[Mod] lib_ecology [lib_ecology][git]

[ShadMOrdre]

_E7,

I've been very busy with updates. I know things have been broken in the process. With that said, I have not yet updated github, but have a copy that you might otherwise dl and try.

If you are willing / able, you can download the map, and all relevant mods, here. https://github.com/ShadMOrdre/MTGameDev/tree/master

I'll try to get the individual mods updated on github this next week.

[DrFrankenstone]

I got MTGameDev working and it looks rather amazing, however I error out in about 15 seconds with out of memory (the machine has 32GB, but this is a LuaJIT limit rather than a native memory issue).

Tip 3 from paramat's Memory use optimisations for Lua mapgens is crucial, and not being used by these mods.

I'm not sure how you're avoiding OOM errors on your machine, and I'd offer a pull request rather than a link if I had more time atm, but those tips of paramat's are gold - as a first step, refactoring the calls to get_voxel_manip() gets large memory returns with little coding effort.

[FreeGamers]

OK, i've submitted a pull request for adding aliases to lib_materials. There is a couple where i couldn't figure out the origin but the bulk is covered. Let me know what you think.

Having had a closer look at the code i have some observations/suggestions:

- I'd split the crafting recipes and probably also the item definitions from the nodes. The code is pretty massive and it's hard to get an overview so breaking it down into smaller chunks would improve maintainability a lot in my opinion.

- There is quiet a bit of duplication among the nodes. I guess you just didn't get to it yet but given some nodes are just differentiated by numbers there is probably some that could simply be deleted without losing any noticeable variety.

- Some nodes could/should be added to more generic groups. For example there is like 4+ marble/granite nodes. If all of them get used in the map it would be somewhat annoying to just have a single one of them work with technic. Even if just one of them gets used being able to refer to a group instead of hardcoding the actual node would be advantageous. I think patching the target mod would be the lesser evil here.

- The mapgen/biome code feels out of place in a library dedicated to providing materials. It's not much of a problem though since this code is confined to it's own files and doesn't really get in the way. It just feels wrong.

Edit: I know that lib_materials is ment to be a general resource but since it's main (public) use seems to be to provide nodes for lib_ecology it might be a good idea to make it into a specialized world node library. People who want to use lib_ecology won't get much out of the building nodes since they often aren't even craftable.

I agree with most of these suggestions. To me they seem sensible. I've spent a few hours trying to digest what this project is about and what it can do. I'm new to the Minetest community and have spent that last week going through the forum and various gits and trying new mods to see what this game and engine are capable of.

To me it seems that one can grab various pieces of mapgen chunk mods and start throwing them into the default engine, hoping that they work, to create more unique maps or customizing to ones creative intent or vision. However this requires juggling a lot of parts, code, redundant nodes, dependencies and complexity. Thats mostly fine considering how well this engine handles old mods and interoperability in most cases.

However, I agree with ShadMOrdre that the original map generator, while sufficient, can be a bit stale. The dynamics of it over time are not very compelling to one who looks with a scrutinizing eye. The biomes are repetitive. I was originally using sets like MoreTrees, Plants_modpack, Biome_lib, nsspf (not so simple plants and fungi) along with a few others (like lightning, rain/thunder, ambient sounds, etc )to spruce up the game-play. The result is much added foliage and more naturalness, which I really enjoyed. But this ends up really making a complex beast. I have an interest in the natural world, so I'm glad projects like this (and mapgen geeks, if I may) exist. But it took me awhile to understand what this project is about and what it is hoping to accomplish, as its quite complex at first glance. Perhaps others may feel the same way considering that this thread as over 2000 views and a few thread contributers. Looking deeper into this, I now understand that it is essentially an effort by several people to make mapgen more natural and unique while incorporating the fundamental basic nodes and materials to generate this world. Most people interested in mods and maps have committed to some collection or map already and its quite a bit component to set into play.

This may hamper the project this substantially limits this mod from its modularity. It locks out other mods.

I actually pulled the master git last night for both lib_ecology and lib_materials, wiped my server world, plugged these components into it and attempted to generate a world. The versions from master, last night, generated interesting hills and valleys but were devoid of trees or other plantlife. It also had nodes with really noticable noisy transitions all over. I dumped the master versions and moved to the V2 release that came out, that version worked fine apart from a few missing texture errors. I'm currently looking forward to a V3 release that you mentioned right now. I think additions like xoceans would be interesting.

Right now as it stands, without the aliases that _E7 is mentioning, it will be a very complex process to get this mod "in production" for admins that want to add anything else besides the default. Getting mobs, for example, requires specific nodes to be specified for spawning. With your current usage of dirt_1, dirt_2, etc and the numerous nodes for each biome set, its very difficult to envision how I will customize a new server using this library. I think the group alias idea may have some potential. But most mods still build themselves for a standard set of Minetest biomes or ethereal biomes, so even with alias groups, many of the generated biomes will likely not have an appropriate group to bind to. I see the more expansive biomes a great thing. But I'm still struggling to wrap my mind around how to use this generator and make a compelling game, server, and community around it without great difficulty in adjusting existing and new mods to this. Surely player interaction and artificial life goes hand in hand with a natural-feeling world?

This generator does indeed make some very interesting biomes. I found myself flying around exploring things much longer than I do with v7 or Ethereal. Does this mod change cave generation? I only explored that a bit and I would explore the use of things like Nether, Cave Realms, and df_caverns with underground generation. For above ground I like to use cloudlands and otherworlds. I'm not sure how these will all inter-operate or if thats possible. Ultimately it may not be a goal for the developer of this project to add such functionality, and that'd be fine to me. These are more my passing thoughts as a curious user. How deeply will crafting be affected? Are wood, stone, grass, blocks all grouped so that one doesn't need to program numerous recipes for new additions? How will I set mobs and items to spawn in various sensible regions?These are the types of concerns I have when looking at using this project to make my world and server be more natural, interesting, and unique, while offering extendability to the game. They are not necessarily your problem, but easy solutions to problems like this could perhaps lead to more use of this really unique and interesting utility. Thanks to all involved regardless.

I'll try out the last push you did from earlier today too.

[FreeGamers]

Oh, my personal interest and perhaps goal is to use a tool that will great a world that will make playing with it, long term, more compelling and interesting than vanilla v7 or other mods. I'd like to craft a cohesive and challenging survival / default mt game around this and add what I like over time. My sister and I are both naturally into biology and ecology so getting her to eventually play this game with me regularly (since she now lives long distance) is eased with a compelling natural world. I'd also like to create a server that could be used long term and retain a user-base by offering something unique and substantive to players. Innovation, stability, and compatability are usually quite difficult to obtain simultaneously in software projects it would seem.

I was obsessed with crawling this community for all the interesting mods I could find this last week. I've recently graduated university and am in the process of applying, interviewing, and getting hired (which I sense will be very soon). For now, I've dedicated a good chunk of time recently to familiarizing myself with the modding community here at MT. Currently, I think this one of the more compelling world projects because a lot have been depreciated or have a much smaller scope. I will test it out a bit more.

[ShadMOrdre]

Dr. Frankenstone,
The Lua optimizations are well received. The only places that I can recall that use v_manip are in the lakes lua file in lib_materials, and the geomoria mod. Both mods were written by others, but it is my goal to integrate them better into the lib_materials mod.

Please note that I use a version of MT compiled without LuaJIT, and run on Lubuntu 17.10. I disabled LuaJIT long ago, simply because I find it too constraining for any serious collection of mods. Try running Advtrains and mg_villages. That combo is what inspired me to quit LuaJIT. I still use Advtrains, but do not use mg_villages any longer, preferring my own combo included in the DevWorld download, which is based on Rochambeau's Settlements mod, IIRC.

Also, please note, the version included with the DevWorld download is the most recent version of lib_materials / lib_ecology. Hot from my keyboard, as of this morning. The master, v2, and v5 branches are now officially out of date, and I personally discourage their use. I keep the code around for historical purposes. The master branch is my first release version, and is essentially a mashup of valleys_c and ethereal. The v2 branch includes some updates, but I wasn't happy with the results. The v5 branch is only a couple months old, but I noticed large areas that were devoid of plant life. This has been corrected in the version included in the DevWorld download.

jeremyweston,

Lots to comment on here. Please be advised, as per the above statement, that all previous versions of lib_materials and lib_ecology are outdated, with the DevWorld download including the most recent versions. I will post these updates to the respective github repositories this next week.

The DevWorld download is a basic, playable world. It's release is merely for the purpose of allowing others to quickly explore a playable world that includes lib_materials and lib_ecology. I had hoped to include the entirety of the mods that I use, and develop against. That download, unforetunately, came in at about 350MB. The truth is, I don't have to include them, just make a mod list available. But that kinda defeats the purpose of the DevWorld download, which is a light weight playable explorable world.

I grew tired of the MC simplicity long ago. My initial inspiration for this was to be able to use ethereal and valleys_c together. It just grew from there. I have a background in anthropology/geography/history educationally, and a professional career as a systems and data analyst/developer with an eye towards pattern recognition and AI. There is much I want to accomplish, and am frantically working various pieces of code.

The ultimate goal is to provide a collection of mods from which users can tailor a game. I believe, with the mods in my current collection, that anything from Diablo/Skyrim to AoE/Civilization can easily be accomplished. And, this is my ultimate goal. Whether or not I'll achieve this remains to be seen, but I love to play MT, and love to code.

Any and ALL help in creating real world ecosystems, (essentially grouping plants, trees, fauna, and soil types) would be very beneficial. I've spent countless hours scouring Wikipedia to get plants and trees appropriately placed, while leaving enough of the older garbage code in place, so as to at least have playable, and enjoyable spaces to explore. Currently, there are 5 dirt types that generate as ores within each biome. 5 humidity points x 5 temperature points x 6 conigured elevation ranges x 16 dirt types x 8 "grass" cover types = 1 large range of possibilities. In the DevWorld downloaded map, a v7 map that disables caves, floatlands, mountains, and ridges, while scaling most default values by a factor of as much as 8, provides the "continental" feel of the provided map. I use this map daily for all development, so some things may get a little stale on this map. You can always create a new world, and either copy the world/game folder to the new world folder, or use that new worlds map_meta.txt file in the DevWorld folder, to generate any other preferred world.

Again, the DevWorld download is mainly intended as a quick release test / debug version. There is much on my end, that I was not able to include, that helps answer some of the larger questions regarding integration into the larger modding ecosystem, (no pun intended). I personally feel that it is past time to move beyond MTG as the only game that most modders develop against. I know Wuzzy has complained of this very thing in regards to MCL2.

Things will need to be broken in order to be rebuilt. Backwards compatibility is not a goal or concern. Rather than try to maintain compatibility with a rather large number of existing mods, it is my opinion that compatibility mods should arise that help this mod work with that mod, in a general sense. After all, it was that "compatibility mod" thinking that started my efforts on these very mods. In this way, neither of the affected mods need to ensure anything other than providing nodes and objects, and at most, a "don't break the working stuff" mentality. Compatibility mods then arise that help to bridge the gap between mods, for those that want to keep using the less modular, more dependent mods.

The game mod that I include in the DevWorld download is one effort to help to correct what is a very big problem, IMO. In MTG, default does entirely too much. Every other mod then needs to depend on default, often for nothing other than a singular part, and sometimes, a very trivial one at that. The game mod should provide anything that would be need by almost any other mod, but only from an API perspective. I do not believe that the game mod, which is essentially the game loop, should provide any game content, but only help to link up the various parts that make up a game. Thus, lib_system, lib_player, and such. Each of those modpacks should be combined. I should not need more than a singular point of reference for all things player related. As it is now, I need a multitude of mods to provide a decent player object/experience. All the lib_system mods should just be part of the basic package. And, if possible, all of it should be referenced from a single point in code, the game object. Neither modders nor users/players should have to remember from which mod a particular node comes. I should just be able to see a single nodelist.

I ramble..... Please use the DevWorld download, available on the screenshots thread here.

I do all this alone, not with a team. It is only I, a single father of a very active young boy. I code while he's at school, if I can, and late into the night, if I can. I enjoy this, and I am really liking the results. I do hope all of you do as well. I see some educational potential in this, if it is done in a way that is server/singleplayer friendly (read lightweight), and accurate, (if it is to contain educational value).

Thanks for your comments, suggestions, and criticisms.

Shad

[_E7]

_E7,

I've been very busy with updates. I know things have been broken in the process. With that said, I have not yet updated github, but have a copy that you might otherwise dl and try.

If you are willing / able, you can download the map, and all relevant mods, here. https://github.com/ShadMOrdre/MTGameDev/tree/master

I'll try to get the individual mods updated on github this next week.

OK, i think i am getting the bigger picture a bit better now. I haven't tried the world yet but rather just looked over the code and what can i say? You think big and i like it. Thing is i am only interested in the map generation aspect but i understand now how lib_ecology isn't really ment for standalone use. That's somewhat sad news for me but it's your project.

I'll check out the changes to the world generation later when i have the time to actually look around a bit.

[ShadMOrdre]

_E7

The map generation aspect is one of two things in the DevWorld download.

The included map_meta.txt file changes the default mgv7 parameters. You can copy this map_meta file to any new world, change the seed to your desire, and have the new mgv7 world generate with the modified parameters that I supply in map_meta.txt.

The other part, and probably the one that you are interested in, is the lib_materials / lib_ecology mods. These are standalone. Both mods can safely be extracted from the download, placed into your mods folder, activated for a world, and you should see these mods in action. The only caveat, is that lib_materials clears registered biomes, ores and decorations. Any other mod that also does this could potentially override lib_materials, thus disabling the biomes and ores generated by lib_materials, as well as, the flora generated as decorations in lib_ecology.

Please understand that while I am programming this with a bigger goal in mind, the goal includes enough modularity and cohesion so as to make integration with the exisiting ecosystem possible. While backwards compatibility isn't a particular goal, that doesn't mean it isn't important. I do want this mod used. Widely. I want to disrupt the simplicity that makes MC/MT seem dull after a relatively short time, at least for an adult. I hope to expand some horizons while drawing from, what IMHO, is the best of previous efforts. All media content is drawn from other mods, therefore it is only right to assume that others would want to integrate these mods with those mods, therefore requiring me to consider compatibility at all times.

I do hope that the code has been written straightforward enough as to make compatibility a rather trivial effort. While some node defs have been changed from the originals, this has mostly been done to make nodes act consistently, where necessary or desired.

Please do try extracting lib_materials / lib_ecology mods and using them in any other world of your choosing. This is the only way I'll really be made aware of compatibility issues, since I do play from a modified game, and my mods list is stable and doesn't include a lot of other popular mods. I'm working on setting up another github repository that includes the larger mod framework within which I develop. While Technic has not been in my mod list, I do want to support a very popular mod. To not, in my view, would only harm this mod, instead of help it see adoption.

I do see where a compatibility mod, or perhaps a pull request for me, to include support for technic and other mods is in demand. I'm hoping, with the Devworld downloads, to attract other modders, who may be willing to take on some of the ideas to help make all this happen faster.

This link, DevWorld_FULL, provides a more complete picture of my goals and direction.

Thanks for the continued interest. Please don't dismiss me yet.

Shad

[Gael de Sailly]

Glad to see about this mod again, looks amazing!
I will definitely try it out tomorrow, and maybe involve a bit in it (I always say everywhere that I will, but very often I don't find the time to actually do it...)

[ShadMOrdre]

Gael,

I'd LOVE (**wipes the drool) to have your involvement.

Your mod inspired this. I know you have expertise. I know you wannnaaaa......

Anywhere you can help....

And I edited the previous post to include a link to the more complete picture.

[_E7]

Well, to be honest i wouldn't even be completely opposed to contributing on a broader scale since in general i really like the direction you are going. I just don't think it would work all that good. The full scope of your project is just straight up massive and for an outsider it's very hard to get an overview of the actual details. In the end i need something i can work with and i fear i'd overextend myself trying to use the whole thing as a base. That's why i am concentrating on what's really useful to me: map generation.

Don't worry it takes way more for me to actually dismiss something. The main question for me right now is how much sense it makes to try to contribute. In it's current state lib_materials/lib_ecology make beautiful landscapes but aren't overly useful to me beyond that and changing this would probably need quite a bit of patching. Of course i'd rather contribute than maintain a private fork (or at least keep local patches to a minimum) but i wonder how well that would work. What i am after is not a primary goal (or none at all since i'd probably want to restructure some stuff) and you seem to be taking kind of a lone wolf approach (which i totally respect). I simply fear we'd slow each other down and waste time.

[ShadMOrdre]

_E7,

Goals are simple, I hope.

Immersion. Anything that adds to the immersive aspect. By this, I mean, things that happen inworld, versus, via a formspec. Crafting, in my opinion, is a big culprit here. I prefer mechanisms that make materials, versus crafting. Instead of dropping a tree node into the craft grid and pulling out 4 wood nodes, I prefer a saw, that could be added to a larger workshop, but also stands alone, modular. This kind of immersion.

The lib_materials / lib_ecology mods are meant as a base. Any mapgen requires nodes to generate. MTG supplies some, but they are basic, few, and really only mimic MC. My hope is to provide a broader base for mapgens. A way to generate any world that a user could conceive, and configure. Modders, especially mapgen modders, should be able to spend more time with the actual mapgen, and less trying to decorate. Materials/Ecology supplies the nodes, biomes, ores, and decorates. Hopefully this finds support, but if not, thats ok.

The lib_mapgen mod, included in the DevWorld download, is meant to be expanded to make world gen more intuitive for basic players. I envision a dialog where players can select more meaningful options, such as continental, archipelago, flat, mountainous, or variations to that effect, or to select various lua mapgens on the fly. Most users would get lost with the mapgen noises. I've spent months "perfecting" the mgv7 params that I use, and I am still not satisfied. And most of this time was spent just trying to figure out what a particular setting does.

All this aside, I am by no means a lone wolf. Just a very bored, former IT pro and now stay at home single dad. I may not be an extrovert, but I do keep comms open. As a mapgen, materials/ecology can stand on their own, but there would need to be massive support from other mods, or a very health community of folks willing to make mods cooperate. This is mainly the fault of modders who make specific mods that use MTG as a base. This isn't wrong. It just limits the ability of any other mod to integrate without also accomodating MTGs default mod. I would prefer to make a base that is game independent, instead of assuming that MTG is the default game, and thus relying on it, or having to accommodate it just to have the functionality of some third party mod.

Shad

[_E7]

Immersion. Anything that adds to the immersive aspect. By this, I mean, things that happen inworld, versus, via a formspec. Crafting, in my opinion, is a big culprit here. I prefer mechanisms that make materials, versus crafting. Instead of dropping a tree node into the craft grid and pulling out 4 wood nodes, I prefer a saw, that could be added to a larger workshop, but also stands alone, modular. This kind of immersion.

That's a perfectly valid approach. I'm not really bothered by crafting itself though. If there is something about it that bothers me it's the uglyness of the formspecs. Besides given amount of available nodes it's probably impossible to completely get rid of crafting but i guess i am just stating the obvious here.

The lib_materials / lib_ecology mods are meant as a base. Any mapgen requires nodes to generate. MTG supplies some, but they are basic, few, and really only mimic MC. My hope is to provide a broader base for mapgens. A way to generate any world that a user could conceive, and configure. Modders, especially mapgen modders, should be able to spend more time with the actual mapgen, and less trying to decorate. Materials/Ecology supplies the nodes, biomes, ores, and decorates. Hopefully this finds support, but if not, thats ok.

At least before you get to something stable i wouldn't bother with supplying "extras". It might not be a problem for you but in my opinion it creates needless confusion for other people trying to work with the code. Not to mention is probably way easier to fully flesh out and organize a limited set than having to take a bunch of surplus nodes into account that might be never used at all. It's not like people are going to write custom mapgens against what is currently a moving target anyways.

The lib_mapgen mod, included in the DevWorld download, is meant to be expanded to make world gen more intuitive for basic players. I envision a dialog where players can select more meaningful options, such as continental, archipelago, flat, mountainous, or variations to that effect, or to select various lua mapgens on the fly. Most users would get lost with the mapgen noises. I've spent months "perfecting" the mgv7 params that I use, and I am still not satisfied. And most of this time was spent just trying to figure out what a particular setting does.

That's certainly a nice touch but to be brutally honest i don't care much about users ;)

All this aside, I am by no means a lone wolf. Just a very bored, former IT pro and now stay at home single dad. I may not be an extrovert, but I do keep comms open.

Sorry, i guess that was easy to get wrong. It wasn't really directed at you personally but more your release style where you publish code that might reflect months of work done in private. I mean the alias code i wrote was probably already outdated by the time i started, wasn't it? ;)

As a mapgen, materials/ecology can stand on their own, but there would need to be massive support from other mods, or a very health community of folks willing to make mods cooperate.

I don't see this as much of a problem was long as the mods where compatibility can not archived with aliases accept patches since the needed changes would likely be trivial. The part where i see more work needed is actually organizing and fleshing out the defined nodes. Right now there are a lot of rough edges. Like there is a red clay variant but no red lumps and obviously nothing that drops them while white clay depends on default:clay when it seems it should be related to the white clay dirt(?). A bunch of stuff references default:dirt which as i understand it shouldn't exist at all. The "deco" nodes are somewhat unfortunate in my opinion. Missing groupings. And so on.

That's also what i was getting at with stabilizing a limited set of nodes first. I know you only have 2 hands but as is there are quite a couple of loose ends and then there is also the points concerning structure i mentioned earlier.

[_E7]

.

[Gael de Sailly]

I've just tested it today, and I must say, it goes far beyond my expectations. You basically managed to do more or less what I was trying to do in Valleys Mapgen or 30-biomes but never came up to something satisfying: a truly various, yet consistent, biome/plant system, and it looks very good.
Your biomes look rich, diverse and not stereotyped. This latter was my constant guideline when developing my mapgen mods, but I can't say it was achieved, or at least not as good as lib_ecology. So much variety, so many different plants, dirts, rocks, and it seems that after 20 minutes of exploration I've basically not seen twice the same landscape.

+ Screenshots

I'd also like to report some minor bugs/imperfections that I found so far:

+ Issues

• I get many errors like this:

  Code: Select all

  2019-06-02 20:33:31: ERROR[Server]: Map::setNode(): Not allowing to place CONTENT_IGNORE while trying to replace "air" at (-2496,-5,-1761) (block (-156,-1,-111))

  I tried teleporting in the place it occurs, it always involves your water nodes. I think the flowing liquid is not set properly, but when looking at the code quickly I haven't been able to spot the cause.
  
• Snow grass looks enormous, I don't know if it's intended (I suppose it's not). This reminds me much of a bug VanessaE experienced a while ago. I don't remember what was the cause and how she fixed it, and I haven't found it yet in the commits/issue tracker (I'll edit my post if I find it).
  
• Multiple layers of dirt with snow doesn't look nice, there should be plain dirt below. Also what are the stone tiles at the left?

Something I'd like to highlight particularly: It looks like you have done a pretty nice job regarding how biomes interact with each other, I mean especially biome transitions. This looks like a detail but is probably the thing that will make your terrain look "natural". However, it seems that at some points, there is still some to be done, some transitions look way too sharp and artificial, see the screenshot. It would be good to add intermediate nodes and/or biome blends, "smooth" a bit the conditions needed to spawn a particular plant, and also work on textures to make different nodes look more harmoniously together (in particular avoid too bright colors). I don't know if that's the direction you're taking for the development of lib_materials/lib_ecology but it would definitely be mine if I was you. Just a piece of advice and a suggestion.

+ Too sharp transition

I've seen your recent posts, but I'm not sure to exactly get your point (to be honest i'm too lazy to read all in details). It looks like you have a very clear outline of your project, this is a key for success. I see that you are taking distance from the default game, this is a good thing too.

Please understand that while I am programming this with a bigger goal in mind, the goal includes enough modularity and cohesion so as to make integration with the exisiting ecosystem possible. While backwards compatibility isn't a particular goal, that doesn't mean it isn't important. I do want this mod used. Widely. I want to disrupt the simplicity that makes MC/MT seem dull after a relatively short time, at least for an adult. I hope to expand some horizons while drawing from, what IMHO, is the best of previous efforts. All media content is drawn from other mods, therefore it is only right to assume that others would want to integrate these mods with those mods, therefore requiring me to consider compatibility at all times.

The idea to make it compatible with every mod seems way too ambitious, and in my opinion, is unlikely to give good results since mods are made by different people with different goals and different ways to achieve them, and simply different styles, that won't fit well together (the most flagrant being textures from different mods that often look ugly when used together).

In my opinion, the best thing to make out of your mod collection is turning it into an actual (sub)game, on which your mods would completely replace default. This is the only way to get rid of default's materials, the game would be based on your lib_materials so no need for any overriding of nodes or biomes (I HATE overriding. It is often presented as a feature to provide compatibility between mods, but usually results in the opposite, when mods override each other until we can't even understand what's going on).

All the games that you can see and download from the Minetest forums are either awful copies of Minetest Game with several random mods added (what someone I don't remember the name was calling "mod salad"), or simplistic things made from scratch. IMO the only good existing game is Sofar/Nore's InsideTheBox, but it's extremely specific and for other purposes, there is simply nothing worth being cited.

I think you're holding one of the rare occasions to make an actually good and enjoyable game, I mean, something that differs from the basic and cheesy Minetest Game, and that is very complete and polished. This isn't done by adding compatibility to 50 random mods like Dreambuilder but by selecting carefully what you include in the game or not, modifying them as much as necessary so that they fit well in the whole, both technically and stylistically, and make as much content as possible by yourself.

Your project seems definitely very ambitious by I do hope that enough people will realize the potential of your work and involve to the extent of their time and capacities. I myself more and more think I will involve in it, since this is extremely close to what my overall modding project was, some years ago.

I surely need to try your demo world to understand your plan better, will do. This world thing seems pretty mysterious for me, for now.

I now realize I gave you 4 hours of my time to give you this feedback, and I need to sleep…

[ShadMOrdre]

Gael,

Thanks for such kind words and advice.

I get the same errors regarding liquids, and am working on a solution.

Snow grass is overused. Most of the cold biomes have this currently as the standard cover. Stereotyped. This is one of the areas I'm cleaning up. All biomes should present as unique, and the ecosystems that generate within them should also be unique. I've also modified the humidity / temperature values used by lib_materials to generate the biomes. Previously I used 90, 70, 50, 30, 10, as the various values to define hum/temp. I've reset those to 100, 75, 50, 25, 0. This should knock back the cold and hot biomes a bit, but should also even out the distribution a little better. I myself was finding cold biomes a little more often that I thought I should be encountering them.

The arid areas are also getting a makeover, as the transitions were "ugly" in my opinion. They still might not look right, but I do think they are getting better.

As for the bigger picture.

The goal is to simply make lib_mat / lib_eco a drop in replacement for default, as far as I can. Default includes far more than just a few nodes to spawn. Compatibility with the mods from which these mods draw should be a minimum level of compatibility. I can't say I'm too worried about most mods, since I am reworking some, but I know I won't be touching mods like mesecons, pipeworks, advtrains and the like. But I would like to ensure some level of compatibility, where necessary. Mainly craft recipes. I use the above mods, but have not verified if crafting works with lib_mat / lib_eco.

A (sub)Game is the goal. I agree with your statements regarding the existing modding landscape, and while all MT games will always be not much more than mod soup with side of mod salad, I do believe that the "right" mods can make a cohesive, coherent, consistently playable game. If MT is to grow, it needs compelling reasons for people to want to download and play. Regurgitated MC is not compelling. Mod soup can be, if.....

Honestly, the VAST majority of the ideas that I am trying to implement, are the thoughts and opinions of folks here on the forums. "Man I wish MTG had this..... " or "MTG really needs..." I've been reading these forums for a few years now, and there are great ideas all over. It just seems that very few modders are listening.

The other part of that equation is that there is no clear, consistent, easily digested style guide. I've worked hard on these mods, to ensure that graphics are 16 color, 16x16pixel. This has gone a long way in consistency, as well as, performance. So many modders use 32x32 RGB textures, and then wonder why there is lag, then spend hours, days, weeks, digging through code to find the lag, and ultimately giving up because of the lag. Default jungles are a BIG reason for lag. Why? It's not the shear number of trees. It is the 5.5kb jungle tree texture and the 5.1kb jungle tree top texture, multiplied by the shear number of trees.

BTW, thanks for the four hours. I have a feeling you've given these mods some more time too..... ;) Remember, you are partly responsible for the very existence of these mods.

And again, thank you much for the constructive ideas and advise. I hope you continue enjoying "our" efforts.

Shad

[ShadMOrdre]

Major update.

Saplings spawn trees.

Plants grow.

Minimal, if any, ABMs. All plants and saplings use node_timers. Runs light, with little if any background processing.

Relies soley on the engine to generate biomes, ores, and decorations.

[CalebJ]

Thanks for this update. This is simply awesome to combine all the mapgens into one ;) lib_materials, however, is quite bulky, is there a way you could make it more optional in the next commit? Thanks!

[CalebJ]

Sadly I got quite a few bugs in the init.lua when I tried it with ethereal, cool_trees, aotearoa, and australia, and biome_lib. Once I fixed all the bugs, then I got this bug:

Code: Select all

A serialization error occurred:
Malformatted inventory list. list=main, read 1484 of 4134 ItemStacks.
The server is probably running a different version of Minetest.

If you could help resolve this bug, I'd be grateful :) There are more errors in the init.lua, please debug it all.

[ShadMOrdre]

CalebJ,

About the lib_materials dependency, I thought about this long and hard. Other than the current configuration, I essentially have only two other choices. I can combine the the two mods or try to accommodate all the other mapgens where lib_ecology would / could be used.

lib_materials isn't as heavy as one would think. The biomes used to get defined here in lib_eco, but I realized it made more sense to define both the dirts and the biomes in lib_mat. Previously, lib_mat only really supplied nodes and ore_defs, but now, it seems more balanced.

The error you are encountering could be due to the MT version you are using. When 5.0 was released, I ran into this issue straight away. I do use a LOT of mods. A 5.0.1 patch has been released that "should" fix this for you. Please start with this, as the malformatted inventory list is due to a fix for a previous issue, that then overlaid with inventory transfers. The 5.0.1 version fixes this. If it does not fix the issue, consider running from a modless setup, adding mods until the error reproduces. This is error is related to really large amounts of nodes getting defined. As an example, I currently use mods that define upwards of 19000 nodes.

Shad

[CalebJ]

Yes, I was using v5.0.0, I'll test in v5.0.1. Thanks for the info!

[ShadMOrdre]

MAJOR UPDATE

[ShadMOrdre]

Posted another update to address some issues I've been finding.

Also include now, are nsspf mushrooms. Other nodes were also fixed, so should now be available.

Check it out. I doubt you'll be too disappointed... ;)

Shad

[ShadMOrdre]

I have pushed some fixes to missing textures and other peculiarities.

Please let me know what you think.


Shad

[StarNinjas]

This looks cool, but I am getting errors saying "Map::setNode(): Not allowing to place CONTENT_IGNORE while trying to replace air at(xyz) (block(xyz)" So the mapgen is not working :|

[ShadMOrdre]

StarNinjas,

I've noticed these errors also. I thought I had tracked the cause of them, for the previous update, but alas, I was wrong.

These errors are being generated by some decorations that I'm still tracking down. I do apologize for the log spam, but I have no control over those messages, and they are too vague to attempt debugging.

Thank you for reporting this issue. I'll get an update to this posted soon.

Shad