimension (multiple worlds).

Currently the license we are considering is Apache but we are not necessarily settled on that. What's most important to me is that someone is not able to easily bundle my work with some other junk and put a price tag on it so as to make easy money from something that I intend to be used for free.

Third question is too broad. It really depends on circumstances. Violating the license matters little to me. What matters is how my work is being (ab)used.

ronoaldo
Member
Posts: 177
Joined: Mon Dec 07, 2020 01:04
GitHub: ronoaldo
IRC: ronoaldo
In-game: RonoaldoKakashi
Location: São Paulo, Brasil
Contact:

Re: Infinite world size

by ronoaldo » Post

On the topic: I guess that some other core devs at the Minetest project may be working/thinking about the implementation of such feature, specially in support to alternative dimensions.

In tech, infinite is quite vague, most of the time it is not actually infinite. I usually think of it as "large enough to the upper/lower bounds work well for the majority of use cases". Like, such implementations usually require a space/compute trade-off and a storage serialization like protobuff for the map can eventually help with the potentially larger coords in such a way that one could stick with the fixed limites of now without increasing space unnecessarily.

About the project 0siribix mentioned we are working on, I'm a step back myself now: learning OpenGL while doing some coding experiments before I dig into Godot side. About licensing, it will probably end up being a dual license, as we don't want commercial usage. Doing something if infringement happens is a matter of how much effort we want to put on it but I personally don't plan to sue anyone; it does empower us to fill DMCA take downs tho.
Servers: Mercurio | Tools: ContentDB CLI | Mods: minenews

ronoaldo
Member
Posts: 177
Joined: Mon Dec 07, 2020 01:04
GitHub: ronoaldo
IRC: ronoaldo
In-game: RonoaldoKakashi
Location: São Paulo, Brasil
Contact:

Re: Infinite world size

by ronoaldo » Post

Astrobe wrote:
Fri May 06, 2022 13:34
Blockhead wrote:
Fri May 06, 2022 06:06
Linuxdirk wrote:
Thu May 05, 2022 17:54
Don’t reinvent the wheel. There are multiple stable Voxel engines for Godot.
As much as the maxim is useful sometimes, I like to add an exception: 'Don't reinvent the wheel, unless you're competent and dedicated to building a better wheel than we already have'. We still need new better wheels all the time - cars used to have spoked wheels similiar to a bicycle or motorcycle!
Same, except I hate this maxim because it is too often used to justify (or rather disguise) risk aversion (obviously not the case here).

The wheel has been invented circa 3500 years ago from the potter's wheel, and has evolved during all this time. Its most recent "re-inventions" are the mecanum wheel (1972) and the Sbarro wheel (1989).
If you're reinventing a wheel for researching purposes, I would always say go for it. You should't kill your project before it is born. Your wheel don't need to be better; you can improve production aspects of it for instance and the resulting product be still the same. At least, you will build knowledge yourself in the process and that is already priceless value being generated. I avoid reinventing wheels most of the time (specially because I'm not in the automobiles industry lol) but sometimes I do just for the process itself, not for the result.

Knowledge is power.

EDIT: fixed typo (s/well/wheel/g)
Servers: Mercurio | Tools: ContentDB CLI | Mods: minenews

DOOM_possum
Member
Posts: 172
Joined: Sat Mar 27, 2021 22:06
In-game: DOOM_possum

Re: Infinite world size

by DOOM_possum » Post

well just like any Game, ((( that runs into These Problems )))

we can only hope someone comes up with something hot, and smart to give us more Freedom
i'm having similar problems in Ternaria on the STEAM platform, but infinite world just makes no sense (There)

DOOM_possum
Member
Posts: 172
Joined: Sat Mar 27, 2021 22:06
In-game: DOOM_possum

Re: Infinite world size

by DOOM_possum » Post

any word yet on a good MOD implementing THIS into the Game yet (?) (?)

User avatar
debiankaios
Member
Posts: 910
Joined: Thu Dec 03, 2020 12:48
IRC: debiankaios
In-game: debiankaios Nowe
Location: germany
Contact:

Re: Infinite world size

by debiankaios » Post

You can compile minetest32 but with mods that's not possible :(. But we all hope that it will implemented in future

User avatar
j0j0n4th4n
Member
Posts: 249
Joined: Tue Jan 26, 2021 06:45

Re: Infinite world size

by j0j0n4th4n » Post

What about worlds with different shapes? Right now Minetest is a giant cube with a side of of 60 000 km, would it be possible to say, lower the depth and increase the width of the world?
cdb_894a100ddd76

User avatar
Blockhead
Member
Posts: 1622
Joined: Wed Jul 17, 2019 10:14
GitHub: Montandalar
IRC: Blockhead256
In-game: Blockhead Blockhead256
Location: Land Down Under
Contact:

Re: Infinite world size

by Blockhead » Post

j0j0n4th4n wrote:
Sun Jul 24, 2022 21:14
What about worlds with different shapes? Right now Minetest is a giant cube with a side of of 60 000 km, would it be possible to say, lower the depth and increase the width of the world?
*60 km, or 61 824 m to be precise. If it were 60 000 km everyone might already be happy with it..

That would still require major changes to the coordinate system, for not much benefit. In particular it would get tricky to let you choose which dimensions to sacrifice. Of course, you are always free to limit the mapgen in a particular direction to be less, but that's not the point.
/˳˳_˳˳]_[˳˳_˳˳]_[˳˳_˳˳\ Advtrains enthusiast | My map: Noah's Railyard | My Content on ContentDB ✝️♂

c56
Member
Posts: 307
Joined: Wed Apr 21, 2021 03:05
GitHub: tigercoding56
In-game: bm5 or bemo5 also sell_her_on55

Re: Infinite world size

by c56 » Post

debiankaios wrote:
Wed Jul 13, 2022 16:59
You can compile minetest32 but with mods that's not possible :(. But we all hope that it will implemented in future
well it is not possible unless you want a 0 fps game (a mod implementing an infinity world would need to change the blocks around the player (entities if you want smooth walking ) and keep the player at 0 0 0 and store generated blocks somewhere and generate the map itself so basically doing everything minetest is doing but using minetest as display )
this is a signature not a place to post messages also if i could change my username i would change it to sell_her_on55

Astrobe
Member
Posts: 570
Joined: Sun Apr 01, 2018 10:46

Re: Infinite world size

by Astrobe » Post

Blockhead wrote:
Mon Jul 25, 2022 00:35
j0j0n4th4n wrote:
Sun Jul 24, 2022 21:14
What about worlds with different shapes? Right now Minetest is a giant cube with a side of of 60 000 km, would it be possible to say, lower the depth and increase the width of the world?
*60 km, or 61 824 m to be precise. If it were 60 000 km everyone might already be happy with it..

That would still require major changes to the coordinate system, for not much benefit. In particular it would get tricky to let you choose which dimensions to sacrifice. Of course, you are always free to limit the mapgen in a particular direction to be less, but that's not the point.
Nah... "Nobody needs" 32Km above and 32Km below. There's no option to offer outside of "how many".

I guess most would be alright with only 8Km (16Km total), that's two bits you can reallocate to double the "length" of the two other axis.

Some would probably be ok with only 2Km (for another 4x of the surface), but it would be less comfortable if you want to squeeze floatlands and/or Nether.

Going down to 0.5Km up/down (that's still huge without drilling or flying machines) would in practice forbid "layering" mods, and one would then sacrifice diversity over quantity; you'd have plenty of horizontal space.. Filled with more of the same mapgen and decorations. I think that would only satisfy servers with a large (past or present) populations.

Speaking of them, I believe that popular servers with infinite world size would collapse under their own weight - computer resources are never infinite - so that they would eventually disable it in a way or another.

User avatar
Blockhead
Member
Posts: 1622
Joined: Wed Jul 17, 2019 10:14
GitHub: Montandalar
IRC: Blockhead256
In-game: Blockhead Blockhead256
Location: Land Down Under
Contact:

Re: Infinite world size

by Blockhead » Post

Astrobe wrote:
Tue Aug 02, 2022 17:47
Nah... "Nobody needs" 32Km above and 32Km below. There's no option to offer outside of "how many".
The SI "metric system" unit symbol for kilometre is km. You should also leave a space between the unit value and the symbol.
m = metre
K = Kelvin
k = thousands prefix

32,000 metres = 32 kilometres = 32 km
Another example to help you understand SI
Watts / (metres * Kelvin) -> W/m·K - notice the dot, which stands for multiplication but also helps to separate units where confusion may occur due to m being a unit prefix and a unit.

You should not let the fact that k is not uppercase despite being more than the base unit disturb you. Just think, every prefix above k is uppercase. Also the kilogram is the base unit of mass. The SI has strict rules, and following them greatly aids communication in science, trade and international communications. Breaking the rules brings us back to the dark age of a plethora of non-standard units and symbols. These are the rules, not your rules.
Astrobe wrote:
Tue Aug 02, 2022 17:47
I guess most would be alright with only 8Km (16Km total), that's two bits you can reallocate to double the "length" of the two other axis.

...
You have not understood the main crux of the argument: The complexity it takes to store coordinates in Minetest's world database in the first place. The coordinates are packed together into a 64-bit integer. Each coordinate is 16 bits; the 64-bit container limits the total size to about 20 bits for each coordinate.

The fact that all three coordinates are constant size is an important point of simplicity. The functions that unpack this 64-bit structure into 3 16-bit values are:

Code: Select all

  def getBlockAsInteger(p):
      return int64(p[2]*16777216 + p[1]*4096 + p[0])

  def int64(u):
      while u >= 2**63:
          u -= 2**64
      while u <= -2**63:
          u += 2**64
      return u
      
  def getIntegerAsBlock(i):
      x = unsignedToSigned(i % 4096, 2048)
      i = int((i - x) / 4096)
      y = unsignedToSigned(i % 4096, 2048)
      i = int((i - y) / 4096)
      z = unsignedToSigned(i % 4096, 2048)
      return x,y,z

  def unsignedToSigned(i, max_positive):
      if i < max_positive:
          return i
      else:
          return i - 2*max_positive

What you are suggesting is that we take those constants 16777216 & 4096 and make them variables. This would entail recording how many bits are assigned to each coordinate for all coordinates in the local database, and keeping that in mind when unpacking the coordinates.

Code: Select all

  def getIntegerAsBlock(i):
      x = unsignedToSigned(i % 2**(x_bits), 2**(x_bits-1))
      i = int((i - x) / 4096)
      y = unsignedToSigned(i % 2**(y_bits), 2**(y_bits-1))
      i = int((i - y) / 4096)
      z = unsignedToSigned(i % 2**(z_bits), 2**(z_bits-1))
      return x,y,z
Intuition tells me this is going to have worse performance, even if we use a closure. Furthermore, we have not gained spaced, only rearranged it. We have also broken compatibility with any program outside minetest that reads its mapblock format, requring it to reimplement the more complex algorithm. I therefore put it to you that if we must introduce such a breaking change, we should at least not do it a way that will be compromised: we should increase the world size drastically if we are going to change the mapblock format at all.
Speaking of them, I believe that popular servers with infinite world size would collapse under their own weight - computer resources are never infinite - so that they would eventually disable it in a way or another.
Storage space is not too terribly expensive in the scheme of things, it's just a lot of servers are running in cheap VPS with only a small allocated storage amount in the plan. We already have a setting to disable generating the world beyond a certain area; that is how people do reduce the available area already. However, that doesn't mean that if the world's potential size got larger they'd decide to still limit it to less than 60,912. They might well decide to limit it to 100,000 or 250,000.
/˳˳_˳˳]_[˳˳_˳˳]_[˳˳_˳˳\ Advtrains enthusiast | My map: Noah's Railyard | My Content on ContentDB ✝️♂

DOOM_possum
Member
Posts: 172
Joined: Sat Mar 27, 2021 22:06
In-game: DOOM_possum

Re: Infinite world size

by DOOM_possum » Post

yes, it isn't a matter of how DEEP you can go, as most worlds end up with a bed of molten lava anyway at it's core, just how much space you can actually use to develop

(which is The Most Important Thing)

and sure, we'd all like to get to space some day too, but i agree, that is very taxing on low end PC's

Astrobe
Member
Posts: 570
Joined: Sun Apr 01, 2018 10:46

Re: Infinite world size

by Astrobe » Post

Blockhead wrote:
Tue Aug 02, 2022 18:48
Astrobe wrote:
Tue Aug 02, 2022 17:47
Nah... "Nobody needs" 32Km above and 32Km below. There's no option to offer outside of "how many".
The SI "metric system" unit symbol for kilometre is km. You should also leave a space between the unit value and the symbol.
Yeah, right, thank you for the lecture, master. A quick remark about how I misused SI notations with a link to wikipedia just in case would have been enough, though.
What you are suggesting is that we take those constants 16777216 & 4096 and make them variables. This would entail recording how many bits are assigned to each coordinate for all coordinates in the local database, and keeping that in mind when unpacking the coordinates.
Certainly I was not clear, but I actually was not considering a configurable packed coordinates format. I learned that there are 16bits left unused, enough to store a 4th dimension, that's probably why people talk about "dimensions" here and there.

BTW I think your constants are wrong, they should be ~4 billions (2^32) and 65536 (2^16) - unless your p argument doesn't carry the usual world coordinates or QBasic2000 has weird semantics - that's the constants used by hash_node_position(). In C, stealing a bit or two from the Z coordinate is a matter of masks, shifts, and ORs, no weird loops and modulus required, so it wouldn't incur a big performance hit.
Intuition tells me this is going to have worse performance, even if we use a closure.
"Memoization" to be accurate, I guess.
Furthermore, we have not gained spaced, only rearranged it.
That's the idea of OP, I think. Make the basement and the attic more useful by moving them to the left and back side of the house.
We have also broken compatibility with any program outside minetest that reads its mapblock format, requring it to reimplement the more complex algorithm.
I don't think this format is "contractual", i.e. I don't think guaranties are given anywhere that it will stay the same. When you base your work on undocumented implementation details, your work will sooner or later endure the same fate as an MC mod (and every new version will sound like bad news for you).

User avatar
Blockhead
Member
Posts: 1622
Joined: Wed Jul 17, 2019 10:14
GitHub: Montandalar
IRC: Blockhead256
In-game: Blockhead Blockhead256
Location: Land Down Under
Contact:

Re: Infinite world size

by Blockhead » Post

Astrobe wrote:
Wed Aug 03, 2022 17:49
Yeah, right, thank you for the lecture, master. A quick remark about how I misused SI notations with a link to wikipedia just in case would have been enough, though.
Sorry, you just hit a pet peeve of mine :)
Astrobe wrote:
Wed Aug 03, 2022 17:49
Certainly I was not clear, but I actually was not considering a configurable packed coordinates format. I learned that there are 16bits left unused, enough to store a 4th dimension, that's probably why people talk about "dimensions" here and there
Oh. Well, well, because you only talked about sizes smaller than 60912, not larger, so I assumed to meant letting people squeeze the existing bits downward as well as bumping others up. As for expanding into the 16 bits available non-uniformly, in just one opinionated way? Hmm it's not as silly as configurable format, but I'd probably throw it in the same basket as I did the other idea, and say that if we're going to change the world format we had better do it right, and so 64 bits is not enough.
Astrobe wrote:
Wed Aug 03, 2022 17:49
BTW I think your constants are wrong, they should be ~4 billions (2^32) and 65536 (2^16) - unless your p argument doesn't carry the usual world coordinates or QBasic2000 has weird semantics - that's the constants used by hash_node_position(). In C, stealing a bit or two from the Z coordinate is a matter of masks, shifts, and ORs, no weird loops and modulus required, so it wouldn't incur a big performance hit.
The python functions are copied from minetest/doc/world_format.txt and yes I probably got them wrong, but it's all a sketch anyway. But I did assume we were sticking with 16 bits per coordinate in that scenario, so don't mark me down if you though I was talking about 32 bits per coordinate :P
Astrobe wrote:
Wed Aug 03, 2022 17:49
Intuition tells me this is going to have worse performance, even if we use a closure.
"Memoization" to be accurate, I guess.
I did mean closures as in Python or Lua closures, but of course it's C++ so we don't really have such a thing. Yes, memoization in this case is a useful technique, but yes, it's still slower since it's not built into the instructions like compile-time constants are.
Astrobe wrote:
Wed Aug 03, 2022 17:49
We have also broken compatibility with any program outside minetest that reads its mapblock format, requring it to reimplement the more complex algorithm.
I don't think this format is "contractual", i.e. I don't think guaranties are given anywhere that it will stay the same. When you base your work on undocumented implementation details, your work will sooner or later endure the same fate as an MC mod (and every new version will sound like bad news for you).
I firmly disagree. The format is well-specified, versioned and documented in Minetest's official documentation. It's subject to community input and not dictated to users like Minecraft's proprietary format. Any change can be flagged a newer version, but version changes are still disruptive and if the change is unnecessary, unhelpful. It was controversial enough that Minetest 5.x moved to a new protocol format and left a heap of iOS and offbrand Android app users behind. So many servers are slow to update as well, so the change just being on the server side doesn't help either. So I imagine a new world format could be very disruptive indeed. Also, you might be surprised how common such external programs are: Minetestmapper or Minetest Mapserver are just about ubiquitous for server operators for instance
/˳˳_˳˳]_[˳˳_˳˳]_[˳˳_˳˳\ Advtrains enthusiast | My map: Noah's Railyard | My Content on ContentDB ✝️♂

Astrobe
Member
Posts: 570
Joined: Sun Apr 01, 2018 10:46

Re: Infinite world size

by Astrobe » Post

Blockhead wrote:
Wed Aug 03, 2022 18:26
Astrobe wrote:
Wed Aug 03, 2022 17:49
Yeah, right, thank you for the lecture, master. A quick remark about how I misused SI notations with a link to wikipedia just in case would have been enough, though.
Sorry, you just hit a pet peeve of mine :)
That explains it. I usually like to write accurately but I was too focused on the topic and I had in the back of my mind the idea that 1 block is a one-meter cube is a loose conventional approximation (stairs with 50 cm tall steps are actually hardly usable... But voxel brings us so much fun that we don't mind that kind of detail).
IThe format is well-specified, versioned and documented in Minetest's official documentation.
Ok then. You were not talking about the things I was thinking about.

DOOM_possum
Member
Posts: 172
Joined: Sat Mar 27, 2021 22:06
In-game: DOOM_possum

Re: Infinite world size

by DOOM_possum » Post

any news, or word, yet, (On This Feature) ???

User avatar
debiankaios
Member
Posts: 910
Joined: Thu Dec 03, 2020 12:48
IRC: debiankaios
In-game: debiankaios Nowe
Location: germany
Contact:

Re: Infinite world size

by debiankaios » Post

I heared nothing :(

c56
Member
Posts: 307
Joined: Wed Apr 21, 2021 03:05
GitHub: tigercoding56
In-game: bm5 or bemo5 also sell_her_on55

Re: Infinite world size

by c56 » Post

every time someone brings up this feature it ends in a looong discussion about why people dont want it (even though according to polls they do ) instead of progress in implementing it (among with dimensions )
this is a signature not a place to post messages also if i could change my username i would change it to sell_her_on55

User avatar
LMD
Member
Posts: 1386
Joined: Sat Apr 08, 2017 08:16
GitHub: appgurueu
IRC: appguru[eu]
In-game: LMD
Location: Germany
Contact:

Re: Infinite world size

by LMD » Post

c56 wrote:
Thu Sep 29, 2022 02:16
every time someone brings up this feature it ends in a looong discussion about why people dont want it (even though according to polls they do ) instead of progress in implementing it (among with dimensions )
Finding people who want a feature (usually players) is a lot easier than finding developers to implement the feature.

This feature pretty clearly is high effort (nontrivial to implement, would create merge conflicts with practically all PRs and likely introduce subtle bugs to linger around for a decade), low reward (no one can even store a full world with the current limits, so workarounds like stacking dimensions are definitely feasible) and thus understandably low priority in the eyes of most devs.

There always is a trade-off to be made: How important is a feature to players & developers? How hard is it to implement? Prioritization is necessary. Minetest development slowed down to a crawl because of lack of proper prioritization. Fork Minetest and implement it yourself if you want it so dearly.
My stuff: Projects - Mods - Website

User avatar
debiankaios
Member
Posts: 910
Joined: Thu Dec 03, 2020 12:48
IRC: debiankaios
In-game: debiankaios Nowe
Location: germany
Contact:

Re: Infinite world size

by debiankaios » Post

LMD wrote:
Thu Sep 29, 2022 10:56
Finding people who want a feature (usually players) is a lot easier than finding developers to implement the feature.
Really? https://github.com/proller/minetest/tree/minetest32
LMD wrote:
Thu Sep 29, 2022 10:56
This feature pretty clearly is high effort (nontrivial to implement, would create merge conflicts with practically all PRs and likely introduce subtle bugs to linger around for a decade), low reward (no one can even store a full world with the current limits, so workarounds like stacking dimensions are definitely feasible) and thus understandably low priority in the eyes of most devs.
Yes, you don't really think it'll change in future? And in 10 years it will be harder to do, as if we would do it now because there will be many more features.
LMD wrote:
Thu Sep 29, 2022 10:56
There always is a trade-off to be made: How important is a feature to players & developers? How hard is it to implement? Prioritization is necessary. Minetest development slowed down to a crawl because of lack of proper prioritization. Fork Minetest and implement it yourself if you want it so dearly.
I don't know what we do wrong. Mindustry is also so much fun like minetest and here are stats of minetest and Mindustry:
  • minetest has 625 cotributors, Mindustry has 482 contributors
    In minetest are 994 of 6551 Issues closed and in Mindustry are 3283/3283 Isues closed
    minetest-devlopment begann at end of 2010(12 years ago) and Anukens(maintainer of Mindustry) made his first commit in Mindustry at end of 2017(5 years ago)
That are only 3 of many stats which show that other repos do anything other and get them goals. I know last time that a update(I think it was 5.5.0) comes 5 months to late and some features which should been added weren't added. I think if we do anything other there wouldn't be so much problems.

Post Reply

Who is online

Users browsing this forum: No registered users and 2 guests