Minetest_game: Split MTG into two games - one a game, the other a mod base

[rubenwardy]

Our users consistently say that they want a more polished and complete base game.

So I suggest maintaining two games:

Minetest Foundation (feel free to suggest a better name)

A base for modding.
I suggest removing:

• Carts
• TNT
• Keys (should be made a separate mod in MTG, it wasn't merge long ago so there shouldn't be any mods relying on it)

Minetest Game

A full polished game in its own right. Modifications to mods in MTF should be minimised (in order to avoid conflicts), and changes should instead be done as separate mods. We should work to add mobs etc

[tobyplowy]

@rubenwardy totally agree we really need something to show that minetest is a good engine and the only way to do that is to make a good game (NO minetest game how it is right now is not a good game!) We need something called gameplay so People Don't get bored instantly
(you shouldn't be forced in to installing mods to have fun!)

And please Everyone who reads this please don't say this is a mod thing take it with an open mind this can improve minetest a lot!

[stujones11]

I presume by 'in its own right' you mean that it will not depend on MTF because we had something like that once before and it did not go terribly well. If you do not depend then there is going to be much code duplication though I do like the idea in principle.

[rubenwardy]

code duplication

That is an issue - I'm open to suggestions

[stujones11]

Could something like this not be done with sym-linked repos (I'm no github expert) so a single base modpack could be linked to each game?

[rubenwardy]

Would require --recursive download however, and scripts would need to be updated

Maybe Minetest Game could just be a fork of Minetest Foundation that cherry picks changes every so often? Cherry-picking is pretty easy

[stujones11]

Yeah, I think the core mods would be unlikely to change that much other than odd bug-fixes as most of the focus would be on the new game anyway. I'd guess that back-porting fixes might be the bigger maintenance headache but I agree that it's not difficult if you keep on top of it.

[alexeysvrv]

I agree. I want to assume full basic game with mobs, automation, machinery, etc. But it is necessary to maintain a database of modding.
I propose writing mods official (by developers base game) to be included initially in the game. A good idea in general

[sfan5]

We've had this in the past and it seems it wasn't optimal since we no longer do

[Fixer-007]

Or current full MTG becomes "Minetest Classic", and "Minetest Whatever" becomes full blown with much more features, mobs and other.

I'm against removing TNT and Carts from the Base.

[stujones11]

I'm against removing TNT and Carts from the Base.

I think that would be missing the whole point, who would be using the foundation game other than those wishing to create their own unique game or server experience?

[rubenwardy]

We've had this in the past and it seems it wasn't optimal since we no longer do

Well, the alternative is to ditch "modding base" when it gets in the way of making MTG a fuller game

The question is - is it still a problem? Would it stop us from adding mobs or weather (if stable enough)?

[ghost]

As cool as it is, I don't really think nyan cat fits very well. Maybe that can get split off too.

[sofar]

keys are really not implementable as a separate mod, it cuts too deep into default as is.

[rubenwardy]

sofar: #1486

[paramat]

MTG has been developing a lot recently and optional mobs are intended for the future, it is already intending to be a more polished and more complete subgame and maintaining 2 subgames is not needed for this to happen, because it's already happening. There is no conflict between the need for a modding base and the direction of MTG, there would be conflict in a very specialised subgame but MTG is not going to be that sort of subgame.

However, if some people want to create a stripped-down version or a much more specialised version that's fine by me, we need new subgames, but MTG is developing well and should be left to continue in it's current direction and with no name change.

Essentially i object to the suggestion that the development of MTG has to be changed or that it has to be 'split', it's going well so lets leave it alone and new subgames can be added either side of it: either stripped down or much more specialised.

I can't see much need for a stripped-down version, i feel most modders would prefer to base their mods around a MTG with more content, i certainly do. I can't remember any modders complaning that MTG has too many features to be moddable. It can be a good subgame and be the base for modding, with carts, TNT and keys etc. left in and simple optional mobs added.

Well, the alternative is to ditch "modding base" when it gets in the way of making MTG a fuller game

It doesn't get in the way.

[C1ffisme]

Alternatively, what about removing minimal and replacing it with Minetest Base? Minetest Base would still work as a minimal development environment, but would have all the functions and blocks needed for basic modding. Plus, it means that we have two useful games for the average (average meaning "at least knows how to use the lua API") Minetest user, not one game for the average Minetest user and one for Engine Developers.

@paramat

It doesn't get in the way.

Yeah, it really does sometimes. It's taken us a long time to add carts to the game, we still don't have anything in the way of mobs, etc. "MTG is simple" is another way of saying "Too many people depend on our subgame's minimal approach for us to add anything that would get in the way of mod testing, and we leave everything to third party developers."

[TechnoPrometheus]

After 7 years, finally you got it! Minetest 0.4.15 was a total disappointment. The skeleton keys? you have added unnecessary things, when I suggested adding simple decorative blocks, and content farming (all already developed by Tenplus1). Minetest is an engine, it is true! but users want to play, they want a complete game, at least the basic features of MC.

@rubenwardy never speaks, but when he writes is pure legend ^__^

@tobyplowy Everything you wrote is absolutely true! Minetest is a good project, but the end user (player) wants to play, does not have to download 10 mods for a complete gameplay.

[tenplus1]

+1 for a compatible BASE game for server owners and users to built onto, as well as +1 for having an adventure game for new players that doesn't require them downloading mods to enjoy.

[ghost]

I think @C1ffisme is absolutely right. I don't think anyone really uses minimal dev test, certainly not modders. I'm not even sure if engine devs use it. Since players don't use it, we might as well make it useful for you know, development.

[paramat]

Our users consistently say that they want a more polished and complete base game.

That's normal and expected, who wouldn't? But there isn't a problem because that is what we are working towards with MTG, and many reasonable people can see that.
Remember that those who complain tend to be more vocal than those who are quietly content with development. Comments above by tobyplowy and TechnoPrometheus are ridiculous and very negative, it's not as if we're being lazy and refusing to develop MTG, can you not see the rate of development and the commitment to a richer subgame?

In the short time sofar has been with us he has made more commits than celeron55 ever did, and i have made twice as many. The biomes are numerous and much richer, we have added carts, many new nodes and intend to add optional simple mobs, weather and ambience sounds too once we have some suitable engine code to do that well. The more special features will be optional and have parameters so that MTG can be made more simpler if wished.

A more complete game is not a problem for modding, a mod just depends on certain parts of MTG of it's choosing. MTG can be more polished and complete and still be the base for modding, so although i welcome new subgames they are not actually essential to having these 2 things.

I suggest that we don't have a stripped-down version, since that's not really needed and will just add workload, i suggest that any additional subgame is more complex and specialised.

Concerning minimal, i have suggested we trash it or remove it from the engine, but devs have said they use it and find it useful placed in the engine.

[paramat]

Yeah, it really does sometimes. It's taken us a long time to add carts to the game, we still don't have anything in the way of mobs, etc. "MTG is simple" is another way of saying "Too many people depend on our subgame's minimal approach for us to add anything that would get in the way of mod testing,

This is nonsense. The delay of carts and mobs has nothing to do with MTGs simplicity, it is due to dev time and other problems. There is nothing getting in the way of adding new features, and the minimal approach certainly isn't. In fact MTG is not really minimal anymore.

[C1ffisme]

This is nonsense. The delay of carts and mobs has nothing to do with MTGs simplicity, it is due to dev time and other problems.

Unless my memory is failing me, that would not be the case if I asked for this two years ago.

[rubenwardy]

if I asked for this two years ago.

Because MTG was in a feature freeze then, which has since ended. And also carts were not good enough, technically

[stujones11]

While this does seem like a popular idea, I think most people would be equally happy to see the continued development of the current game so long as any new additions are kept as modular and configurable as possible.

[alexeysvrv]

@paramat You opened my eyes. I began to look differently on this issue. Thank you

[tobyplowy]

@paramat I agree with you it's true we are complaining and minetest is Moving along Good but it's not going Fast Enough ( I know that There is Only A Small team of Developers) but seriously there's things that people just keep asking for but you don't give it like for example mobs when are we going to see them the only thing I Really Want is mobs, this is an engine yet one of the most important Things Is Not There (I know there's mods for this) but a propper mob ai is way overdo!

Also Sorry that I was so mean yesterday but I'm just tired of this, minetest just has so much potential but it's not doing Anything with it ( I even think minetest would be able to kill Minecraft in the long run) but Before that happens It Still needs a lot of stuff

You are one of the people who improve the future of minetest (and you do alot of work to do so all of us do) but please look at things with an open mind it's better for everyone, for you, for me, and most important for minetest (thank you for all the time you put in this project I really appreciate it)

[rubenwardy]

I think most people would be equally happy to see the continued development of the current game so long as any new additions are kept as modular and configurable as possible.

I'd be happy with this, if others are

[C1ffisme]

Because MTG was in a feature freeze then, which has since ended. And also carts were not good enough, technically

Too clarify, I didn't mean _exactly_ two years ago.

[paramat]

but it's not going Fast Enough ( I know that There is Only A Small team of Developers) but seriously there's things that people just keep asking for but you don't give it like for example mobs when are we going to see them

I'm just tired of this, minetest just has so much potential but it's not doing Anything with it

Ridiculous childish rantings. We're working our asses off on MTG.

[tobyplowy]

@paramat I'm not saying you're not working hard enough im saying that we need more people, development is not going fast enough

[stujones11]

Going off-topic a little, I also consider the minimal test game something of a relic. I can't remember ever using that game other than by mistake. Maybe that could be replaced with something useful like a truly minimal sub-game template.

[ghost]

@stujones11 I completely agree. By replacing just the devtest, we'd let users keep enjoying the stock survival experience while still providing a nice base for modders and developers.

[paramat]

I admit i'd like to remove minimal too, but met dev resistence. I have never used it, it's a pain to maintain and is so different from MTG it's not very useful.

[C1ffisme]

@paramat Even updating minimal to behave with mods the same way as MTG (Besides crafting recipes, obviously) and update the textures (Since, as nostalgic as they are, they are quite an eyesore) would be just as good as splitting MTG.

[sfan5]

it's a pain to maintain

Nobody is maintaining minimal, how can it be a pain?

[rubenwardy]

A MTF would be easy to maintain - it's just cherry-picking when bug fixes are made

Nobody is maintaining minimal, how can it be a pain?

Agreed

[Zeno-]

@paramat

I admit i'd like to remove minimal too, but met dev resistence. I have never used it, it's a pain to maintain and is so different from MTG it's not very useful.

Minimal is essential for performance profling, finding mem leaks, etc, etc, etc. It's very useful and required.

[paramat]

Ok, it's fine by me to keep it then.

[stujones11]

Back on topic, maybe there could be an officially sanctioned sub-game similar to the old next branch but given its own repo and possibly other, non core developers can have some say in its direction and in some cases even push access. It could be a bit more liberal in its merge policy and provide a kind of testing ground for the official game.

[C1ffisme]

@stujones11 I _do_ like the idea of a way to build a subgame with more of a community feedback.

[stujones11]

If enough MTG developers were to agree to oversee such a project then there should be no need for outside push access, instead there could be a committee of trusted contributors with up-down voting rights. It should be easy to find the best likely candidates by looking at the recent commit history.

[sorcerykid]

I would like to third stujones11's suggestion. The Next branch was a step in the right direction. Perhaps a similar approach should be taken for this next phase (no pun intended)

[Wuzzy2]

If Minetest Game is going to be a … well … game, this raises some obvious questions:

• How do you imagine Minetest Game to be in version 1.0.0?
• Will Minetest Game have (gameplay) challenges and goals?
• What will be the objective/goal of the game (i.e. goals for the player)?
• What new core game mechanics are planned (e.g. stuff like hunger)?
• What's the focus of the gameplay (survival, creativity, MMO, etc.)?
• What will be the genre?
• What will the player be able to do (think of a “typical play session”)?
• To which kind of player do you want to cater (e.g. casual, hardcore, LEGO builder)?
• How will Minetest Game be distinct from similar games?
• Most importantly: What's the overall vision of Minetest Game?

These technical questions may also be of interest:

• How important will not breaking compability be in order to introduce new features? I.e. are new features more important than providing perfect backwards-compability?
• Is an in-game help system desired?
• Will the 2-approval system stay in place?

If you are able to give a good answer to all questions, then Minetest Game may be successful, after all.
I'm indifferent about the proposed subgame split.

Oh, and for all who are still disappointed in Minetest Game: Try out other subgames (look in the forums). There are some good things to find, like Pixture, Lord of the Test and rpgtest.

[paramat]

It's the wrong approach to call it a game and then have a list of requirements for a 'game'.

There won't be a version 1.0.0, Minetest isn't like other games, it's never finished or complete, or already has been for a long time.
We have challenges and goals, but should not have a 'final objective', MC adding 'The End' was a mainstream / commercially-driven mistake that is out of place with the open-ended nature of these games, it was the point where i feel Notch lost his way with MC, and indeed soon after stopped developing it.
We don't need a focus, MTGame is multi-focus.
We can't have a genre, it's flexible and defined by mods.
The overall vision is just how it is, there's too much emphasis being put on having to sit around stressing out about having a 'vision' or 'direction', maybe there isn't one and it develops intuitively, if so that is no problem, often the best art develops that way.

We welcome new subgames, these can be the place for goals, focus, genres, vision and plans, MTGame is fine developing the way it always has in it's low-stress, intuitive way.

If you are able to give a good answer to all questions, then Minetest Game may be successful,

So MTGame is a failure, and the way to success is answering these ridiculous questions.

[MarkuBu]

MTGame is multi-focus

That's not totally true. MTG tends to Minecraft like games (Trees, Pickaxe, Minecart...). If you want to make a futuristic game you have to remove/overwrite a lot of the predefined stuff. And some stuff like leave decay ABMs can't be removed without hacking the "mod base"

I would define a "Modding Base" as something, where you just have to add stuff for your game and not remove everything first before you can start.

To be honest: sometimes it is easier to write a subgame from scratch than using MTG

So you should define "mod base" before anything else. Maybe the whole concept should be reconsidered. Is a Picaxe or a Tree really default or is this already part of the subgame? And ores? Maybe a subgame don't need ores, or different once. Biomes? I don't think there are jungles on the Mars Survive Server

From my point of view even minimal dev is not a modding base because it defines stuff that is typical for a Minecraft like game. A modding base should be as minimal as possible. Everything else, even biomes, should be added as mod

The default mod in MTG is a big mistake. It defines a lot of stuff and it is difficult or impossible to make changes without changing the code.

[Wuzzy2]

It's the wrong approach to call it a game and then have a list of requirements for a 'game'.

This is an important answer. I was exactly looking for answers of this kind. This means Minetest Game is officially a sandbox, and ironically, NOT a game, although it is CALLED “Minetest Game”. xD (Note: I'm not saying that sandboxes are bad, I'm just saying that I see games and sandboxes as distinct things.)
You're right, all game-related questions are obviously moot because of this.

We have challenges and goals, but should not have a 'final objective

I don't see any challenges and goals. At least not any meaningful one. The focus on sandbox gameplay is very strong in the current version. I don't call “try not to die in lava” a real challenge. ;-)

But explicitly saying on record that there will not be a final objective is another good answer. Please more answers like this.

We don't need a focus, MTGame is multi-focus.

So what are the multiple focuses of Minetest Game, then?

We can't have a genre, it's flexible and defined by mods.

That's bullshit. Every sandbox or game has some genre, if you like it, or not.

Maybe we just don't have a name for it yet. Minetest has brought a couple of Minecraft-like games/sandboxes, and they are pretty similar: Voxel-based, main gameplay is digging, building and crafting and (maybe) some fighting. I'm pretty sure Minetest Game falls into this (unnamed) genre, and that this is also one of the (unstated) goals.

The overall vision is just how it is

So basically nobody knows where Minetest Game is heading to. It's just a random bunch of stuff a random bunch of people wanted to have.
Do you maybe want to say, just like the gameplay, the development is open-ended? Maybe I slowly seem to understand what's all the fuzz with Minetest Game.

There won't be a version 1.0.0, Minetest isn't like other games, it's never finished or complete, or already has been for a long time.

That's a very disappointing answer. Sure you can't even list some list of core features you want to definitely have? E.g. if you think hunger is a “must have” feature (just an example). Or do you think all major features are already there, and you do not plan to add any more major gameplay changes?
1.0.0 does not have to mean “final version”, the “1.0.0” could simply mean that Minetest Game is stable and “complete” enough for the average gamer not to throw it immediately out of the window. Oh wait, the version number is synced with Minetest's, this may be a problem. :-(
Intentionally keeping a “0” version forever, even if Minetest Game has become high-quality, is not a good idea since basically everyone will suspiciously look at everything with a leading 0, as this almost always means “beta”, “alpha”, “buggy”, “for testing only”, etc. Yeah, it's more a symbolic act than everything else. But IMO Minetest Game is not good enough for a 1.0.0, but opinions may differ on this.

If you are able to give a good answer to all questions, then Minetest Game may be successful,

So MTGame is a failure, and the way to success is answering these ridiculous questions.

You said that. ;-) Joke aside, the questions are not ridiculous. I meant every question seriously. I just wanted to know what you actually want. I wanted to understand where this is heading to. But at least you made good-faith effort to answer the questions (thanks), although I don't understand why you skipped some of them. To be honest, I'm currently not sure if Minetest Game will be a success or failure, but seeing the recent heavy development gives me hope.

Oh, MarkuBu gave me the idea for another question:

What about the “mod base” thing? Many people around Minetest Game talk about how Minetest Game is or should be a “base for modding”. Is this really an intended goal (or side goal) for Minetest Game? To be a “base for modding”? And how much so? (Note: As a modder, I would be fine either way.)

[MarkuBu]

MTG is not a mod base at all. It is a game as the name says. Period

A mod base would create an empty world without any items, tools and nothing more than stone

[rubenwardy]

The default mod in MTG is a big mistake. It defines a lot of stuff and it is difficult or impossible to make changes without changing the code.

So true. Unfortunately, we need to worry about backwards compatibility

[rubenwardy]

Have you seen this, @Wuzzy2? #1488 http://dev.minetest.net/User:Rubenwardy/Minetest_Game_Development

I don't think that games need to have a definite goal (ie: the end) but they should work towards tropes/gameplay values and be fun

What about the “mod base” thing? Many people around Minetest Game talk about how Minetest Game is or should be a “base for modding”. Is this really an intended goal (or side goal) for Minetest Game? To be a “base for modding”? And how much so? (Note: As a modder, I would be fine either way.)

celeron55 originally said this (around the time of MTG's feature ban). I believe that this has since changed - we still want to be very moddable (and so provide APIs), but we shouldn't worry so much about being bare-bones minimalistic

[MarkuBu]

Unfortunately, we need to worry about backwards compatibility

True, but...

I wouldn't focus to much on compatibility. I would focus on more flexibility and, wherever possible, add aliases, but declare them as deprecated by default with a deadline to force modders to write new mods for the new mod base or modify existing mods as soon as possible.

Sometimes it is necessary "to cut off old ties", even if that means that some mods wont work. MTG is still there for backwards compatibility, so I don't see a reason for to much compatibility

It's better to make a painful break than draw out the agony.

[paramat]

Yeah although the name may be MT'Game' it doesn't have to be, and sort of isn't, it's a basic universal core for modules. The word 'Game' being incorrect doesn't bother me at all, just as 'Minetest' doesn't, they're just names that stuck and it was too much trouble to change them.

Sorry, my understanding of 'genre' and 'focus' is whether it is fantasy, steampunk, medieval, modern day. What i mean is it's best not to define this because it tends to be any of these depending on what you build or what nodes you use or what mods you use.
Because MTGame is a basic core for modules and is quite universal i can't see how a specific 'focus' is possible.
I agree that it has a 'genre' in terms of being a MC-like sandbox.

I can certainly list features we intend to have and are working towards, see #515 optional simple mobs, weather and ambience sounds (once we have good engine code for these), underground decorations and biomes.

What i mean by 'no 1.0.0' is that development is open-ended, once we have these desirable features we will inevitably be thinking of the next desirable features, so it's almost impossible to define a state of 'finishedness'. In another way MTGame has been usable and playable for a long time so 1.0.0 could also be in the past.

Having a version 1.0.0 in mind is problematic and pointless. I like the idea of either always being 0. ... or jumping far past 1.0.0 by going to 4.16.0 or 5.0.0.

Other types of computer game are easier to define a 1.0.0 for, but this is a basic and universal core for modules in an open-ended voxel game.

1.0.0 does not have to mean “final version”, the “1.0.0” could simply mean that Minetest Game is stable and “complete” enough for the average gamer not to throw it immediately out of the window

Again that would be impossible to decide, some feel it has been enjoyable for years, some feel it needs years of work.

Mods can always be added so it's always effectively a base for modding however complex it gets.
However it's not the 'mod base' many want so maybe a simpler subgame can be added?

[paramat]

Maybe I slowly seem to understand what's all the fuzz with Minetest Game.

Indeed it is fuzzy Wuzzy.
Hehe.

[MarkuBu]

I agree that it has a 'genre' in terms of being a MC-like sandbox.

That's the point. To be a mod base it must have no "genre" at all. That's what I want to say with my post.

[MarkuBu]

I try to make a "mod base" based on MTG. I will ignore any mod compatibility. I want to know what makes sense and what doesn't. No idea why furnace and torch are in the defaults mod. Successfully striped them into separate mods.

[C1ffisme]

Again that would be impossible to decide, some feel it has been enjoyable for years, some feel it needs years of work.

Coming from someone with a small attention span, I feel the game got boring for me around 0.4.9. Perhaps it was the fact that I had gotten Minecraft at the time, or I thought that with the release of 0.4.10 this game wasn't really going anywhere, but I don't really play Minetest the way I would play Minecraft anymore.

(Just as a reminder, 0.4.x has been the latest major version for almost 5 years now. And while version numbers don't always mean anything, I would like to see some 0.5.0 or some other version that brings something new, whether it be strictly for MTG or for the Minetest Engine.)

[Wuzzy2]

Well, I guess I just have to lean back and watch how things will be going around Minetest Game. :P

[sorcerykid]

Perhaps it is time to phase out minetest_game altogether. After all, why does Minetest need an "official" subgame that is endorsed by the Minetest core developers? Is that really the vision for this project?

Wouldn't it be more beneficial to let the community as a whole take the reigns on publishing a variety of subgames which players can download and install via a "Game Store" (analogous to the appstores in Windows 10, iOS, etc.), either within the engine itself or via a Web interface? There could even be a featured "Game of the Month" published on the Minetest homepage, thereby showcasing real-world examples of what is possible in Minetest.

I think having a Game Store would be far more attractive to potential players, rather than trying to promote a "default" game with a single objective. It would also lessen the workload on the development team, so they could focus on enhancing the engine for the plethora of newly emerging subgames.

If there is to be any type of official subgame bundled with the engine, then I would recommend that it be an interactive tutorial for beginners to familiarize themselves with the feature-set of Minetest.

[C1ffisme]

@sorcerykid The problem with that is that which mod works with which subgame(s) would become a nightmare.

• Some internal subgame mods often share names similar to MTG mods, because nobody ever uses mods with their subgame. (At least, last time I checked, which was probably years back.)
• Some people will beg for mod creators to port to one subgame, while others will complain that the mod author is 'spending too much time on porting the mod to a worthless subgame'.
• Mods that currently work with MTG will suddenly be left grasping at straws trying to find which subgames they prefer to get their mod to work with.

And it's not just mod developers either. Subgame developers will have to deal with:

• Without a main subgame, there will be very little "example code" to work with.
• Over time, people will decide that certain internal default functions that mods depend on must be added into the game, despite how they may or may not fit in.
• Having a 'Game Store' would be very convenient, but it also means that naive players will often disregard things like licensing, etc. Besides that, we've already tried doing that... (More details here.)

If there is to be any type of official subgame bundled with the engine, then I would recommend that it be an interactive tutorial for beginners to familiarize themselves with the feature-set of Minetest.

Also talked about before.

Also, about a game store, the same naive players who don't read more than the list of features and maybe supported subgames might not read the 'where to go for support'. In fact, such a place might not even be specified by beginning modders. I forsee minetest/minetest being filled with issues related to 3rd party mods or subgames.

[rubenwardy]

Quality

Just look at boost carts vs carts in mtg

[C1ffisme]

@rubenwardy Oh yes, I forgot about that too in my counter-argument. If we don't have a main subgame, then community created subgames will define their own version of quality. I can't wait to play new subgames based off of the 'meme mod' and the 'mob mod with incredibly bizarre, unfitting models and textures from 2014'. (I'm looking at you, MobFramework.)

[octacian]

Before I get started, these are just personal opinions so I hope that nobody will take offense at anything I say below.

I'm beginning to think this is ridiculous. Though some of this conversation is constructive, while reading through I find most of it _destructive_. Here we are, wasting the developers' time asking ridiculous questions when we could be opening _constructive_ issues and/or pull requests to deal with each issue one by one. Yes, quality is an issue, but rather than just talking about it, do something. This is where open source comes into play.

Above, there was some discussion as to how people don't like the fact the the version has been 0.* for "the past 5 years." Well, I believe there is a reason for this. Minetest and MTG seem to use a modified version of Semantic Versioning. Though this may not be directly purposeful, I see this as being the case. Just because the version starts with a zero, does not mean that it's entirely bad. According to Semantic Versioning, it just means that there have not been major API changes that were not backwards compatible. Though this may not directly be the case, I see it in some aspects of Minetest's versioning. The 0 isn't bad, it just means that the software is subject to change. Yes, I believe that we should have goals to achieve for a major release in the future (e.g. 1.0.0), but zeros aren't bad :P

I personally don't believe that Minetest should directly have an end objective, at least nothing weird like "slaying the dragon." Maybe some sort of achievement system could be added through some sort of mapgen structures. Players do need something to help them push forward, but not so much that it's no longer interesting. For example, I believe that MTG needs a crafting guide, but not in the style that most want. I think the crafting guide should be something that you have to obtain, with basic hints given at the beginning as to how to obtain it. It shouldn't be hard, but it shouldn't be so easy as just clicking a button.

I think that a mod, game, and texture pack store is something that would be very useful and even important. Yes, it was done before, but as mentioned in the link shared by @C1ffisme, it had some major issues. However, rubenwardy is currently developing the backend for a new, what I like to call content database, and I'm trying to make time to work on the frontend. I've also considered (and will soon) writing a Minetest launcher using NodeJS. This way the mod/subgame/texture pack store could easily be implemented, including the automatic download of mod dependencies. A launcher would also allow newbs (and others) to easily compile new versions of Minetest.

Anyways, yes, MTG does need work. But I think this issue has gone further than I would hope rubenwardy originally intended. If there are other issues, I think we should open separate (constructive) issues/pull request to actually _improve_ the game, rather than just talking about doing it.

[C1ffisme]

According to Semantic Versioning, it just means that there have not been major API changes that were not backwards compatible.

As I've said before, while version numbers don't mean much, they still give an indication of a beta stage or testing field. And this further proves my point:

...I would like to see some 0.5.0 or some other version that brings something new...

that we haven't done anything incredibly major for a very long time.

Yes, quality is an issue, but rather than just talking about it, do something. This is where open source comes into play.

I feel as though I'm going to get called out on this, but I agree on you with this point. I feel like we do spend a bit too much time complaining about our lack of workers and not enough writing code. Even the code that gets written gets far too much complaints, as if it can't be modified later. We can't have fun, little features because they must be part of an API. (Okay, I suppose my code is pretty trash, I'm not an expert at dealing with other people's code, much less a whole community's. :P)

[tobyplowy]

that we haven't done anything incredibly major for a very long time.

A good mob api?

[Wuzzy2]

Let's say, if Minetest Game:

• Has met all requirements of https://github.com/minetest/minetest_game/issues/515
• Has at least some mechanism for complete newbies to get started, so they are not completely helpless at the beginning
• Is bug-free or almost bug-free and very stable
• Is very stable for use on servers with no or very few complaints of server operators
• Has a mature and stable API with no or little complaints from modders; breaking future changes are unlikely
• Has a massively reworked health system (The current one is just cheap as fuck: Eat bread for INSTA heal? Just stockpile tons of bread and you can survive even in lava for a long time. It also means that PvP battles are mostly decided on who has more food)
• Does not run out of new gameplay content after 6 hours of playing in a new world (excluding Nyan Cat and Nyan Cat Rainbow)
• Doesn't have features which look incomplete/pointless (e.g. rails but no carts, swords but no enemies, items with no use)
• Less than 50% of players who have an opinion of Minetest Game perceive Minetest Game as incomplete
• Compared to 0.4.15, has at least one additional major gameplay concept (like farming)

I would call it “1.0.0” (ignoring the fact that the current version is coupled to the Minetest version), if I were a core developer here. From this point on, Minetest Game will probably be a pretty decent game already. This is basically just a milestone. Of course, after version 1.0.0, there can still be many future versions, so I see 1.0.0 just as a reference point.

[tobyplowy]

@Wuzzy2 if everyone would think Like You minetest development would be way better and faster and more efficient!

[C1ffisme]

@Wuzzy2

Does not run out of new gameplay content after 6 hours of playing in a new world (excluding Nyan Cat and Nyan Cat Rainbow)

I definitely have to agree with this. Generated structures, new dimensions and algorithm-based mechanics go a long way to making gameplay a little bit more about exploring and experimenting rather than just getting materials off of a checklist.

[BrunoMine]

This is not the solution to the problem of lack of mods.
We need to re-establish MMDB as fast as possible and allow mods, textures, subgames to be downloaded.
An MMDB home page can advertise good subgames. (This would be the solution to this problem).

I'm afraid a better-defined default subgame might create a minetest still image. Currently minetest has many possibilities and it is important that the user knows this.

Edit:
I believe that farming must also be segregated. This is not a very polished mod.

[rubenwardy]

This is not supposed to be a solution to a lack of mods

[BrunoMine]

More access to mods is a solution for a complete base game.

[C1ffisme]

@BrunoMine Handing the game over to 3rd party content, in my opinion, is just the same as giving up. Minetest Game should not require mods, nor any other 3rd party content to be a fun game. Many other open-source games are already fun without mods, and I believe Minetest can do the same. Mods should be there to enhance or change the gameplay after someone believes that they have discovered everything about the game that can be discovered.

[Fixer-007]

Right now MTG is basically a very limited sandbox, you do not even see a wielded tool, Karl...

[numberZero]

This issue is called “Split MTG into two games - one a game, the other a mod base.” Why not just DO THAT? Split off all the APIs, keeping no blocks, even dirt or stone, letting mods define everything. That might be called like Minetest Framework.

The next step is to check which APIs should and which should not form the core framework, and split off the latter. It may appear that too little remains in the core, though. But that’s not a problem, really, that just means the subgame form is not suitable for the framework, it should be collection of libraries that may or may not be included in a full subgame.

And, finally, the third step: ~rewrite~ refactor MTG using these libraries. Now mods should depend on libraries and not on some default behavior, and voilà—MTG is free! As a game need not to be MTG-based to be mod-compatible, MTG can be developed in any direction (and faster)!

So what do you think on that?

[C1ffisme]

Split off all the APIs, keeping no blocks, even dirt or stone, letting mods define everything.

My single complaint about this is that you might as well just stop making MTG a mod foundation and let mods work with their own subgames, because defining their own functionality would probably work better in the long run anyway, rather than having generic libraries that may or may not get used. (Although, I suppose you wouldn't want to force a generic dirt or stone node on a realism subgame, unlike the short lived basic subgame.)

[sorcerykid]

The problem with that is that which mod works with which subgame(s) would become a nightmare. (snip)

C1ffisme, that seems like a remarkably unlikely doomsday prediction, just for abandoning MTG. Am I really to believe that if there was no longer a default subgame, mods would cease to be inter-compatible and subgames would fall into complete disarray? No offense, but I don't share that same outlook. I think the community has a lot to offer, but maybe there just aren't enough incentives.

There is clearly an abundance of developers right here, so it's not for lack of talent nor interest. Perhaps the continuation of MTG is itself deterring others from attempting to develop their own. After all, Minetest comes bundled with its own "minetest_game", and there's no easy way for the average user, without a lot of technical know-how, to download and install and configure another subgame to override "minetest_game" in the first place. So why should anyone bother?

Without a main subgame, there will be very little "example code" to work with.

There are tens of thousands of lines of suitable code examples out there, most of which are available right here on GitHub. Add to the fact, if MTG were separated, it would still remain an independent subgame. Perhaps the Minetest NeXt project could be resurrected:

https://forum.minetest.net/viewtopic.php?id=9066

Quality Just look at boost carts vs carts in mtg

I don't understand the association you are drawing about quality. Boost carts were not even part of MTG until recently. That merely affirms that MTG is always behind-the-times and inadequate.

Oh yes, I forgot about that too in my counter-argument. If we don't have a main subgame, then community created subgames will define their own version of quality.

I would not rely on MTG as a barometer for quality. I think it is disingenuous and self defeatist to suggest that MTG is a pillar of good subgame design whereas community-driven projects are automatically inferior by virtue of not being endorsed by an exclusive team of core developers.

I've actually had to rewrite much of MTG for my own server due to inconsistent, redundant, and/or inefficient code. And, it is not only the LUA side of Minetest that suffers from countless examples of poorly thought-out programming logic and style. With all due respect, I think it's important to recognize that being "official" really has no bearing on the production quality whatsoever.

[Fixer-007]

Just look at boost carts vs carts in mtg

problem? carts = modified boost carts :) Quality can always be improved.

[rubenwardy]

I think it is disingenuous and self defeatist to suggest that MTG is a pillar of good subgame design whereas community-driven projects are automatically inferior by virtue of not being endorsed by an exclusive team of core developers.

It's not the core developers that give the higher quality, but more eyes on the code

[C1ffisme]

C1ffisme, that seems like a remarkably unlikely doomsday prediction, just for abandoning MTG. Am I really to believe that if there was no longer a default subgame, mods would cease to be inter-compatible and subgames would fall into complete disarray? No offense, but I don't share that same outlook. I think the community has a lot to offer, but maybe there just aren't enough incentives.

If you are a modder, you'd be lucky enough if three or four subgames were significantly more commonly used than other ones. If that isn't the case and there is no main subgames of any kind, you might as well give up.

In order to make your mod work with other subgames in a clean way, you would almost certaintly have to rewrite any and all code related to item names and remake any textures in order to fit them within the subgame's art. The texture thing may not sound essential, but in the case of ores, not doing it will look really bad.

Now, this is only supposing that this is a mod contains very simple content, such as tools, items, ores or other blocks. If the mod wants to use sounds (which are normally supplied by the subgame), that's another rewrite required, or a whole different set of sounds has to be bundled with the mod. That isn't too bad an option if you intend to only download one mod, but having multiple mods all with their own sound for "stone-like" nodes means a much larger toll on the loading time.

I could go on, I haven't even mentioned other features such as leaf decay and balance, but I think I'll move on to your other arguments.

There is clearly an abundance of developers right here, so it's not for lack of talent nor interest. Perhaps the continuation of MTG is itself deterring others from attempting to develop their own. After all, Minetest comes bundled with its own "minetest_game", and there's no easy way for the average user, without a lot of technical know-how, to download and install and configure another subgame to override "minetest_game" in the first place. So why should anyone bother?

I don't know if this was only directed towards me, but I kinda agree with you on this, although on the sake of subgames _and_ mods. Of course, my (and many other's) solution is to fix the underwhelming gameplay of MTG, which I think can be done.

There are tens of thousands of lines of suitable code examples out there, most of which are available right here on GitHub. Add to the fact, if MTG were separated, it would still remain an independent subgame. Perhaps the Minetest NeXt project could be resurrected:

You would be surprised how few people use their browser's search bar, either out of paranoia, or laziness. But besides that, telling the user to search for examples is an incredibly bad way to introduce someone to an API. (I know this because this is, unfortunately, the situation with trying to learn how to mod Minecraft. Whilst books and tutorials do exist, I don't even know where people find the material to learn how to create mods like Reika or Team COFH) But besides that, there will always be places with outdated, badly styled or just plain broken code.

I would not rely on MTG as a barometer for quality. I think it is disingenuous and self defeatist to suggest that MTG is a pillar of good subgame design whereas community-driven projects are automatically inferior by virtue of not being endorsed by an exclusive team of core developers.

If you can come up with a subgame of reasonable quality that I can play for at least an hour without feeling bored, I'll agree with you on this one.

[sorcerykid]

If you are a modder, you'd be lucky enough if three or four subgames were significantly more commonly used than other ones. If that isn't the case and there is no main subgames of any kind, you might as well give up.

I think this brings up an important point: Why are similar arguments not raised for why there needs to be an official Minetest server, if there has to be an official Minetest subgame?

After all many server operators already maintain their own subgames, and sometimes that requires significant customization beyond what some existing mods provide. Yet plenty of servers are thriving despite being independent of minetest.net. For sake of analogy, I can only imagine if there was suddenly an official "minetest_server" that players were automatically connected to upon launching Minetest, and the only way to login to other public servers was to use a complicated installation and configuration procedure (like what we have now for subgames).

If you can come up with a subgame of reasonable quality that I can play for at least an hour without feeling bored, I'll agree with you on this one.

The "just test tribute" subgame used on the JT2 server seems to have a high degree of player retention. And, while I admit to being somewhat biased (since I'm a developer for that subgame), I would consider it decent quality. The server consistently has 30-50 users at all hours of the day.

If, however, the gauge is specifically "Does C1ffisme get bored?", well that's not a particularly justified point of reference for the entire Minetest userbase :P

[BrunoMine]

An hour is indeed an exaggeration. But I felt bored after about 3 hours. Probably on account of not knowing anything about the game, only that it could be a clone of minecraft.
The servers were the real fun of the game until this day.
(particularly)

[Wuzzy2]

Just Test is more than just a server. It's a different world. It's a different game. :-)

After all many server operators already maintain their own subgames, and sometimes that requires significant customization beyond what some existing mods provide.

There are no server operators in Minetest, only game developers who disguise themselves as server operators. XD I hope more server operators would also publish their subgame somewhere, in the spirit of free software.
So far, I don't know if anyone has actually published a subgame optimized for servers/multiplayer. The things you find on forums are mostly singleplayer-only.
Plain Minetest Game is not really for multiplayer, because there is no real protection system. So it is very chaotic in nature, no building is safe. It is for LANs at best. I also can't recall if I have ever seen a pure Minetest Game server, without any additional mods.

Maybe we should write down a list of things or mods which should be added into Minetest Game. And maybe a list of things which shouldn't.
This begs a question: Should a protection mod be added? Or do you think it is good as now? Maybe do you actually WANT a chaotic world? :-) This is a serious question.
But it's a bit strange that a lot of code in Minetest Game is actually compatible with protection mods but does not actually include a protection mod on its own...

a complicated installation and configuration procedure (like what we have now for subgames).

This is the second time you now rant about subgames, but I don't understand you at all. What is complicated about extracting a single Zip file into games/? o_O

It's actually even easier than installing and using mods since you don't even have to get the dependencies right.
Maybe the problem is that subgames, or at least the concept of subgames, are just not advertised well enough. They are only briefly mentioned on the homepage, but there isn't even a single screenshot of a subgame or even a mod which is not Minetest Game. This is very sad because would-be players are missing out on interesting things beyond Minetest Game.

The main menu could still use some improvements in general because it has many gotchas, but I think Minetest has WAY bigger usability problems than subgame handling.

[sorcerykid]

So far, I don't know if anyone has actually published a subgame optimized for servers/multiplayer. The things you find on forums are mostly singleplayer-only.

You do realize that maikerumine has published all of his subgames on GitHub? They are tuned almost specifically for multiplayer, as they were the basis for all of his servers to date, altho to my understanding, they function perfectly well in singleplayer too :)

https://github.com/maikerumine

This is the second time you now rant about subgames, but I don't understand you at all. What is complicated about extracting a single Zip file into games/? o_O

Sorry, my concern about installation of subgames was moreso reflecting the average user that just wants to start playing Minetest without delving into the directory structure (which I think we can all admit is a bit unorthodox). In Windows, after all, people are accustomed to installers and App Stores. I can't even imagine what the procedure is for Minetest on Android devices.

When I first started playing Minetest, for example, I had no idea there even WERE alternatives to the default subgame, moreless the ability to override the default subgame. There is really nothing about the user interface within the Singleplayer menu that makes it clear what a subgame truly is. That to me is a drawback for attracting new players to Minetest, each seeking a slightly different gameplay.

In contrast, you just click over to the Servers tab and BAM. There's a list of new and exciting worlds you can begin exploring and building. Nearly every one presents a unique gaming experience. That's how easy (and fun) it should be to switch between subgames.

[numberZero]

[WIP] https://github.com/numberZero/minetest_game/tree/dev-framework

[MisterXtreme]

OK, I'm gonna toss my hat into the ring. :)

First of all, for the default subgame/engine we ask to add a feature, and we are told "no" because It's "too much like Minecraft", Minecraft is one of the best selling video games, has lots of modders/players even with a not so good mod API, why? because of they're ideas.

I'm not saying "lets clone MC", we have subgames for that, but not having dimensions, or double chest because Minecraft does, sounds like backwards thinking.

Also, the launcher needs an overhaul, we are a mostly mod-focused game, so why not have a launcher made for that, like FTB, or Technic, etc?

From it you could:
Download modpacks
Download texture packs
Download maps
Options to enable/disable/add mods to a modpack.

OK, I'm done. :)

[tobyplowy]

@XtremeHacker If everyone would Think Like This minetest Development wuld accelerate and Go beyond the level that anyone would have imagined (it can even kill minecraft) but sadly It can't be

[octacian]

Well, here is my final opinion.

Yes, MTG should be split. Minimal should obviously be kept, but MTG should be split into a base for modders. It should contain only the basics, with only a few tools or weapons. I would even suggest that many of the other blocks be removed so as to keep it as a minimal (yet not as minimal as minimal) example and testing game.

With MTG split into a second mod base game, we wouldn't have to be held back because the features people wanted did not comply with the idea of being a mod base. This would allow us to actually make Minetest "Game" into Minetest Game.

Anyways, somebody should fork MTG and remove those extras and replace some default things with things that are specifically for the purpose of being examples. I thought somebody had started on this before, but I haven't seen any progress. If nobody does it, I'll do it at some point but I'm already quite busy with other things ATM.

[MarkuBu]

And my opinion is that MTG is neither a good game nor a good mod base.

To complex and monolithic for a mod base, no real gameplay content for a good game

[MisterXtreme]

@MarkuBu Pretty much, and no one knows what should actually be added because of that.

[MarkuBu]

That's why I try to split MTG into parts to make a real minimal game for modding

[paramat]

We're working on making MTG a reasonable subgame, it's not as though we are resisting doing this, it's not so easy when we're understaffed and overwhelmed with work, unless you're a core dev you will not understand the workload. The review process slows things down too, a subgame made by an individual or small group can progress many times faster and will be more consistent in design, which is why we need a variety of subgames.

I agree a new mod base is a good idea and MTG is not this anymore.

So, let MTG continue improving as it has been, it doesn't need to be split or disrupted, but lets have additional subgames and a mod base.

[Desour]

I'm not sure if a new modding base subgame would be useful.
The players would want to have the convenient features of the normal game and additional the features of mods. => Mods would still base on normal game and not on modding base.

[MarkuBu]

A modding base is not so much for players. It's more for subgame makers

[numberZero]

@MarkuBu Do you like my approach? https://github.com/numberZero/minetest_game/tree/dev-framework
@DS-Minetest If several games would share some APIs, mods would rely on that API and not on game-specific things, when possible.

[MarkuBu]

@numberZero can't tell yet. My goal is to remove as much as possible out of the default mod and make separate mods

[numberZero]

@MarkuBu That’s one of my goals) But I’m moving code out of other mods now, separating API implementation from game-specific stuff.

BTW, the fact that mod is called default is very strange. MT is an engine, it should support many games with different APIs, why should any be default? And why should it be a game and not a framework?

[MarkuBu]

I have a serious question about MTG: what is it's goal? It can be used as modding base, but it is far from an ideal modding base. But with the current developing speed and the conservative strategy it will never become a game that catches newcomers.

I think, Minetest lost a lot of players who downloaded Minetest and MTG, play for an hour and dumped it, because they didn't understand the concept.

[MisterXtreme]

@MarkuBu I've noticed that too, maybe we could go a route like this:
Minetest Game, game features will be added like music, game items, ores, etc
Minetest Base would have any APIs/features that get added to MT game, but skip the nodes, ores, etc
Minetest Minimal is for people who just want an engine.

[Wuzzy2]

Maybe the real solution is to stop promoting/treating Minetest Game as the “real”, “default” or “canonical” subgame. Instead, just treat it as a subgame like every other subgame, and also put other subgames into the spotlight. Technically, there is absolutely nothing special about Minetest Game. The fact that engine developers also develop Minetest Game does not make it special in itself. Minetest Game has FAR too much attention compared to other subgames. I'm not saying Minetest Game should not be promoted, I'm just saying that it should not be promoted as the “one and only subgame”. ;-)

Currently, the only way to get real information about other subgames is from the forums and NOT the homepage. This sucks. Yes, there's a subgame section but is is very small, as if subgames would be just an insignificant side feature. Thus, other subgames stay in the shadow. This is sad because there are a few interesting subgames out there.
minetest.net leaves the impression that Minetest = Minetest Game. Hell, even the very first sentence claims that Minetest is the engine AND the subgame (this needs to change). Although some subgames are mentioned, they are are only mentioned briefly, and also don't have any screenshots.

Minetest.net should give other subgames at least a chance. Also, from the homepage it should be MUCH clearer that Minetest is much more than just Minetest Game.

Also, I think Minetest Game is not a good choice for the default subgame anyway because it lacks in depth. It slowly gets better over time, but I think it is lack of depth does not make it ready for “prime time”.

Another benefit of this solution would be that there is no longer a reason for us players to pressure the poor developers to any particular development direction. ;-) Also, a split into two subgames would not be neccessary, Minetest Game will just be what the developers want it to be.

[MisterXtreme]

@Wuzzy2 I get where you are coming from, and you make good points. :)
My 2C.
The website _and_ launcher need to show that there are other subgames, kinda a place in the launcher to download sub-games.

Also, Minetest game is the base of most sub-game, is Minetest game a game, or a barebones modding starter, I say bare bones because mods made for a game similar to Minetest (Minecraft) has a much more developed main game, and has lots of mobs _because_ the game is well developed.

Also, I'm sure more people would contribute to take a load off the core dev's back if we had an idea on what actually will get merged, and what won't. Which at the moment, doesn't happen _because_ minetest_game has no real "direction"

OK, that was my 2 cents.

[rubenwardy]

because minetest_game has no real "direction"

It does #515 #1488

[paramat]

I think a modding base is useful, it will be small anyway so might as well ship it as a subgame and see how it gets used. MTG is no longer a mod base.

I have a serious question about MTG: what is it's goal?

It used to be considered a modding base and we were waiting for more exciting subgames to be shipped with MT, but they didn't appear so we (or at least i and sofar) decided to work hard on adding content to it to work towards something more impressive and more playable. So it's sort of inbetween mod base and complete subgame.

But with the current developing speed and the conservative strategy

The slowness is frustrating, for me too, it's due to the branch, PR, review process and lack of devs and dev time, an individual writing a subgame outside the process can develop something much faster.
It's not particularly conservative, see what has changed over the last 2 years and all the new biomes and new nodes.

I think, Minetest lost a lot of players who downloaded Minetest and MTG, play for an hour and dumped it, because they didn't understand the concept.

Yes, they didn't realise mods are needed.

I agree that new subgames, probably developed fast by an individual outside the review process, are needed to impress, then MTGame will be the carefully-devved but less exciting subgame in the middle, then add a mod base too.
So as before i'm saying leave MTGame to progress as it has been and add 2 more types of subgame, a 3-way split that avoids MTGame trying to either be a mod base or an exciting complete subgame. That is why i disagree with the title of this thread.

So i agree with Wuzzy here. MTG is carefully devved with a strict process, it can be the 'classic, dependable but not particularly exciting subgame', something like this is valuable to ship alongside other subgames.

[MisterXtreme]

@paramat What do you think of the idea of adding the ability to downloads mods/modpacks from GitHub, or an other place with a direct link?

Then we could have "featured" subgames that get more downloads, and have an "on-official" Minetst full on game.

Any thoughts on that idea?

[paramat]

I don't like the idea of direct internet links from the game, security issues etc. But there is such a thing as the 'minetest mod database' that may be revived.

[MisterXtreme]

@paramat Maybe we could ask if any community members would like to help with a vetting system (I would)?

EDIT: Or even just "pages" in the launcher that lead to the forum thread at the least?

[paramat]

I guess the Mmdb is meant to be that safe place to direct-download from. There should be an issue or PR thread for it's revival somewhere here.

[MisterXtreme]

@paramat Exactly, then we can have modpacks, but not have to have them bundled.

[MarkuBu]

As I mentioned in #1606 is was easy for me today to update sfinv and screwdriver in my subgame, but adding the new blocks is difficult because I had to modify default. It would be easier to add the new nodes if they where introduced as mod.

So I had the idea to split MTG into mods _step by step_ instead of developing a new modbase, which is a lot of work. This could be done while MTG moves still forward and without breaking compatibility to existing mods

[paramat]

split MTG into mods step by step

Rubenwardy has been trying to do this already and has found it difficult to go further than where we are now. Still worth trying though. Doing this for the sake of avoiding developing a mod base subgame is the wrong reason though, i feel a separate mod base subgame is probably a good idea.

[MarkuBu]

I don't think it is difficult. It is just a lot of work. And I think rubenwardys* approach was not optimal (also mine)

lets the modding base aside. A more modular approach would be also good for MTG. As I mentioned, adding the now sandstone nodes is a lot of work if you don't want/can't update the complete MTG folder. If they are added as mod everyone could add them to his game or server

• Not rubenwardy. I meant numbesZeros dev framework

[MarkuBu]

I have some other serious questions: why do we need a modding base? Is'n the Minetest Engine itself a modding base?

What are most subgames currently? They are mod collections based on MTG, sometimes not even with a singel change in the code.

What should a modding base provide to the subgame developer? Nodes? Biomes? Plants? What? Everything in a subgame could be different so everything has to be replaceable just like Lego

If we discuss about a modding base we should discuss what this modding base should provide to the developer.

It used to be considered a modding base and we were waiting for more exciting subgames to be shipped with MT, but they didn't appear

To be honest, I think a modding base wouldn't be a success, except it comes with a functionality more or less like MTG, but build like a Lego toolbox.

So the answer for my fist question: Laziness of devs, time saving for the devs and easy starting point for newbies.

[rubenwardy]

It'd difficult to do it whilst maintaining compatibility

[MarkuBu]

Well, this is another reason why a good modding base will fail. A lot of mods depend on default and I highly recommend that there is no "default"

@rubenwardy do you have a branch? What exactly was difficult?

[rubenwardy]

See the books pr. Aliases aren't all powerful, and you need to maintain the dependency tree roughly

[paramat]

Having a separate mod just for sandstone nodes seems to be an excessive level of modularity, there would be hundreds of mods and dev work would become more difficult.

[MarkuBu]

I really didn't talk about single mods for each node, but a mod for nodes. If new nodes are added we just need to update that mod and not multiple files in stairs and default mod. And we could have started with the sandstone nodes

[sorcerykid]

I agree 100% with paramat. Moreover, I'm not sure if a mod for nodes is even the ideal approach. What about nodes that have a very distinct purpose, like say the protection block, that include an extensive code base and a wide-variety of assets and wouldn't really make sense being consolidated with any other mods. At that point, we're back at square one.

That's not to suggest your idea doesn't have merit, but I think this type of modular approach you are proposing will require significant forethought to account for all manner of scenarios.

[MarkuBu]

Well: then another serious question: what are we talking about? A modding base has to be modular to make any sense. But this would mean that a lot of mods won't work

I think we have two choices:

1. we make a modular modding base and force the modders to adept their mods to that mod base or
2. we close this issue and continue as before

Just to clarify. I'm not upset. I only sum up the facts

Currently I start a new approach for a modding base from scratch, without any default mod

[MarkuBu]

Here is an example what I mean if I talk about a modding base

https://github.com/MarkuBu/minetest_modbase

[Ferk]

a mod for nodes

There's a file for nodes, "default/nodes.lua"
if your intent is to copy the nodes mod from MTG into your game you could as well just copy nodes.lua.

Maybe it could at least be separated in files inside of the default mod (like a "nodes_mineral.lua", "nodes_organic.lua", "nodes_crafts.lua"). This way no dependency problems would be created and you can still decide which files to copy over.

Also, Minetest is very flexible, you can override/unregister whatever you don't want in your game/modpack.

[MarkuBu]

I don't want several "files" in one mod. They are difficult to maintain/update and I have the same situation like now

What I don't like at Minetest Game is, that it is often not only one file.

For example stone. The node registration is in nodes.lua, the crafting recipes are somewhere in crafting.lua, the cooking recipes are also in crafting.lua, but at the bottom, stairs and slabs are registered in the stairs mod. So if you want to change something you have to maintain 3 files.

I put everything in one mod. Everything stone related is in one mod. Registration, crafting, cooking, textures and maybe even sounds

[Wuzzy2]

The default mod is a terrible bloated clusterfuck with way too much content.
The reason why we have mods in the first place is to have some sort of dependency management.

That default is bloated is very bad. It forces modders to almost always depend on “everything and the kitchen sink” for minor things.

Just think of these (seemingly) simple use cases:

• You want to register a node? de facto default, because of the sounds. You probably don't want to ship your own sounds for every node, or even have a silent node.
• Need the player model? default!
• Want to add some standard GUI? default!

Basically, no matter what you do, you almost always need to depend on this mod, which is bad, because you can't JUST depend on the sounds or JUST depend on the player model. No, you have to depend on dozens of nodes, sounds, the player model, fences, ladders, GUI definitions, and so on.
The only times where you can actually get away with not depending on default is when you add custom GUI or chat commands or some other things not related to gameplay.

Another big problem with default that its scope is undefined. There is apparently no limit to what can be added to default. As a consequence, this mod just grows and grows and the bloat is becoming worse. And once a thing is in default, it rarely gets out.

If mods were classes like in object-orientated programming, default is a prime example of an anti-pattern known as the “God Class”.

But I also think a mod just for nodes is not a good idea either. A mod should more like contain things which are closely related and/or make dependency management easy. One mod = one task. This is my rule of thumb. Good examples for this are doors, tnt, xpanes.

These were just my 2 cents.

[MarkuBu]

Doors is a good example. Do you add the registrations to the doors mod or the wood? Or steel?

So if you want to make a stoneage world without iron you still have the steel door registration in your doors mod

Such a mod should only provide the API, not the node registration itself. Same for stairs, fences ...

[numberZero]

@MarkuBu

Such a mod should only provide the API, not the node registration itself.

https://github.com/numberZero/minetest_game/tree/dev-framework/mods/libdoors

It seems we should cooperate =)

[C1ffisme]

@Wuzzy2 You forgot to mention that not only is default the mod where oddly chosen 'mod necessities' are stored, but also several miscellaneous items, such as furnaces, and for a while, nyan cats. As well as things that would better belong in mods like underground, overground, plants, etc. And the mods that are seperated from default are usually too small to warrant a whole mod on it's own.

I'm thinking there might be a second option to splitting MTG, which is to create a new subgame entirely, without some of the bizarre choices that have been made over the years. It would be a painful process, but it would have significant benefits. Just gonna say that it's an option, not the easiest or the least painful, though.

[Wuzzy2]

While I hate the design of the default mod, we also have to acknowledge the sad truth there is a ton of mods which heavily depend on this mod. So any major redesign is not without consequences. :-(

[MarkuBu]

So any major redesign is not without consequences.

That's the point. As I said yesterday

https://github.com/minetest/minetest_game/issues/1485#issuecomment-284760052

@numberZero yes, we can. But your approach is a little bit different. You just move the registration into a new mod. My idea is different.

First I want to add doors for each wood. We have fencegates for each type of wood, but only one wooden door and one wooden trapdoor. How lame is that?

Than I want the registration and recipes to the related wood. So Acacia wood door, acacia wood trapdoor and acacia wood fencegate to acacia wood, pine wood door, pine wood trapdoor and pine wood fencegate to pine wood and so on

[0-afflatus]

Haven't read the whole of this yet ...

In order to make your mod work with other subgames in a clean way, you would almost certaintly have to rewrite any and all code related to item names and remake any textures in order to fit them within the subgame's art. The texture thing may not sound essential, but in the case of ores, not doing it will look really bad.

This is correct. There is a variable amount of work to make a mod work with games other than MTG. It depends on the dependencies [sic]; most mods are really MTG mods.

So far, I don't know if anyone has actually published a subgame optimized for servers/multiplayer. The things you find on forums are mostly singleplayer-only.

@Wuzzy2 you clearly are not seeing what other people are working on - you have to dig quite deep into the forums in order to discover what other game designers are doing.

In order to make a different game you really need to start from scratch, which does make for a lot of work. Even if you base your game on minimal you will end up making something that looks and feels like a MC clone. This is partly because the MT engine is not purely an engine, it's a fudge containing quite a lot of game code and development is primarily focused om MTG. Many core functions are not generalised enough to make really original subgames possible.

All the above is really an engine issue and beyond the scope of this discussion.

[rubenwardy]

I agree wholeheartedly the the default bloat is a problem. However, when trying to rectify this I've encountered resistance - mainly from either not being able to maintain backwards compatibility, or from those who feel like it's not worth it. Aliases aren't all powerful, mods still compare against the item strings which causes problems

[0-afflatus]

@rubenwardy

It'd difficult to do it whilst maintaining compatibility

I'd say it's almost impossible.

[0-afflatus]

This is OT but also relevant:

I spent a week recently creating a stripped-down 'minimal' type game, which I plan to base the next version of Grailtest on. The 'default' mod contains only player.lua and some gui stuff. I have refactored most of the actual nodes into other mods (I'll push this out sometime soon if anyone cares to look). A reasonable degree of compatibility with MTG mods can be achieved using aliases, so generalised mods with few dependencies are fairly easy to integrate.

From my experience of doing this I'd say it's pointless splitting up MTG. Either accept the mess that it is and carry on developing it into a complete game or start from scratch and consider modifying some of the core functions in builtin as well. If you're going to dig up the roses, you may as well redesign the whole garden.

[MisterXtreme]

I'd say that we should just embrace Minetest_game as a game, Minecraft, which is a similar game is more complete, and has more, not _less_ mod developers because of that, as for people who want something more minimal, there is the minimal game for using MT as an engine, any flaws with my logic, I'm OK with anyone pointing it/them out. :)

[ghost]

Dependencies on default are definitely a problem, I think at very minimum the player needs to be moved into a separate mod. Everyone making a subgame will likely want to have a 3d player, right now that means needing the rest of default as well.

EDIT: I opened an issue for this #1615

[MarkuBu]

Well, dependencies are not a problem. Most mods need dependencies to work.

Mods will have much more dependencies to my mod base approach. If you think, it will be possible to make universal mods for MTG and a whatever looking mod base, you will be disappointed

The truth is, MTG would have to be ported to this mod base too

The only way would be to split everything into mods, but overwrite the mod names with default

[tobyplowy]

People just need a good game so they can make good mods (they need something to bace them selves on)

[MisterXtreme]

@tobyplowy Precisely.
No good base - mod authors implementing what should be a standard feature themselves, and then everyone has the "best" idea on how to implement the feature, causing the massive fragmentation we have (hint, hint, mobs, hint).

[rubenwardy]

Well, dependencies are not a problem. Most mods need dependencies to work.

Misuse of dependencies is a problem - this is what happens with default. By having everything in one mod, it makes it hard for mods to depend on certain features - thus making multi-subgame support harder

For me this is less of a organisational problem, or an update problem (git is quite good at managing changes) but a dependency problem, and a subgame development problem - having it all in one mod makes it hard to disable/rip out a certain feature.

[sofar]

A long breath.

There is an (un?) surprising amount of support for doing a split.

If I just quickly look at the calendar, we're not going to do anything quick, so this is something we should do post-0.4.16.

However, we can decide already what to do going forward. And we could do so already now, even.

We've got several choices to make, though:

• archive minetest_game or not? (1)
• continue minetest_game as the aforementioned "foundation" (2)
• name, and start a new singleplayer game (3)

As we want to assure that existing worlds remain functional, I think (1) archiving mtg would be critical, and we would have to keep shipping it in the future. Therefore, we can't use the name going forward.

If that is the case, then (2) means that the foundation becomes a new git project.

And then (3) also becomes a new git project, and needs a new name.

The foundation project can continue to support existing multiplayer servers as long as it exports default and the other interfaces.

The new survival game can start with an entirely clean slate.

I'm willing to put my time into this after 0.4.16. I think we should just do it as I described above. Thoughts?

[Fixer-007]

minetest_game is still a pure sandbox without goals or any survival and most of server owners mod it to hell.
minetest_game can be renamed to minetest_sandbox or minetest_legacy, but new game will need to have proper, hard survival with stuff mentioned in https://github.com/minetest/minetest_game/issues/515 (modular enough to be usable for modders ofc to change parts/pile on top/whatever), game should be playable in singleplayer or vanilla state too. I'm not sure if it needs to be necessary linked to minetest_game, it could use it as a base or depart in some ways.

Brainstorming.

Maybe there must be a new github topic that will have some roadmap/outline/TODO/how exactly it will be done thing?

[sofar]

We can't rename it (we want old worlds to remain playable!) so why would we copy it verbatim? It seems that we might as well lose some weight and clean it out.

[MisterXtreme]

I'd say that starting a new subgame, and just cherry picking a bit for a brand new (real game) subgame would be good, mtg would remain for legacy purposes, and we could have a game, with less weight, and no bloated default mod, and (I'd hope) polished gameplay, I wouldn't mind giving a hand, I'm not good at coding, but I'm pretty good at art, and just trying to think out things.

[paramat]

As we want to assure that existing worlds remain functional, I think (1) archiving mtg would be critical, and we would have to keep shipping it in the future. Therefore, we can't use the name going forward.
If that is the case, then (2) means that the foundation becomes a new git project.

Why 'archiving' MTG? It can continue to be refined and improved while being kept simple.
I'm ok with MTG staying reasonably simple so that it can continue to be our modding-base game, my intents for MTG are just basic stuff like improving biomes and textures. Because MTG is and has been our modding-base i would be ok with not trying to make it a complete and highly featured subgame.

I like the idea of 2 new subgames: a simpler modding-base and a more fancy subgame, being added either side of MTG.

[MisterXtreme]

Just a question, how does a "simplistic modding base" work, do you mean a simple environment for testing mods/developing mods, because that would probably be minimal, most mods want to build on top of a game, and would prefer to use a "newer" "better" MTG, and for those making mods that aren't like that, there is always minimal, or starting with a completely clean slate, and making your own subgame.

I guess I don't see the point, or reasons behind a "modding base", can anyone give me some good uses, that minimal can't do?

[numberZero]

As for me, a modding base is not a game, it is a set of libraries (that is, mods providing little to no content but good API).

[numberZero]

And that should not be called a modding base, but a game base instead: it’s purpose is to let people create good subgames (not based on MTG) compatible with many good mods, and mods compatible with all these games. (MTG should be based on it too)

[Ferk]

That kind of "game base" wouldn't even need to be included in the default install, imho. Since it would be too "minimal" for the average player and it'll be only useful for subgame makers.

Maybe it would make sense to strip MTG to the basics and move its more meaty content to a modpack.
Then the new MTG would be the combination of the "game base" + "new mod pack", fused together in each release.
There would be no code duplication, division of efforts or conflicting APIs between both projects.

I imagine people will want to make their mods for whichever is the most popular subgame, even if it's not the one labelled "foundation" / "modding base" / whatever.

If there's "a more polished and complete base game" (as per the issue description), that's probably the one that will get the attention of modders.

I have the feeling that the main reason why there's so much support for this split is because for a long time whenever there was a proposal for including some new content it was shot down with the argument of "this should be in a mod, minetest_game should stay basic, no need for new ores/fences/whatever".
If that's no longer the case then not sure if it matters too much anymore.

[MisterXtreme]

Ferk, you hit the nail on the head for me, now, what do some core devs think of actually doing this?

[Wuzzy2]

minetest_game is still a pure sandbox without goals or any survival and most of server owners mod it to hell.

The official name is “Minetest Game”, not “minetest_game”.

[Ferk]

The official name is “Minetest Game”, not “minetest_game”.

"Minetest Game" is how it's called now. My sort of "quote" is from the past :P

[sofar]

@paramat nothing against "maintenance" and perhaps small incremental changes, but I think it'll be clear later on that making progress in a less restrictive environment will be more appealing and rewarding.

[paramat]

Yeah no problem. I would be happy to just maintain and refine MTG as it is now, no big additions, just improve biomes, flora etc.

[rubenwardy]

Looks like we're going to keep with one game, but split up default to make it easier to disable / override things.